The present invention relates to a code division multiple access communications system (CDMA) using spread spectrum modulation which is able to separate and remove intra-cell and inter-cell-interference-noise having an especially detrimental effect among noise admixed in the transmission process, to enhance the frequency-utilization-efficiency, and to decrease the power-bandwidth-product. Here, a mobile communications system where the spread spectrum modulation is applied to BPSK signals is taken as an example, to describe such technology as to avoid intra-cell and inter-cell interference and to describe multi-ary modulation technology, etc.
A Spread spectrum communications system is of using spread spectrum modulation technology where spreading sequences are modulated by transmission data. Due to this spread spectrum modulation, a data spectrum with a relatively narrow band width is spread to a wide frequency band and then the spectrum spread signals are transmitted. And in a region (cell or sector) where a base station (BS) provides communications services, there are plural user stations. Such a communications system is excellent in that a low transmission power per unit frequency is consumed, disturbance to other communications can be kept at a relatively low level, and the system has inherently strong capability resistant to ambient noise (external noise like CW noise) admixed in the transmissions process, and to inter-station-interference-noise incoming from interfering mobile stations other than a desired station. Generally this interference-noise includes delayed waves incoming via multipaths from the interfering-stations. However, because communications being performed by numerous stations share the same band, there is a problem such that increasing the number of users to be accommodated becomes difficult, due to the disturbance made by the inter-station-interference-noise. That is to say, it means that so many users can not be accommodated that a frequency-bandwidth to be occupied increases due to spread spectrum modulation, and it results in a decrease in frequency-utilization-efficiency or an increase in power-bandwidth required for 1 bit transmission.
A receiver RX sends a spread-spectrum-modulated signal to an amplifier 3 via an antenna (not shown in the figures), amplifies the signal up to a required power level, and then frequency-mixes the amplified signal with a local signal fL=(=f0) which a local oscillator 4 has provided. Receiver RX then produces a low frequency component of the resultantly mixed output-signal as a baseband received demodulated signal r(t), by passing it through a low pass filter 5. In this case, coherent demodulation is generally performed.
Baseband demodulated signal r(t) is inputted to a multiplier 7 with the same spreading sequence generated by a sequence generator 6 as the sequence used by above-described transmitter TX. The resultant multiplied output is then integrated by an integrator 8 over the period of the spreading sequence length (1 symbol-frame), to obtain a matched filter output. A received binary data {circumflex over (b)} is detected by making this output on the hard decision. This function compares it with a threshold level by a detector 9 at the end of the above-described frame period (correlative demodulation). A control signal made based upon this detected data is inputted into the control terminal of sequence generator 6 via a synchronization detector 10, and the generating timing of the spreading sequence is controlled so that the sequence phase may be synchronized with that of the received signal. Furthermore, in receiver RX shown in
where ci(i=0,1,2, . . . L−1), δ, and q(t) are the i-th chip amplitude of a spreading sequence, the delta function, a square-wave-function, respectively. An actual radio-band transmit-wave sa(t) is generally produced by modulating a carrier-wave by baseband transmitting waveform s(t)[g(t) or −g(t)] having a bandwidth which is limited less than Te=Te−1.
Furthermore, when a square waveform is used as q(t), inter-symbol-interference takes place. To avoid this phenomenon, for Eq. (3) such a waveform is used that the auto-correlation function may take zero at the adjacent sampling points (called an f-type-sampling-function whose DFT conversion has cosine roll-off characteristics). In this case, if a receiver prepares the same waveform q(t) as that of the transmitting side, and performs correlative demodulation by multiplying a received symbol-frame by q(t) and a local carrier-wave. As a consequence, the desired wave components of the received signal will be obtained as an impulse sequence indicated in Eq. (1), where it is not subjected to the interference due to preceding or succeeding chips. That this impulse sequence is dispread with spreading sequence g(i) and then integrated. The n-th transmitted information bn can be detected, by making the integrated output on the hard decision. If an f-type-sampling-function with a roll-off factor of α (generally taking a value between 0.1 and 1) is used, the radio bandwidth BS of the above-described spreading signal and the radio bandwidth BD of the transmit-data is given by:
BS=(1+α)/TC
BD=(1+α)/(LTC)=BS/L. (4)
However, L>>1 is required in a practical system, and in spite of using a bandwidth L times larger than that of one user, for the maximum number K of simultaneous calls (the number of above-described user-station), K<<L is required. Therefore the simultaneous transmission capacity/Hz is (K/L) times as large as that of a time-division-multiple-access radio system (TDMA). Consequently, there is such disadvantage that existing CDMA systems have frequency utilization-efficiency not higher than that of time-division-multiple-access system.
Thus, the reason why the number K of user stations compared to the sequence length cannot be set to a large value is that the cross-correlation-coefficient between sequence g0(t) assigned to a desired station and a different sequence gk(t)(k≠0) assigned to the other mobile station cannot be sufficiently small. Furthermore, considering delayed waves due to multipath, the cross-correlation-coefficient between many delayed sequence groups coming from respective user stations and sequence g0(i) increases. That is, that the code-error-rate considerably deteriorates, because many interference waves with high cross-correlation come in, and therefore it is impossible to increase the number of user stations K and to raise the frequency-utilization-efficiency.
To reduce the disturbance due to the above-described interference-noise, many methods have been studied so far, the sufficient noise suppression effect has not been obtained. Here let's explain 4 kinds of preceding techniques having close relationship with this invention.
(A) Selective transmission system with parallel combinatorial sequences [Reference: S. Sasaki, H. Kikuchi, J. Zhu, and G. Marubayashi, “Error Rate Analysis of Coherent and Differential Multiphase Parallel Combinatorial Spread Spectrum System,” IEICE Trans. Fundamentals, Vol.E80-A, No.7, pp.1196–1203, 1997-07.]
(B) Isolated pilot assisted analytical demodulation system [Reference: Mitsuhiro Tomita, Noriyoshi Kuroyanagi, Naoki Suehiro, Shinya Matsufuji, “A Pilot Frame Assisted CDMA System using Gold Sequences”, Proceedings of WPMC'99, 9.4, pp.346–353, Amsterdam, 1999-09]
(C) Rake reception system [Publication: Ramjee Prasad “CDMA for wireless personal communications”, Artech House 1996, Sections 11.5 and 11.6 (p348, pp365–366)]
(D) Space-time coding system [Reference: Ben Lu and Xiaodong Wang, “Iterative Receivers for Multiuser Space-Time Coding Systems”, IEEE JSAC, VOL.18. No.11, November 2000]
In system (A), a transmitter prepares a sequence-set consisting of M kinds of basic spreading sequences, selects γ sequences in it, multiplies them by a set of binary information, and adds up all the resultant sequences to transmit. In this case, because a binary information-set of I bits to be transmitted corresponds to combinations which are determined by M and γ, the information quantity per symbol-frame is given by:
I=γ+log2(MCγ) (5)
Consequently, this system is an example of Multi-ary systems and the improvement in frequency-utilization-efficiency has been expected.
When a transmit-symbol-frame arrives at a receiver, generally it turns out to be (J+1) waves consisting of the main wave and J delayed waves that are caused by multipaths. If the number of users is assumed to be K, the receiver receives (γ−1) waves (sequences) from the desired station as a self-interference-component, and γ(K−1)(J+1) waves from the other interfering stations as an inter-station-interference-component. Therefore the disturbance due to these interferences is considerably large in general. And this system requires many (MK kinds) spreading sequences, and the average value of cross-correlation-functions between an arbitrary pair of sequences in the sequences increases, as the family size MK increases. That is to say that the interference power increases in proportion to M, K, and J. Consequently, it is impossible to increase M and K to a large extent, in order to obtain desirable error-rate-characteristics by using system (A), as a result, it is impossible to raise the frequency-utilization-efficiency.
In system (B), a transmitter of each user transmits isolated pilot-frames so as not to suffer the disturbance of interference waves due to the other users. A receiver receives these isolated pilot-frames and always prepares highly accurate pilot-response from all the users. Each user transmitter produces a transmit-data-frame using both a sequence which is allocated to the user and a carrier wave which is common for all the users, and the produced frame is transmitted.
That is to say that system (B) is such a system that a base-station BS accurately acquires the transmission-path-characteristics (channel) from each user to BS by means of that each user transmits isolated pilot-signals to BS. Consequently, BS can obtain the pilot-response (channel-response) {psk}(s: the number showing the shift position of the correlation-function for s=0,1,2, . . . L−1), corresponding to the transmission-path from the k(=0,1,2, . . . K−1)-th user. If a transmit-information of the k-th user is bk, the correlation response {Φs} of a received-frame r(t) is given by
{Φs}=bk{psk}+Δbk{psk}. (6)
The second term on the right-hand in the above equation is an error component corresponding to a white noise included in r(t). By solving Eq. (6), a detected output {circumflex over (b)}k=bk+Δbk of the transmitted information can be obtained. This system can remove the effect of interference waves almost completely. Because {psk} includes an interference component due to the white noise, perfect separation of interference waves can not be achieved, however, for CDMA systems to which interference waves give stronger disturbance than that white noise does, the advantage which almost all the interference wave components can be removed is valuable.
However, since system (B) uses the correlation responses, {psk} makes a matrix with a size of L×L, as a result, for the parameter k(=0,1,2, . . . K−1), a relation given by K≦L is required. That is to say that the number of users who can be accommodated is limited by the spreading sequence length, and it raises a problem that the frequency-utilization-efficiency can not be sufficiently increased.
System (C) is a demodulating detection method used for a multipath environment where many of delayed waves generated by a desired user's transmitted frame are contained in a received frame. This system has a function such that not only a main wave but also the delayed waves are demodulated by multiple fingers (demodulation circuits operating synchronously with delayed waves) and their outputs are summed up. It is reported that the error increases considerably due to the inter-station-interference-disturbance as the number of users increases, because the above stated finger output includes inter-station-interference-components consisting of main waves and the delayed waves both coming from interfering stations.
System (D) is a CDMA communication system using multiple transmit- and receive-antennae. In this system, respective user transmitters produce plural frame-sequences by applying convolutional coding to a transmit-data-frame-sequence, and then transmit these coded frame-sequences via plural antennae. A base-station receiver obtains demodulated outputs for respective users with a multi-user demodulator, using a signal group received via plural (Nd) antennae, and then produces the first detected value bn on the n-th frame. The time sequence of the first detected value bn has a high error-rate. Based on these plural (Nd) sets of similar time sequences, a convolutional decoder performs decoding to produce the second detected values. By feeding back the detected values to the above-stated multi-user-demodulator and the convolutional decoder, the above-stated demodulation and decoding are performed to obtain the third detected values. This is such a system that the final detected values are decided by repeating the above stated processing multiple times.
System (D) requires a considerably complicated processing and a long delay time. It is the problem of this system that nevertheless it fails to increase the frequency-utilization-efficiency. In contrast, an actual sequence increasing technique of the present invention to be stated later is a system capable of obtaining the above stated first detected value bn with a low error-rate, to acquire an excellent performance by using a simple processing, and it is a quite different technique from system (D).
The present invention has been done to overcome the performance limit in terms of the frequency-utilization-efficiency or power-bandwidth-product, as CDMA system evaluation measure, which the preceding systems have indicated, by providing multiary system design techniques with spreading sequence sets.
In order to solve the above-described problems, the invention of the present invention is characterized by the fact that, in a direct-sequence spread-spectrum CDMA communications system, a transmitter of the k-th user uk(k=0,1, . . . K−1) comprises such a function as to transmit information of I bits per symbol-frame by providing; means of that a sequence-set consisting of M kinds of basic spreading sequences g0, g1, . . . gm, . . . gM−1, each sequence having code length L, is prepared, an information-set consisting of γ components, {βS}=(β0, β1, . . . βγ−1) not including the value 0 is made of a modulated sequence-set which are made by multiplying the said M respective basic spreading sequences by an original information-set consisting of M components {β}=(β0, β1, . . . βM−1) including (M−γ) pieces of the value 0, resulting in production of a selected sequence-set consisting of γ selected spreading sequences, means of that a modulated synthesized sequence sD is produced by adding up all of the said selected spreading sequences, that a repeated core spreading sequence sDN with a sequence length of NL chips is produced by repeating the said respective selected sequences an integer (N) times, that an enveloped sequence eD is produced as an baseband transmitting data-frame by arranging the rear and the front parts of the said repeated core spreading sequence as the guard sequences at the front and the rear outsides of the said repeated core spreading sequence, and that said data-frame modulates a carrier-wave having the k-th orthogonal frequency fk which is orthogonal each other on k and the resultant output is transmitted, and means of that a similar enveloped sequence ep is produced based on a sequence pgp which is made by multiplying a basic spreading sequence gp for pilot by a pilot-information p as a baseband transmit-pilot-frame, and that the said pilot-frame is transmitted so as not to interfere with similar data-frames and pilot-frames of other users, and
a receiver comprises such a function as to detect the above-described transmitted information of I bits per symbol-frame by providing; means of that a received signal is demodulated into a baseband signal using a local carrier-wave {circumflex over (f)}0 of the 0-th orthogonal frequency f0 which the transmitter of a desired user u0 has used, that a synchronously received frame period with a length of NL chips is set at the position which is synchronized with the principal wave of signal components coming from the desired user included in this demodulated output, that the baseband signal component on this period is extracted as a repeated synchronously received data-frame r*DN, and a repeated synchronously received pilot-frame r*pN is also obtained in the same way, and that by applying an averaging operation to the said respective repeated frames, a synchronously received data-frame r*D and a synchronously received pilot-frame r*p both with reduced length L chips are produced, means of that the said synchronously received data-frame r*D is analyzed using a channel response {μ} which is the demodulated output of the said synchronously received pilot-frame, resulting in production of an estimate-set {β′}=(β′0, β′1, . . . β′M−1) corresponding to the said original information-set which has been transmitted by the desired station, and means of that a detected output-set {{circumflex over (β)}S}={{circumflex over (β)}0S, {circumflex over (β)}1S, . . . {circumflex over (β)}γ−1S} consisting of the information which the desired station has transmitted obtained by judging the said estimate-set.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, in a direct-sequence spread-spectrum CDMA communications system, a transmitter of the k-th user uk(k=0,1, . . . K−1) comprises such a function as to transmit information of I bits per symbol-frame by providing; means of that a sequence-set consisting of M kinds of basic spreading sequences g0, g1, . . . gm, . . . gM−1, each sequence having code length L, is prepared, an information-set consisting of γ components, {βS}=(β0, β1, . . . βγ−1) not including the value 0 is made of a modulated sequence-set which are made by multiplying the said M respective basic spreading sequences by an original information-set consisting of M components {β}=(β0, β1, βM−1) including (M−γ) pieces of the value 0, resulting in production of a selected sequence-set consisting of γ selected spreading sequences, means of that a modulated synthesized sequence sD is produced by adding up all of the said selected spreading sequences, that a repeated core spreading sequence sDN with a sequence length of NL chips is produced by repeating the said respective selected sequences an integer (N) times, that an enveloped sequence eD is produced as an baseband transmitting data-frame by arranging the rear and the front parts of the said repeated core spreading sequence as the guard sequences at the front and the rear outsides of the said repeated core spreading sequence, and that said data-frame modulates a carrier-wave having the k-th orthogonal frequency fk which is orthogonal each other on k and the resultant output is transmitted, and means of that a similar enveloped sequence ep is produced based on a sequence pgp which is made by multiplying a basic spreading sequence gp for pilot by a pilot-information p as a baseband transmit-pilot-frame, and that the said pilot-frame is transmitted so as not to interfere with similar data-frames and pilot-frames of other users, and a receiver comprises such a function as to detect the above-described transmitted information of I bits per symbol-frame by providing; means of that a received signal is demodulated into a baseband signal using a local carrier-wave {circumflex over (f)}0 of the 0-th orthogonal frequency f0 which the transmitter of a desired user u0 has used, that a synchronously received frame period with a length of NL chips is set at the position which is synchronized with the principal wave of signal components coming from the desired user included in this demodulated output, that the baseband signal component on this period is extracted as a repeated synchronously received data-frame r*DN, and a repeated synchronously received pilot-frame r*PN is also obtained in the same way, and that by applying an averaging operation to the said respective repeated frames, a synchronously received data-frame r*D and a synchronously received pilot-frame r*P both with reduced length L chips are produced, means of that M kinds of fabricated pilot-frames r*pm(m=0,1,2, . . . M′−1) each having a length of L chips are produced by using both the channel response (μ) which is the demodulated output of the said synchronously received pilot-frame and the above-described sequence-set consisting of M kinds of basic spreading sequences, resulting in production of a pilot matrix ρ with a size of L×M, and means of that a system of M linear equations in M unknowns is made of the said pilot-matrix ρ, a received data-matrix d′=(d′0, d′1, . . . d′L−1)T consisting of the chip-elements of the above-described synchronously received data-frame r*D, and an unknown matrix β′=(β′0 β′1 . . . β′M−1)T, that an estimate-set {β′}=(β′0, β′1 . . . β′M−1) corresponding to the original information-set is produced by solving the said system, and that a detected output-set {{circumflex over (β)}S}=({circumflex over (β)}0S, {circumflex over (β)}1S, . . . {circumflex over (β)}γ−1S) consisting of the information-set which the desired user has transmitted is obtained by judging the said estimate-set.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, in a direct-sequence spread-spectrum CDMA communications system, a transmitter of the k-th user uk(k=0,1, . . . K−1) comprises such a function as to transmit information of I bits per symbol-frame by providing; means of that a sequence-set consisting of M′ kinds of basic spreading sequences (g0k, g1k . . . gmk . . . gM′−1), each having a sequence length of L chips, is prepared, an information-set consisting of γ components, {βkS}=(β0kS, β1kS, . . . βγ−1kS) not including the value 0, is made of a modulated sequence-set which is made by multiplying the said M′ respective basic spreading sequences by an original information-set consisting of M′ components {βk}=(β0k, β1k, . . . βM′−1k) including (M′−γ) pieces of the value 0, resulting in production of a selected sequence-set consisting of γ selected spreading sequences {βkSgkS}=(β0kSg0kS, β1kS, g1kS . . . βγ−1kSgγ−1kS), means of that a modulated synthesized sequence sDk is produced by adding up all of the said selected spreading sequences, that a repeated core spreading sequence sDNk with a sequence length of NL chips is produced by repeating the said synthesized sequences an integer (N) times, that an enveloped sequence eDk is produced as an baseband transmitting data-frame by arranging the rear and the front parts of the said repeated core spreading sequence as the guard sequences at the front and the rear outsides of the (respective sequences) sequence, and that the said data-frame eDk modulates a carrier-wave having a frequency fD which is a common carrier for data, and the resultant output is transmitted, means of that a similar enveloped sequence epk is produced using a sequence pgp which is made by multiplying a basic spreading sequence gp for pilot by a pilot-information p as a baseband transmit-pilot-frame, and the said pilot-frame modulates a carrier-wave having an orthogonal frequency f′k, and the resultant output is transmitted, and means of that all of the frequencies fD and f′k are set so as to be different each other by integer times of a frequency fG corresponding to the reciprocal of the synchronously received frame-period with length NL chips, and a receiver comprises such a function as to detect the above-described transmitted information of I bits per symbol-frame by providing; means of that a received signal is demodulated into baseband signals using local carrier-waves {circumflex over (f)}′k and {circumflex over (f)}D respectively having orthogonal frequencies f′k and fD which the transmitter has used, that a synchronously received frame-period with a length of NL chips is set so as to satisfy a quasi-synchronous condition, that a repeated synchronously received data-frame r*DN and a repeated synchronously received pilot-frame rpNk* are produced by extracting the baseband signal components on this period, and that a synchronously received data-frame r*D and a synchronously received pilot-frame rpk* both having a reduced length of L chips are produced by applying an averaging operation to the said respective repeated frames, means of that M′ kinds of fabricated pilot-frames rpmk*(m=0,1,2, . . . M′−1) each having a length of L chips are produced by using both the channel response (μ) which is the demodulated output of the said synchronously received pilot-frame and the above-described sequence-set consisting of M′ kinds of basic spreading sequences, resulting in production of a pilot-matrix ρ′ with a size of L×M using M=KM′ pieces of the fabricated pilot-frames, and means of that a system of M linear equations in M unknowns is made of the said pilot-matrix ρ′, a received data-matrix d′=(d′0, d′1, . . . d′L−1)T consisting of the chip-elements of the above-described synchronously received data-frame r*D, and an unknown matrix βk′=(β00′β10′ . . . βM′−10′, β01′β11′ . . . βM′−11′, β0K−1′β1K−1′ . . . βM′−1K−1′)T, that an estimate-set {βk′}=(β0k′β1k′ . . . βM′−1k′) corresponding to the original information-set which the k-th user has transmitted is produced by solving the said system, and that a detected output-set {{circumflex over (β)}kS}=({circumflex over (β)}0kS, {circumflex over (β)}1kS, . . . {circumflex over (β)}γ−1kS) consisting of the above-described transmitted information for the k-th user is obtained by judging the said estimate-set.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a transmitter of the k-th user uk(k=0,1, . . . K−1) comprises; means of that a transmitting pilot-frame ep which is an enveloped sequence produced based on a basic spreading sequence gp for pilot described above is transmitted over a carrier-wave with an orthogonal frequency f′k, that the said orthogonal frequencies f′k and fk(k=0,1,2, . . . K−1) used for transmitting the above-described pilot-frames and data-frames respectively are set so as to be different on k each other by integer times of a frequency fG which is the reciprocal of the repeated synchronously received frame period TG, resulting in providing of the frequency-division-pilot-frame-transmission-function by setting the sequence length of the above-described guard sequences so that any boundaries of the respective received flock-flames rpfk and rDfk coming from all the users may not be included in the synchronously received frame period, and a receiver comprises; means of that the said synchronously received pilot-frame is demodulated using a local frequency {circumflex over (f)}0 corresponding to the orthogonal frequency f′0 which the transmitter of desired station u0 has used, and means of that an isolated-pilot-response for a desired station without being subjected to interference due to both the above-described data-frame received from a desired station and all the interference-waves received from interfering stations is produced.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a transmitter of the k-th user uk comprises; means of that a synthesized frame-sequence is produced by providing empty slots in the above-described baseband transmit-data-frame sequence, and inserting pilot-frames in said data-frame sequence in time-division manner, and means of that the said synthesized frame-sequence is transmitted over the above-described k-th orthogonal carrier wave having frequency fk, and a receiver comprises; means of that the parts corresponding to the pilot-frames in a frame-sequence made of the above-described synchronously received frames are extracted in a time-division manner and then the extracted parts is demodulated, and an isolated-pilot-response without being subjected to interference due to the above-described interference-waves is produced.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a transmitter of the k-th user uk(k=0,1 . . . K−1) comprises; means of that a transmit-pilot-frame for uk which is produced by modulating a common carrier-wave for pilot having an orthogonal frequency fp by the k-th baseband transmitting pilot-frame described above epk is transmitted over such a transmission timing as to receive in time-division-manner with a similarly received pilot-frame which the k′-th user uk′(k′≠k) has transmitted, and a receiver comprises; means of that a baseband pilot-frame for uk is extracted in time-division-manner from a pilot-frame-sequence demodulated by a carrier-wave having the said orthogonal frequency fp, and the above-described channel-response is produced using the resultant demodulated output.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a receiver detects the transmitted information of I bits per symbol-frame by comprising; means of that the 0-shift-correlations p′nm between the m-th fabricated pilot-frame r*nm and M data analyzing sequences yn(n=0,1,2, . . . M−1) each of which differs each other are produced, and that a pilot-response-matrix P′ with a size of M×M constituted of the said correlations is produced, means of that a data-response-matrix Φ′=(Φ′0, Φ′1, Φ′2, . . . Φ′M−1)T is produced by such a method as to obtain the 0-shift correlation Φ′n between the above-described synchronously received data-frame r*n and the n-th data analyzing sequence yn, and means of that a system of M linear equations in M unknowns are made of the pilot-response-matrix P′, the data-response-matrix Φ′ and an unknown matrix β′=(β′0, β′1, β′1, . . . β′M−1)T, that an estimated set β′=(β′0, β′1, β′1, . . . β′M−1) corresponding to the said original information-set is obtained by solving the said system, and that a detected output-set {{circumflex over (β)}S}={{circumflex over (β)}0S, {circumflex over (β)}1S, . . . {circumflex over (β)}γ−1S} consisting of the above-described transmitted information is obtained by judging the said estimate-set.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a receiver detects the transmitted information of I bits per symbol-frame which is transmitted by each user by comprising; means of that M′ fabricated pilot-frames rpmk* and M′ data analyzing sequences ynk(n=0,1,2, . . . M′−1) corresponding to the k-th user uk(k=0,1,2, . . . K−1) are produced, means of that a pilot-response-matrix P′ with a size of (M×M) constituted of the 0-shift-correlations pnmk* between the said respective fabricated pilot-frames rpmk* and the said respective data analyzing sequences ynk is produced, and means of that a data-response-matrix Φ′=(Φ00′, Φ10′, . . . ΦM′−10′, Φ01, Φ11, . . . ΦM′−11, . . . Φ0K−1, Φ1K−1, . . . ΦM′−1K−1)T is produced using such a method as to obtain the 0-shift-correlation Φnk′ between the above-described synchronously received data-frame r*D and the said data analyzing sequence ynk, that a system of M linear equations in M unknowns is made of the said pilot-response-matrix P′, the said data-response-matrix Φ′ and an unknown matrix β′=(β00′, β10′, . . . βM′−10′, β01′, β11′, . . . βM′−11′, β0K−1, β1K−1, . . . βM′−1K−1)T, that a estimate-set {βk′}=(β0k′, β1k′, . . . βM′−1k′ corresponding to the transmitted information-set consisting of M′ components for each user is obtained by solving said system, and that a detected output-set {{circumflex over (β)}kS}={{circumflex over (β)}0kS, {circumflex over (β)}1kS, . . . {circumflex over (β)}γ−1kS} corresponding to the transmitted information of each user is obtained by judging the estimate-set.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a transmitter of the k-th user uk comprises; means of that a sequence-set consisting of M kinds of basic spreading sequences having L chips is prepared, being transmitted, and a receiver detects the transmitted information of I bits per symbol-frame which is transmitted by each user by comprising; means of that double length fabricated pilot-frames r*pmE and a double length synchronously received data-frame r*DE both having 2L chips are produced using the real part and the imaginary part of the respective chip-elements of both the fabricated pilot-frames r*Pm(m=0,1,2, . . . M−1) and the synchronously received data-frame r*D, that a system of M linear equations in M unknowns is made of a double length received pilot-matrix ρ′ with a size of 2L×M made of the double length fabricated pilot-frames r*PmE, a received data-matrix d′ which consists of the chip-elements of the double length synchronously received data-frame r*DE, and an unknown matrix β′, that an estimate-set {β′} corresponding to the original information-set is produced by solving the said system, and that a detected output-set {{circumflex over (β)}S} consisting of the information which the desired station has transmitted is obtained by judging the said estimate-set.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a transmitter of the k-th user uk comprises; means of that a sequence-set consisting of M′ kinds of basic spreading sequences having L chips is prepared, being transmitted, and a receiver detects the transmitted information of I bits per symbol-frame which is transmitted by each user by comprising; means of that double length fabricated pilot-frames rpmEk* and a double length synchronously received data-frame r*DE both having 2L chips are produced using the real part and the imaginary part of the respective chip-elements of the fabricated pilot-frames rpmk*(m=0,1,2, . . . M−1, k=0,1,2, . . . K−1) and the synchronously received data-frame r*D, that a system of M linear equations in M unknowns is made of a double length received pilot-matrix ρ′ with a size of 2L×M made of the double length fabricated pilot-frame rpmEk*, a received data matrix d′ which consists of the chip-elements of the double length synchronously received data-frame, and an unknown matrix βk′, that an estimate-set {βk′} corresponding to the original information-set is produced by solving the said system, and that a detected output-set {{circumflex over (β)}kS} consisting of the information which the desired station has transmitted is obtained by judging the said estimate-set.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a transmitter comprises; means of that transmit-signals are transmitted via one or plural transmit-diversity-antennae ATv(v=0,1,2, . . . ), and a receiver comprises; means of that received signals rd are received via plural receive-diversity-antennae ARD(d=0,1,2, . . . ), that baseband demodulated outputs are produced at pilot and data demodulating parts Dpd and DDd for the d-th received signal rd by using a local carrier-wave corresponding to a carrier-wave which the transmitter of a desired station has used, and thereby the d-th synchronously received pilot- and data-frames r*pd and r*Dd are produced, that the d-th channel response {μ}d is obtained with the d-th pilot-frame r*pd, that a fabricated pilot-frame r*pmd or r*pmd corresponding to the d-th received signal is obtained using the said channel response {μ}d and the m-th basic spreading sequence which the desired station has used, that extended fabricated pilot-frames r*pmE or rpmEk* are produced by using such a method as to increase the frame length by arranging chip-elements of the said fabricated pilot-frames which belong to respective antenna numbers d on a time axis, that an extended synchronously received data-frame r*DE is produced by arranging the said d-th synchronously received data-frame r*Dd on a time axis in the same way, that an estimate-set {β′}=(β′0, β′1, . . . β′M−1) corresponding to the original information-set is produced by analyzing the said extended synchronously received data-frame r*DE with the said extended fabricated pilot-frames r*PmE or rPmEk*, and that a detected output-set {{circumflex over (β)}S}={{circumflex over (β)}0S, {circumflex over (β)}1S, . . . {circumflex over (β)}γ−1S}consisting of the information which the desired station has transmitted is obtained by judging the said estimate-set.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a transmitter comprises; means of that a sequence-set consisting of Nv basic spreading sequences {gm}=(gm0, gm1, . . . gmv, . . . gm,Nv−1) per one transmit-information βm is prepared, that the v-th basic spreading sequence gmv is modulated by the same information βm to produce a core-sequence for data and then an enveloped sequence for data based on this modulated output, that an enveloped sequence for pilot is produced in the same way, that both the said enveloped sequences are lead to the v-th transmit-diversity-antenna ATv to transmit, and a receiver comprises; means of that the components which the v-th transmit-diversity-antenna ATv has transmitted are demodulated using an input received via the d-th received-diversity-antenna ARd, resulting in production of both the d-th synchronously received data-frames r*Dvd and the d-th synchronously received pilot-frames r*pvd, that a channel response {μ}vd from the v-th transmit-antenna to the d-th received-antenna is produced using the said synchronously received pilot-frame, that the d-th fabricated pilot-frames r*pmvd or rpmvdk* are produced using the channel response, that an extended fabricated pilot-frame r*pmE or rpmEk* is produced by such a method as to increase the frame length by arranging chip-elements of fabricated pilot-frames which belong to the respective pairs of transmit- and receive-antenna numbers v and d (mutually different)* on a time axis, that an extended received frame r*DE is produced by arranging the said synchronously received frame r*Dvd in the same way, and that the said extended synchronously received frame r*DE is analyzed with the said extended fabricated pilot-frames r*pmE or rpmEk*.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, such a function is added that, when the chip-elements of the above-described fabricated pilot-frames and synchronously received data-frames which belong to respective transmit- and receive-antenna original numbers are arranged on a time axis, and thereby extended fabricated pilot-frames and extended synchronously received data-frames are produced, a real amplitude component and an imaginary amplitude component of a chip-element are arranged so as to take individual time positions to constitute each frame.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a receiver comprises such a function as to detect the transmitted information of I bits per symbol-frame which is transmitted by each user by providing; means of that instead of solving a system of M linear equations in M unknowns, an optional number MR which is smaller than M, and equal to γ or larger than γ is chosen to produce a reduced size system of MR linear equations in MR unknowns, and NR sets of the reduced size systems are produced, that NR estimate-sets of the spreading sequence-set {β′}=(β′0, β′1, . . . β′MR−1) or {βk′}=(β0k′, β1k′, . . . βMR−1k′) are obtained as solutions of the said respective equations, and that an evaluation function Δβ which shows deviations between the said respective estimate-sets and the original information-set {β} or {βk} is obtained, and means of that an estimate-set {β′} or {βk′} corresponding to the original information-set based on a set with which Δβ takes the minimum is obtained, and that a detected information-set {{circumflex over (β)}S} or {{circumflex over (β)}kS} consisting of the above-described transmitted information is obtained by judging the said estimate-set.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a receiver comprises such a function as to detect the transmitted information of I bits per symbol-frame which is transmitted by each user by providing; means of that instead of solving a system of M linear equations in M unknowns, combination-sets Us(s=0,1, . . . Nγ−1) which consists of γ frames taken out of the fabricated pilot-frames rpm(m−0,1,2, . . . M−1) is made, that a sum of the frames which is made by multiplying the elements of plural virtual sequences Ch(c0, c1, . . . cγ−1) by the respective frames of a said combination-set Us is produced as a fabricated transmitted frame Fsh, and that correlations between respective fabricated transmitted frames and the above-described synchronously received data-frame r*D are obtained, and means of that the maximum likelihood fabricated transmitted frame is obtained based on the said correlations, and that a detected information-set {{circumflex over (β)}S} consisting of the above-described transmitted information is obtained by judging the said estimate-set {β′} using the constituent elements Us and Ch of the maximum likelihood fabricated transmitted frame.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a receiver comprises such a function as to detect the transmitted information of I bits per symbol-frame which is transmitted by the k-th user uk by providing; means of that instead of solving a system of M linear equations in M unknowns, combination-sets USk(s=0,1,2, . . . Nγ−1) which consists of Kγ frames taken out of the fabricated pilot-frames rPmk*(m=0,1,2, . . . M′−1) for the k-th user uk(k=0,1,2, . . . K) is made, that a sum of the frames which is made by multiplying the elements of plural virtual sequences Ch(c0, c1, . . . cKγ−1) by the respective frames of a said set Usk is produced as a fabricated transmitted frame Fshk, and that correlations between respective fabricated transmitted frames Fshk and the above-described synchronously received data-frame r*D is obtained, and means of that the maximum likelihood fabricated transmitted frame is obtained based on the said correlations, and that a detected information-set consisting of the above-described transmitted information is obtained by judging an estimate-set corresponding to the original information-set using the constituent elements Usk and Ch of the maximum likelihood fabricated transmitted frame.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a receiver comprises such a function as to detect the above-described transmitted information of I bits per symbol-frame by providing; means of that the above-described fabricated pilot-frames r*pm or rpmk*(m=0,1,2, . . . M−1, k=0,1,2, . . . K−1) are generated, and such analyzing sequences wm that the 0-shift-correlations with fabricated pilot-frames r*pm and r*pm′(m′≠m) may take 1 and 0, respectively, are produced, means of that plural estimate-sets set {β′}=(β′0, β′1, . . . β′M−1) corresponding to the original information-set are produced based on the 0-shift correlations between the above-described synchronously received data-frame r*D and analyzing sequences wm, that by judging the maximum likelihood set taken out of these estimate-sets a detected information-set {{circumflex over (β)}S}={{circumflex over (β)}0S, {circumflex over (β)}1S, . . . {circumflex over (β)}γ−1S} consisting of the above-described transmitted information is obtained, or means of that plural analyzing sequences wmk such that 0-shift-correlations with fabricated pilot-frames rpmk* and rpm′K′*(m′≠m, k′≠k; m′=m, k′≠k; m′≠m, k′=k) may take 1 and 0, respectively, are produced, means of that plural estimate-sets {βk′}=(β0k′, β1k′, . . . βM−1k′) corresponding to the original information-set are produced based on the 0-shift-correlations between the above-described synchronously received data-frame r*D and the analyzing sequences wmk, and means of that by judging the maximum likelihood set taken out of these sets a detected information-set {{circumflex over (β)}kS} consisting of the above-described transmitted information is obtained.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a transmitter comprises; means of that a binary to multi-level conversion is carried out by using such a method as to correspond a transmit-information of I bits to a case-number obtained by synthesizing the first selection where an optional number γ is selected out of the numbers of spreading sequences, ranging from one to γm, and the second selection where γ spreading sequences are selected out of an above-described sequence-set consisting of M or M′ kinds of basic spreading sequences as the sequence kinds, means of that a selected sequence-set consisting of γ selected spreading sequences is produced from the above-described original information-set {β} or {βk} to produce a transmitting frame, and a receiver comprises such a function as to detect the above-described transmitted information of I bits per symbol-frame by providing; means of that fabricated pilot-frames are obtained based on the above-described channel response {μ}, and an estimate-set {β′} or {βk′} corresponding to the original information-set is obtained by analyzing the above-described synchronously received data-frame r*D with them, means of that while a detected value {circumflex over (γ)} for the transmitted spreading sequence number is obtained using the said estimate-set, and means of that a detected information-set {{circumflex over (β)}S} or {{circumflex over (β)}kS} consisting of {circumflex over (γ)} selected information components is obtained per user.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a receiver comprises; means of that in solving an above-described system of M linear equations in M unknowns, it is assumed that frames produced by multiplying dummy-information-components {βD}=(βMD, βM+1D, . . . βM+h−1D) by h(≧1) dummy spreading sequences which are not included in the basic spreading sequence-set provided in a transmitter may be included in the above-described synchronously received data-frame, means of that M pieces of the said fabricated pilot-frames r*pm(m=0,1,2, . . . M−1) and the (M+h) pieces of the said data analyzing sequences yn(n=0,1,2, . . . M−h−1) are produced and an extended pilot-response-matrix P′+ is produced by adding h additional row vectors A to M row vectors each having (M+h) elements which are produced from the 0-shift-correlations between the above-described fabricated pilot-frames r*pm and the said data analyzing sequences yn, means of that a corrected data-response-matrix Φ″+=(Φ″0, Φ″1, . . . Φ″M+h−1)T is produced using a method of adding correcting terms to the 0-shift-correlations between the above-described synchronously received data-frame r*D and the data analyzing sequences yn, means of that a system of (M+h) linear equations in (M+h) unknowns which is made of the said enlarged pilot-response matrix P′+, the corrected data-response-matrix Φ″+ and an extended unknown matrix β′+=(β′, βD)T=(β′0, β′1, . . . β′M−1, βMD, βM+1D, . . . βM+h−1D)T is produced, and means of that after the said additional row vector A is selected so as to increase the determinant or normalized singular value of the said enlarged pilot-response-matrix P′+ to enhance the regularity, an estimated set {β′} or {βk′} corresponding to the original information-set is obtained by solving the said system of (M+h) linear equations in (M+h) unknowns.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a receiver comprises; means of that in solving an above-described system of M linear equations in M unknowns, an above-described enlarged pilot-matrix P′+ is produced so as to enhance the regularity of the said enlarged matrix using the above-described fabricated pilot-frames r*pm, h additional row vectors A, and h additional column vectors C which are random sequences, means of that an extended unknown matrix β′+ is made by adding h additional dummy-information {βD} to an unknown set {β′} consisting of M unknowns, means of that a corrected synchronously received data-frame r*D+ is made by adding correcting terms to the above-described synchronously received data-frame r*D, means of that a solution set {β′+} is obtained by solving a system of (M+h) linear equations in (M+h) unknowns which is composed of the said enlarged pilot-matrix ρ′+, the extended unknown matrix β′+, and the corrected received data-matrix d′+ consisting of the chip-elements of the corrected synchronously received data-frame r*D+, and plural solution sets {β*+} are resultantly obtained corresponding to the said additional column vectors C by solving the said systems, and means of that the maximum likelihood solution taken out of the solution sets is set to an estimate-set {β′} corresponding to the original information-set.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, an identical sequence-set consisting of M kinds of basic spreading sequence allocated to some cells, and a scrambling sequence inherent to each cell is allocated to a cell, and a transmitter of the cell comprises; means of that a scrambled sequence-set is produced by multiplying respective sequences of the spreading sequence-set by the scrambling sequence inherent to the cell to produce a transmit-frame, and a receiver comprises; means of that pilot and data analyzing sequences are produced based on the above-described scrambled sequence-set to analyze the above-described synchronously received data-frame.
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a transmitter of the k-th user uk comprises; means of that a sequence-set consisting of M=M′=N kinds of basic spreading sequences (g0, g1, g2, . . . gM−1) is produced by multiplying code words (h0, h1, . . . hN−1) consisting of rows of an hadmarl matrix with a size of N×N by a random sequence-set consisting of M′ kinds of spreading sequences (g00, g10, . . . gM′−10).
The invention of a CDMA communications system using multiple spreading sequences is characterized in that, a transmitter comprises; means of that a sequence-set consisting of M=M′L kinds of cyclically shifted sequences is prepared as a basic spreading sequence-set by producing the s=(0,1,2, . . . L−1)-shift-sequences of respective spreading sequences which constitutes a sequence-set consisting of M′ kinds of spreading sequences (g0, g1, . . . gM′) with L chips in sequence length, means of that transmit-information of I bits is converted into an information-set consisting of γ components {βS} selected out of the basic spreading sequence-set, and means of that a transmit-frame is produced and transmitted by using a selected sequence-set {βSgS} corresponding to the said information-set, and a receiver detects the transmitted information of I bits per symbol-frame by producing a detected output-set {{circumflex over (β)}S}={{circumflex over (β)}0S, {circumflex over (β)}1S, . . . {circumflex over (β)}γ−1S} consisting of the information which the desired station has transmitted.
This invention is related to a system, which can overcome the above-described problem inherent to CDMA communication systems, that is, a characteristic vulnerable to the disturbance caused by interference-waves incoming from the other users, and also can enhance the frequency-utilization-efficiency. Let us explain embodiment-examples applied to technically difficult up-link transmission systems. In the invention, the transmitter uk of the k-th (=0, 1, . . . K−1) user transmits an isolated pilot-frame, and the base-station receiver BS has a function eliminating the disturbance component caused by interference-waves in data-frames, similarly to the above-described method (B) by utilizing this pilot-frame or its correlation-response as well as by utilizing the self-interference-waves. Moreover, the transmitter converts the binary data-information-set {b} to be transmitted to multi-value-information-set {β} in advance. In other words, by multiplying this multi-value-information {β} to a multiple sequence-set, a binary data-set may be related to a multi-value. Namely, TX selects γ(≦M) spreading sequences from M kinds of sequences by multiplying M kinds of codes (sequence population) by an information-set {β}=(β0, β1, . . . βM−1) including M−γ pieces of the value 0. Then, γ components composing a binary information-set, not including the value 0, {βs}={β0s, β1s, . . . βγ−1s} labeled by both new ordinal numbers and a superscript S are multiplied by these selected spreading sequences for transmission. The receiver detects the kind and the polarity (±1) of the respective selected sequences. From this detected output, the above-described binary information-set {{circumflex over (b)}} is restored. This is a sort of multi-ary systems which can transmit the information of multiple bits I per symbol frame, and can enhance the frequency-utilization-efficiency.
where x(t) is the Additive White Gaussian Noise (AGWN).
b) shows the down-link transmission-paths. In this transmission, the delayed waves due to multi-paths shown by dotted lines are similarly generated. The received waves received by user station uo contain not only an illustrated transmission wave sD(uo) and its delayed wave, but also transmission waves to the other stations uk(k≠0) and their delayed waves.
When there is an object intercepting the direct wave, demodulation may be performed using one of the delayed waves, instead of using the direct wave, as a principal wave. In this case, some of the interference-waves due to multi-paths precede the principal wave to be demodulated. The following description deals with the system design specifically for the technically more difficult up-link transmission by assuming states where none of preceding waves may take place, for convenience (without loss of generality).
{β}={β0, β1, . . . βm, . . . βM−1}[βmε0,±1]. (8)
The γ components of βm in M components of βm take a value ±1 while the remaining (M−γ) components of βm take the value 0. (Generally, βm, can take one of multi-values including 0, real values, complex values, etc. However, for simplicity, it is assumed to take ternary values as stated above.) On the other hand, pilot-information is fixed normally as p=1. For user uk, a sequence-population consisting of M kinds of mutually different basic spreading sequences gm(i) is provided. gm(i) is a discrete sequence with length L, as represented by
where, gmi is the i-th chip amplitude, which is assumed here to be a binary value (±1). Generally, it may be given by either a multi-value, a real value, or a complex value. (In this invention, any random sequences can be used.) δ and Tc stand for a delta function and a chip period, respectively. A pilot-sequence gP(i) is a similar basic spreading sequence with length L. At respective modulators MOD0, MOD1, . . . MODM−1, this basic spreading sequence gm(i) is multiplied by said ternary information βm, generating a symbol-frame consisting of L components each with a binary chip impulse (hereinafter, an expression omitting the symbol will be used). By synthesizing these outputs in the adderΣ, the multi-value synthesized impulse-sequence sD(i) consisting of L chips is produced. This output is repeated N-times at an illustrated repeating circuit REPD which generates a repeated core spreading sequence SDN(i) with sequence length LG=NL. Moreover, by copying both the rear portion l(i) and the front portion h (i) of SDN(i) and by adding them to the front and rear sides of SDN(i), respectively, an enveloped frame eD(i) with sequence length LE is produced. The respective chip impulses of eD(i) are applied to a convolutional multiplier COVD, where a convolution output with chip-waveform q(t) in Eq. (3) is obtained. This enveloped frame output eD(t) is a baseband transmit data-frame having a continuous time waveform with frame-period TE=LETc as represented by
where gm(t)/TE is the convolution-output made of both the periodical sequence of gm (i) with sequence-length LE and q (t).
On the other hand, at modulator MODp, pilot discrete sequence sp(i) is produced, by using pilot-information p and basic spreading sequence gp(i)[i=0,1,2, . . . L−1.]. Moreover, using repeating circuit REPp, enveloped circuit ENVp, and convolutional multiplier COVp, a baseband transmit-pilot-frame is obtained in the same way as stated above, and it is represented by
ep(t)=p{gp(t)/TR}. (11)
At illustrated modulators MODfp and MODfD, (enveloped) data-frame eD(t) modulates a carrier-wave fk while the (enveloped) pilot-frame ep(t) modulates a carrier-wave f′k. Synthesizing of these modulations outputs sDαk(t) and spαk(t) produces a transmit-frame-output sαk(t) in the radio-band.
A transmit-wave of respective user stations generates J delayed waves due to multi-paths, including the principal wave (the direct wave or one of the received waves with the maximum power). Therefore, an equalizing circuit output rpf0(t) may contain the baseband transmit-wave ep(t) and its delayed waves. DD0 is a data demodulating circuit similar to Dp0. Consider a low pass filter output generated at Dp0 and a similar output of DD0. These baseband received waves are given by the following equations, assuming that p=1
where the first terms of both the equations are flock-frames, each consisting of a sum of direct and delayed waves generated by the transmission pilot- or data-frame of user u0. μj0 and {circumflex over (μ)}j0 are the signal amplitudes of the j-th delayed wave of the 0-th user and each amplitude takes a complex value consisting of I and Q components due to the phase difference between the transmitted and the local carrier-waves. x(t) is AWGN. ra(t) contains the baseband components transmitted from the other user uk(k≠0) on both carrier-waves f′k and fk, and the pilot- and data-frame components transmitted from u0 on carrier-waves f′0 and f0, respectively. The third and fourth terms of above equations are outputs of inter-station and self-interference components which are obtained by demodulating them with the different orthogonal carrier-waves f0 and f′0, as described later. These outputs are made zero due to the above-described demodulation with the orthogonal carrier-waves and an averaging operation performed by an averaging circuit AO in
Note also that em(t) in Eq.(13) is an output of em(i) convoluted with q (t), and em(i) is an enveloped sequence with sequence length LE which is made by adding a header h (i) and tail l(i) to the core-sequence consisting of N-times repetition of gm(i).
Although both rpf0(t) and rDf0(t) are frame-waveforms on a continuous time axis, the following demodulation processing is performed after converting them into waveforms taking values in chip spacing Tc on a discrete time axis. For this purpose correlation outputs between these continuous waveforms and chip-waveform q(t) are produced at an illustrated chip-waveform correlator Cor (q) in every Tc. [The sampling function waveform is here assumed as q(t). Its time expanse is the integrating time of the correlator, and this time is selected to be sufficiently longer than chip period Tc in order to reduce the correlation error.] These correlated outputs become discrete sequences r0pf(i) and rODf(i), each consisting of chip elements arranged in chip spacing. (The function of this correlator is equivalent to that of a transversal filter with the weighting coefficient of the f-type sampling function.)
The time position of synchronously received frame-period TG(=LGTC) synchronized to the principal wave of the received signal for the desired station is specified by an illustrated frame pulse eF0. [See
Pilot-frame rp0*(i) generated in this manner is applied to a matched filter MF(z) matched to a pilot analyzing sequence z(i), generating a cross-correlation-function {μ0} corresponding to desired station uo, as the channel-response (equivalent to TLo in
Using channel-response {μ0}, synchronizing circuit SYN generates the frame pulse eF0 which is used to indicate the position of the synchronously received frame by known means and transmits it to circuits which need the frame pulse. With the aid of this timing, the core-sequence time portion of the principal wave is extracted from ra(t), as described above, to produce baseband outputs rpN0*(i) and rDN0*(i). [This system using accurate pilot-responses, do not need to establish carrier-wave-phase-synchronization.]
A synchronously received data-frame rD0*(i) with length L is applied, together with a set of r*pm(i), denoted by {r*pm}, to β analyzing circuit β-AYZ. (Circuits SQEXP and SQEXD indicated by dotted line frames in the figure will be described later.)
Circuit β-AYZ determines an estimate-set {β′} (β′0, β′1, β′2, . . . β′M−1) consisting of the M components of the information the desired station transmitted, using both data-frame rD0*(i) and pilot-matrix ρ′. ρ′ consists of frame set {r*pm} with a method described later. Moreover, ordinal numbering is newly applied to such γ components of set {β′}, that each may take a value closer to ±1 than those of the other components, generating a set {{circumflex over (β)}s}=({circumflex over (β)}0s, {circumflex over (β)}1s, . . . {circumflex over (β)}γ−1s). The figure shows a case with γ=2. By applying {{circumflex over (β)}s} to a multi-value to binary value conversion circuit p-s, a detection output {{circumflex over (b)}}/u0 corresponding to an information-set {b}/u0 transmitted from u0, is generated. In this transmission system assuming (βmε±1,0), the information quantity I per transmission-frame eD(i) is given as follows
I=γ+log2(MCγ). (14-A)
In the above-described system (A), γ has been fixed. Consider a system with γ not fixed. Let value γ be a natural number less then γmax as the first step, and let γ be components selected from M kinds of sequences as the second step. Consequently, the information quantity of the system is given as follows
(To allow γ to have the degree of freedom of γmax or less, highly-accurate interference elimination technology such as this invention is required.) The description here deals mainly with the case where γ is fixed. If multi-values or complex values were to be used, instead of ±1 for βm, I would become larger than that obtained from the above equation.
In TX (uo), a basic spreading sequence gp(i)=(gp0, gp1, . . . gpL−1) for a pilot with length L (chips) and selected spreading sequences g0s(i) and g1s(i) for data are provided. The latter two sequences are γ (≦M, γ=2 in this case) sequences selected from M kinds of basic spreading sequences gm(i)=(m=0.1.2. . . . . M−1) as described above, with the new ordinal numbers and superscript S. By multiplying these spreading sequences by the pilot information p(=1) and data information β0s and β1s (±1), illustrated sequences pgp(i), β0sg0s(i) and β1sg1s(i) are generated. The former is a basic pilot-frame sp(i). A synthesized sequence with sequence length L, made by summing respective chips of the latter two sequences, is a basic data-frame sD(i).
Similar sequences illustrated as pgp(i), β0sg0s(i) and β1sg1s(i) are made in TX (u1). gp(i) in this case is the same as the basic spreading sequence for u0, but latter two sequences of g0s(i), g1s(i) are normally different from data sequences for u0. This is because these data sequences correspond to transmission binary information-set {b}/u0 and {b}/u1 of u0 and u1, respectively. In TX (uo), based on the basic pilot-frame sp(i) common to users, a core-sequence spN(i)=sp(i)×4 is produced by repeating N times sp(i) (N=4 is set in this case) as shown in the figure. By adding a header sequence hp(i) and a tail sequence lp(i), as guard sequences, to the front and the rear sides of the core-sequence, an enveloped pilot-frame ep(i) is produced. Similarly, an enveloped data-frame eD(i) is produced based on sD(i). By convolutionally modulating said chip-waveform q(t) with these enveloped frame impulse-sequences, baseband enveloped frame waveforms ep(t) and eD(t) are produced, respectively. By modulating carrier-waves f′0 and f0 with these enveloped frames, respectively, and by synthesizing these modulated outputs (addition of respective chip-waveforms on the same time axis), a radio-band transmission-frame sa0(t) of u0 is produced.
In the similar manner, TX (u1) produces sp1(t) and sD1(t). Modulating different orthogonal carrier-waves f′1 and f1 with these frames produces a radio-band transmission-frame sa1(t) of u1.
The sequence length LE (chips) and time width TE of the above enveloped frames are given by the following equations
TE=LETC (16)
where, Lh, LG, and Ll are the sequence lengths of the header, core, and tail, respectively, while N is the number of the basic sequence repetition, and also the number of carrier-waves, while Tc is the chip period. In the above example, N=2K=4 has been set for the user population K(=2).
where TD and TG are the basic sequence-period and repeated sequence-period respectively. The comb-teeth shaped spectrum of
The spectrum of core-sequence sDN(i) [=sD(i)×4] occupies the comb-tooth shaped frequency-slot similarly. Assuming that f00 is an arbitrary reference-frequency, the following relationship may be applied to said carrier-waves used by uk
In this case, let us relate the time waveform and the spectrum as follows
Thus such a synthesized spectrum is obtained that the spectra of the respective frames may be not mutually overlap, as shown in
Discretely distributed multiplexing on the frequency axis by using carrier-waves with orthogonal frequencies having such relationship as expressed in Eq. (18) is thus provided. As a consequence, the frame components modulated with many different frequencies which are contained in a synchronously received frame with period TG become mutually orthogonal. Therefore, the receiver can perform correlative demodulation so as to separate these frame components.
b) shows the flock-frames rpfk(t) received in correspondence with transmitted pilot-signals spk(t) in
In
A time difference τs shown in
As τs and τI are determined by the cell radius and topography, Th and Tl can be designed to meet the above equation. [Note that r00(t) is used as a time reference. Therefore, negative-polarity τs and τI may occur, if there are any interfering waves coming in advance to r00(t) or self-preceding waves r00(t+Tc), because in the time difference calculation, r00(t) is used here as a reference. The time width Tl of the tail plays a role of guard sequence against such negative-polarity time difference.]
b) shows a case in which a quasi-synchronization-condition as above-described is satisfied. In this case, both interference components, such as μ10r00(t−Tc) and μ21r01(t−2Tc) contained in the synchronously received frame rp00*(t), are sequences made by multiplying cyclically shifted sequences of the core-sequence-portions (period TG) of s00(t) or s01(t) by the complex amplitude. If Eq. (19) is not satisfied, the slant portion in rp00*(t) generally results in containing the tail or header of an preceding (n=−1) or succeeding (n=1) frame carrying different information [β0s, etc.] from that of the desired station or the interference station. In consequence, interference components fail to become simple N-time repeated sequences of a certain sequence, so that the orthogonal relationship on the frequency axis explained in
where,
b) shows enveloped data-frame eD(i) shown in
where,
[This averaging operation can be included in the solution processing of Eq. (29) or the correlation processing of Eq. (25) to be described later. Note also that the component shown with the third and fourth terms in, Eqs. (12) and (13) becomes zero through this averaging operation.]
Comparing Eqs. (23) and (24), and noticing that μj and {circumflex over (μ)}j (j=0,1,2) can be converted mutually, and βms=±1 hold good for p=1, one can find that the value of βms may be obtained by analyzing r*D(i) with r*p(i). The specific approach is described.
The matched filter MF(z) in
where gp(i) is a basic spreading sequence for pilot, while − is a symbol denoting the complex conjugate. A cyclic cross-correlation-function {μ} between this pilot analyzing sequence z(i) and synchronously received pilot-frame r*p(i) which omits the subscript of user number k is a channel-response representing the transmission-path characteristics, and it is given by the following equation, assuming AWGN is zero (x(i)=0)
where μj is the received frame amplitude delayed by jTc from the direct wave, as shown in
where [ ] is Gauss symbol. Thus, the channel-response {uk} which corresponds to the transmission-path characteristics TLk in
Now, let us fabricate (synchronously received) fabricated pilot-frame r*pm(i), which would be obtained when a pilot-frame were to be transmitted using a basic spreading sequence gm(i) for data, instead of using basic spreading sequence gp(i) for pilot. This response is L chip components which are produced from both modified channel-response {{circumflex over (μ)}} to be described later, and gm(i) at a fabricated frame generator FFG in
where ρmi(i=0,1,2, . . . L−1) is the i-th chip amplitude and prime symbol ′ indicates that the deviation caused by AWGN is included.
Assume here that enveloped pilot-frame ep(i) has been transmitted simultaneously using carrier-waves for pilot and data, f′k and fk, as shown in Eq. (18). The receiver is assumed to demodulate the received frame with the local carrier-waves {circumflex over (f)}′k and {circumflex over (f)}k, generating synchronously received frames r*p(i) and r*D(i) in Eq. (23) and (24). These frames are assumed to be applied respectively to two matched filters MF(z) similar to those shown in
In this case, fG is the shift frequency in Eq. (17) while jTc is the delay time of the j-th delayed wave. When carrier-waves carrying the signal have different frequencies, a phase rotation θ takes place in their channel-response. Therefore, when producing fabricated pilot-frame r*pm(i) to analyze data-frame r*D(i), the corrected response {{circumflex over (μ)}} must be used.
[Direct Analysis Method]
The data-frame in Eq. (24-A) is represented by means of the fabricated pilot-frame in Eq. (27), as follows
where, x′D(i) is AWGN generated based on xD(i) and xp(i) in Eqs. (24) and (27). Considering that βm consists of M components, the pilot-frame may be rewritten as follows by using an expression with transposed elements in Eq. (27) (changing the subscript order)
where, ρ′, β′, and d′ are respectively the pilot matrix with size (L×M) consisting of ρ′nm as the element on the n-th column and the m-th row, an unknown matrix consisting of β′n as the element, and the received data-matrix consisting d′n as the element. Prime symbol ′ indicates what deviated components (Δpnm, Δβn, Δdn) due to AWGN shown in Eq. (30) are contained. The method of detecting the transmitted information by solving Eq. (29) is called a direct analysis method.
[Maximum Likelihood Judgment Method]
Normally, the solution can be obtained for Eq. (29) when M≦L. Even when M>L, the solution can be obtained according to a method such as the maximum likelihood solution, as long as γ≦L, considering that (M−γ) elements among M elements consisting of unknown set {β} takes zero. Three methods to obtain the maximum likelihood solution are specifically described as follows.
Solution (a):
Decrease M unknown variables in Eq. (29) and select an optional integer MR that satisfies the following equation.
Let the 0-th arrangement-set among NR reduced size unknown sets be {βR}0=(β0R, β1R, . . . βm . . . βM
NR=MCM
Prepare similarly NR(=MCM
Generally, the solution of above equation is obtained under a condition that MR satisfies Eq. (31).
When all of γ components of selected information were to be contained in set {βR}0 stated above, the solution {βR′}0 of Eq. (33) would take the maximum likelihood solution ({βR}0H) that is one of correct combinations [such combinations, that γ elements in the above reduced unknown sets {βR}0 take ±1, while the remaining (MR−γ) elements take 0]. If it is not satisfied, the above solution takes considerably different values from any combination of the correct values. Let us define deviations representing the distance between the m-th unknown (solution) βmR′ among MR elements of {βR′}0 composing of a solution of Eq. (33) and the transmitted information βm=0, ±1 by the following equations
Here, the minimum value of the three deviations for one unknown solution is taken and represented as ΔmA(εΔm+, Δm−) or ΔmB(εΔm0). Let us align ΔmA, and ΔmB in the order from small to large, respectively, provide new superscripts A and B with respective values, provide individual ordinal numbers n=(1,2, . . . ) with the elements composing of respective sets, and represent these deviation sets as {ΔA}0={Δ0AΔ1A . . .
Δγ−1A}0, sets {ΔB}0={Δ0B
Δ1B . . .
ΔM
This is the evaluation function for v being fixed. (Such candidates that the number of members composing ΔnA takes a number other than γ should be excluded.) For NR unknown sets, the similar evaluation functions are obtained. If Sv′ of these sets takes the minimum value, then said unknown set {βR′}v′ is obtained as the maximum likelihood set {βR}v′H. By making hard decisions whether this set is positive or negative, the detected set {{circumflex over (β)}s}=({circumflex over (β)}0s, {circumflex over (β)}1s, . . . {circumflex over (β)}γ−1s) corresponding to {βs} is obtained. Applying {{circumflex over (β)}s} to a parallel to serial converter p-s in
For a case of γ being not fixed, {{circumflex over (β)}S} may be determined based on γ which satisfies the following equation
Solution (b):
Select sets Us(s=0,1,2, . . . Nu−1), each consists of γ pieces of M fabricated pilot-frames r′pm. There are Nu=MCγ sets as the possible combinations. On condition of excluding the case of all polarity inversion, 2γ−1 pieces of virtual (transmitted information) sequences Ch(=c0, c1, . . . cγ−1), each consisting of γ fabricated pilot-frames whose polarity take +1 or −1 can be produced. By multiplying the element frame consisting of each set of Us by the element (polarity) of Ch sequentially, plural pieces of receivable fabricated transmission-frames Fsh are generated. This combination amounts to 2γ−1Nu, in excluding the case of polarity inversion. For simplicity, a case for M=4 and γ=2 is explained as follows.
As the fabricated pilot-frames in this case are r*pm(i)(m =0,1,2,3), let one set selected from those frames be U0[γ*p0,r*p1] according to γ=2. As a result, fabricated transmission-frames Fsh(i) are given by the following equations, based on U0 and Ch, where two virtual sequences C0(=1,1) and C1(1,−1) exist
Similarly, fabricated transmission-frames are obtained for the other sets U1 to U5, producing twelve frames in total. The 0-shift-correlation value between synchronously received data-frame and each of the fabricated transmission-frames is given by
Among the twelve frames, a fabricated transmission-frame whose RSh takes a value closest to ±1 is identified using the following evaluation functions
The subscript s of Ssh± determines βm corresponding to the combination and order of r*pm(i), while h and +, − determine the polarity of {circumflex over (β)}m. For example, if the evaluation function taking the minimum value is S00+ or S01−, the detected value of the transmitted information {{circumflex over (β)}S}=({circumflex over (β)}0S, {circumflex over (β)}1S) may be obtained from the following relationship
Solution (c)
Derive an analyzing sequence wm(i) that satisfies the following equation, in which the 0-shift-correlation value between wm(i) and fabricated pilot-frame r*pm(i) in Eq. (27) takes 1, while the 0-shift-correlation value between wm(i) and each of MT(γ<MT≦M−1) pieces of r*pm′(i)(m′≠m) takes 0:
Since r*pm(i) has L chips in length, wm(i) also takes L in length. Considering [m=0,m′ε0,1,2, . . . MT,MT≦(L−1))<M] as an example of the combination, the following equation is obtained using an analyzing sequence omitting the subscript m, w(i)=(w0, w1, . . . wL−1)T
shown in
where ρ′mn is the element given by Eq. (27). For m fixed, produce multiple combinations of m′, that is denoted by [m′], amounting to Nm′=[(M−1)/(M−1−MT)]+1, ([ ] is a Gauss symbol). Let wm[m′](i) be the analyzing sequence corresponding to each of the combinations. Thus, Nm′ analyzing sequences stated above can cover the (M−1) kinds of sequence areas. Using these sequences, the 0-shift-correlation value given by the following equation is obtained
If all of γ components of transmitted information {βs} is contained in βm and β[m′], these correlation values indicate values extremely close to ±1. Therefore, by identifying a set of γ components taking the nearest correlation values to ±1 among Rm[m′] determined for all of MNm′ combinations, as the maximum likelihood set, using the evaluation function similar to that in Eq. (39), the respective γ correlation values R{circumflex over (m)}[m′]({circumflex over (m)}=0,1,2, . . . γ−1) become the maximum likelihood solution. The detected values of the transmitted information {{circumflex over (β)}s}=({circumflex over (β)}0s, β1s, . . . {circumflex over (β)}γ−1s) can be determined by making these solutions on the hard decision.
[Correlation Analysis Method]
Instead of analyzing r*D(i) in Eq. (24) with r*pm(i) in Eq. (27), both of them are subject to correlative analysis with an identical analyzing sequence ym(i) and the outputs may be used to establish a system of linear equations. The partial circuit of a receiver of the second embodiment-example using this correlation analysis method is shown in
Namely, ym(i) is a sequence orthogonal to fabricated pilot-frame r*pm(i), except the 0-shift value.
Then, fabricated pilot-frame r*pm(i) and data analyzing sequence yn(i)(n=0,1,2, . . . M−1) with subscript changed from m to n are applied to a pilot-response generator PGN in the figure, to obtain the 0-shift-correlation value (correlation coefficient) between both of them
[where r*pm(i) contains AWGN, so that pnm in the above equation contains the deviation due to noise. Here, the deviation is assumed to be 0]. When assuming pnm is an entry on the n-th column and m-th row, the pilot-response (Decorrelation) matrix P with size M×M is generated. P is parameters defining mutual relation between M kinds of data-frames that may possibly be received from u0 and analyzing sequences yn(i). This is also an isolated pilot-response which is not affected by disturbance due to the pilot- and data-frames received from the other stations and modulated by different carrier-waves.
On the other hand, synchronously received data-frame rD0*(i) which are outputs of the data demodulating circuit DD0 in
From Eq. (28), (45), and (46), Φn is expressed as follows
With Eq. (47), a following system of linear equations with M unknowns is obtained
[P][β]=[Φ] (48)
Taking the deviation due to AWGN contained in r*p(i) and r*D(i) into account as in the case of Eq. (30), and representing the value containing such deviation by the following equation is obtained
where P′, β′ and Φ′ (are respectively a pilot-response-matrix with size (M×M) consisting of ρ′nm as the entry, an unknown matrix consisting of β′n as the entry, and a data response matrix consisting of Φ′n as the element. In β-AYZ in
Align Δm+ and Δm− in the order from small to large and provide a new superscript A and an ordinal number n with each value, so as to be Δ0A<Δ1A . . . <Δγ−1A. Select γ components from the forehead and make the hard decision whether the corresponding estimated value {β′} takes negative or positive. Then, the detected values of the transmitted information, {{circumflex over (β)}s}=({circumflex over (β)}0s, {circumflex over (β)}1s, . . . {circumflex over (β)}γ−1s) can be determined as outputs of β-AYZ. Thus, the detected values of the transmitted information-set, {{circumflex over (b)}}/u0 can be determined based on {{circumflex over (β)}s}. When the rank of matrix P′ reduces in solving Eq. (49), any one of said solving methods (a) to (c) can be applied.
According to the above method when a transmit-data-frame consisting of γ components, that is a modulated sequence-set, selected from M kinds of frames is received via multi-transmission-paths, γ binary information {βs} carried by the respective sequences can be detected, by demodulating the received signal to a baseband signal and analyzing it, for example, by obtaining β0sg0s(i) without being disturbed by other multiple sequences βmsgms(i)(m≠0) transmitted simultaneously, and by identifying the above set consisting of γ modulated sequences. In this case, if deviations Δρm, Δdm, Δpnm, and ΔΦn contained in Eq.(30) and (50) are large, this processing may be affected by disturbance of other sequences transmitted simultaneously. But such disturbance can be sufficiently alleviated by selecting a set of M kinds of sequences with small cyclic cross-correction-values.
Note that, when determining the transmitted information estimate value β′0, for example in Eqs. (29) and (49), the energy of the principal wave and all the delayed waves contained in the received flock-frame carrying this information value are used in the form of a row vector, such as (ρ00, ρ10, ρ20, . . . ρM−1,0) or (p00, p10, p20, . . . pM−1,0). On the other hand, since a Rake reception method, which has been described as Method (C), uses a received pilot-response containing interference-wave components, it is difficult to use all the energies of delayed waves. However, the method of the present invention uses the isolated pilot-response according to the principle described above, therefore an ideal Rake reception characteristic may be achieved.
In the above description, z(i) and ym(i) were used as pilot and data analyzing sequences. Any other optional sequences may be used instead of them. For example, the use of gm(i) instead of ym(i) produces the same result if there is no noise.
Since the analyzing method using ym(i) has such a characteristic as to provide high sensitivity in detecting r*pm(i) component contained in rD0*(i), it results in enhancing the regularity of the pilot-response-matrix P and making solution β′ less sensitive to of noise ΔPnm and ΔΦn.
The method described above uses a set of M kinds of different sequences. By extending this method so that the respective τ(=0,1,2 . . . . L−1) shift sequences may be added to M′ kinds of basic sequence gm(i), the number of sequences increases to LM′. According to this idea, consider a cyclically shifted multiple sequence method, with which γ sequences of LM′ kinds of sequences are selected for transmission. When applying the above method to the direct analysis method using
M=M′L, (52)
the solution can be obtained using Eq. (29).
On the other hand, when applying this sequence set to the correlation analysis method, it is required to use matched filters MF[yn(i)](n=0,1,2, . . . M′−1) as the correlators in
On the other hand, the fabricated pilot-frames of this method is given by r*pm(i−τ) which are the cyclically shifted sequences of r*pm(i). The τ shift correlation with analyzing sequence yn(i) is given by the following equation
Using Pmτnτ′ and Φnτ, a system of linear equations with size M×M may be obtained as, similarly to the case in Eq. (49)
[P′][β′]=[Φ′]. (55)
Solving the above equation to obtain γ components of {{circumflex over (β)}s}, a transmitted information-set {{circumflex over (b)}}/u0 is detected.
Therefore, this method achieves large-capacity transmission using the minimum number M′ of the basic spreading sequences. When using self-orthogonal sequences (such that the auto-correlation takes 0 at the shift positions except 0-shift. Two-phase sequences with length L=4, four-phase sequences with length L=16, etc are well known) as gm(i), it is easy to solve the system of linear equations in Eq. (55).
With M′=1 is assumed for the above sequence population, the M kinds of spreading sequences can be made, for example, with only cyclically shifted sequences of g0(i). This is a cyclically shifted single sequence method, and if auto-orthogonal sequences is to be used as g0(i), an excellent error rate characteristic may be obtained.
Now, consider another example of using M kinds of spreading sequences. Let us produce a product of each column (Walsh function) of Hadamarl matrix with size N×N and a partial sequence Ms(s=0,1,2, . . . S−1) with sequence length L, for example, which has been extracted from a long M-sequence. Assuming the product to be a spreading sequence gns(i), a set of NS spreading sequences {gns(i)} can be obtained. As sequences gn′s(i) and gns(i)(n′≠n) are orthogonal each other, this method can generate a sequence set characterized by a large population, small cross correlation, and short sequence length. The present invention can use such sets as stated above.
[Data Channel Sharing Method]
As the third embodiment-example, let us explain a data-channel sharing method, in which data-frames of all users are transmitted using a common carrier-wave fD. The transmitter and receiver can be expressed by using the block diagram in
M=KM′ (56)
kinds of sequences.
The transmitter according to the present invention is designed by replacing the k-th data carrier-waves fk in
thus producing M′ frames in total.
The above channel response {uk} is applied to synchronous circuit SYN in which the frame pulse eF to designate the synchronously received frame-period is produced. eF is set so that the boundary of enveloped frames incoming from all the users may not be included in this period. Similarly to the case in
r*D(i) is applied, together with M=KM′ fabricated pilot-frames, to the analyzing circuit β-AYZ. Generating a pilot-matrix ρ′ using the fabricated pilot-frames, each is made by transposing the user related components (attached by superscript k) of {ρ′mi} in Eq. (27), the following equation is obtained. The analysis is performed by solving the equation.
The size of matrix ρ′ in the above Eq. is (L×M). When M=KM′=L, the solution is normally obtained. For M<L, the previously stated method to obtain the maximum likelihood solution, (a), (b), and (c), can be applied. In this case, the parameters used in the previous method are to be converted so as to be (M→KM′, γ→Kγ).
Thus, in the unknown set {βk′}, γ (γ=2 in the figure) information-components each user transmitted are contained. By solving the above Eq., Kγ information-components in total can be obtained. Namely, the embodiment in
Based on Kγ transmitted information thus obtained, the transmitted information-set {{circumflex over (b)}}/uk of the user uk with circuit p-s shown in the figure can be obtained.
Note that the correlation analysis method may also be applied to the third embodiment-example. In this case, the correlation-outputs p*nm and Φ*n between each of rpmk*(i) and rDk*(i), and data analyzing sequence ynk(i) are obtained, respectively, and a system of linear Equations similar to Eq.(49) is solved, to determine unknown set {βk′}.
In this embodiment-example, if setting M′=1, a single-ary system is realized, in which each user uses a single spreading sequence gk(i) and is supported by the isolated pilot. The above methods can be also applied to the receive-demodulation of this system.
[Complex Number Developed Sequence Extending Method]
As the channel-response {uk} generally takes complex numbers in the above-described system, fabricated pilot-frame r*pm(i) takes complex numbers even when transmit basic sequence gm(i) is a binary phase or real-number sequence. As a result, it has been described that all of data-frames r*D(i), pilot-response p′nm, and data-response Φ′n take complex numbers. In solving Eq. (29), a condition M L is required mathematically, regardless of whether the constituent elements of the equation are real numbers or complex numbers. By noticing here that M is restricted by the length L of basic spreading sequences, the fourth embodiment-example method in which this sequence length L is artificially increased at a receiver is described.
This method is to develop a synchronously received frame with L chips taking complex numbers, and convert it into a frame with 2L chips taking real numbers. Each chip amplitude of fabricated pilot-frame r*pm and received data-frame r*D(i) in Eq. (27) and (24), respectively, is represented by complex numbers in the following equation
where j=√{square root over (−1)}, and the chip amplitude in Eq. (25) is represented with the real part having subscript R and the imaginary part having the subscript J, such as ρ′m0=ρ′m0R+jρ′m0J. By regarding the imaginary part of the above Eq. as another real part, a frame excluding the sum symbol + and imaginary symbol j is made as follows
where the element alignment order can be arbitrary selected. (However, the same alignment order must be used for pilot and data-frames.) Thus, the double length frame consisting of 2L chips can be produced. The important matter is that all of the above double length frames can be regarded to be of made from random sequences because the real part and the imaginary part of channel-response {uk} are mutually independent random variables. [For a transmission-path without multi-path, independence of the above real and imaginary parts is lost. To cope with such situation, it is enough to use a four-phase sequence consisting of random sequences with real and imaginary parts being mutually independent, as the transmission basic sequence gm(i).]
In applying the principle of this sequence extending method to the first embodiment-example, the parameter changes L→2L in Eq. (29). In consequence, when Eq. (29) is to be solved according to a general solution method, a condition M≦L is required in the above description. However, this restriction may be alleviated as follows by using the above double length frame
M≦2L. (63)
If the previous solutions (a), (b), and (c) are to be applied to a case of M>2L, the use of double length frame can improve considerably the error rate characteristics caused by noise. For cases where this principle is applied to the second and fourth embodiment-examples, respectively, it becomes easy to increase M and KM′ as the analyzing dimension increases.
When this is applied to the first embodiment-example, a conversion from r*D(i) in Eq. (61) to r*DE(i) in Eq. (62) is performed at sequence extender SQEXD shown in the dotted line frame in
[Diversity Received Type Sequence Extending Method]
FIG. (b) shows a receiver to detect the signal of the 0-th user u0, as a desired station. The receiver is provided with Nd(≧2) pieces of receive-diversity-antennae ARd(d=0,1, . . . Nd−1). Dpd0, DDd0 and FFGd(d=001) in the figure are respectively the pilot demodulating circuit Dp0, data demodulating circuit DD0, and fabricated frame generator FFG which were used as the circuits of receiver in
Pilot demodulating circuit Dpd0 outputs channel-response {μ0}d, and fabricated frame generator FFGd produces a set of fabricated (synchronously-received) pilot-frames {r*pm}d using both the outputs and basic spreading sequence set {gm} not illustrated. These outputs are applied to the sequence extender SQEXp. When m is assumed to be fixed, NdL chip elements are applied to SQEXp, because each frame of {r*pm}d consists of L chip elements. These elements are arranged in a predetermined order on the time axis. An extended fabricated pilot r*pmE consisting of NdL chip elements, arranged as described above is produced. Since there are M kinds of extended fabricated pilots, M sets {r*pmE} are fabricated.
On the other hand, a synchronously received data-frame rpd0*, which is a baseband output of data demodulating circuit Dpdo, is applied to sequence extender SQEXD. Here, similarly to the case previously described, all the elements (NdL chips) of the respective frames, each of which consists of L chip elements, are arranged on the time axis, generating an extended synchronously received data-frame rDE0* consisting of NdL chips is produced.
Similarly to Eq. (61), the inputs to sequence extenders SQEXp and SQEXD may be expressed as follows
where ρ′mi−d and d′i−d are the i-th complex chip amplitudes including AWGN. For example, taking a method in which the inputs in Eq. (64) are sequentially aligned with regard to d in cascade, the outputs of SQEXp and SQEXD are given by the following Eq. for a fixed value m.
For simplicity, using Eq. (64), the constituent elements of the above Eq. may be expressed as follows, for Nd=2
From a set {r*pmE} of the above extended fabricated pilots, pilot matrix ρ′ with size NdL×M is produced. On the other hand, unknown matrix β′ and the extended synchronously received data-matrix d′, consisting of M and NdL elements, respectively, are produced. Using these matrices, a system of linear equations similar to Eq. (29) is produced. Therefore, the number of dimensions available to obtain the solution with a general method increases to M≦NdL.
In the previous description using the circuit composition in
On the other hand, preparing a sequence set {gp}=(gp0, gp1, . . . gp,Nv−1) for pilots, Nv basic pilot-frames pgpv similarly to the description stated above are produced. Sequences made by synthesizing the above data-frames and pilot-frames, respectively, are denoted by sDv(i) and spv(i) in
At the receiver, pilot demodulating circuit Dpd0 is replaced with Dpvd0, while data demodulating circuit DDd0 is replaced with DDvd0 in
Note that the complex number developed sequence extending functions described with Eq. (62) may be included in the functions of sequence extenders SQEXp and SQEXD described above. Since the number of dimensions of the above matrices increases to 2NvNdL, a condition such as to solve the equations easily can be alleviated as follows
M≦2NvNdL (68)
Note also that the principle of the diversity reception method described with
The transmit-diversity-principle may be applied to the above cases. Namely, by increasing the number of transmit-antennas and extended sequences by Nv fold, the number of dimensions of the equations can be further increased.
[Analyzing Matrix Regularization Method]
Regularity of pilot-response-matrix P′ in Eq.(49) degrades generally when kind M of the basic spreading sequences are excessively large and a condition of M≧L for the sequence length L exists. To avoid this problem, known matrix regularization technology using dummy pilot-response [M. Tomita, et al, “An interference analyzing system for CDMA signals utilizing lone pilot-responses and an additional dummy pilot-response”, Technical resport of the Institute of Electronics, Information and Communication Engineers, SST2000-43] may be employed. This known method is technology in which the kinds of symbol frames (number of information) that can be transmitted on an identical symbol-frame-time-width TE is reduced to (L−1) for the spreading sequence length L. In this case, the remaining one is used to send dummy-information which is not actual transmit-information. Namely, there exists a problem that, despite it has a capability of transmitting L pieces of information, only (L−1) pieces of information can be transmitted to enhance the regularity.
In this invention, the above technology is developed further where a method of using additional vectors along with the Dummy-information and the Dummy Spreading Sequence is explained. Namely, h (generally h≧1) fabricated synchronously received frames which are not actually transmitted, are added for the analysis in a receiver. For this purpose, the receiver increases the size of the sequence population, for example, by adding h dummy spreading sequences, to (M+h), where actually M kinds of sequences are used. In other words, pilot-response-matrix P′ in Eq.(49) is modified to expanded matrix P′+ with size (M+h)×(M+h). Assume here that h=1 and the dummy spreading sequence is gM(i). A case in which let the right-most pilot-response corresponding to a fabricated pilot-frame r*pM(i), not contained in the actual synchronously received frame r*D(i), be an additional column vector (P01, P1M, . . . PnM, . . . PM,M)T=(α0, α1, . . . αn . . . αM)T is described.
The receiver produces fabricated pilot-frame r*pm(i) from both dummy spreading sequence gM and channel-response {μ0} between desired station u0 and base station BS, using Eq. (27). Based on this frame, an additional row vector P′M=(P′M0, P′M1, . . . P′Mn. . . P′M,M−1) using Eq.(45) is produced. Column vector A does not contain a noise component while row vector P′M contains a noise component because it is affected by the noise of {μ0}. Column vector A is selected by the receiver so that the regularity of the extended pilot-response-matrix P′+ may be enhanced. In addition, the M-th information is set to be a dummy-information βMD=(βγS)=1 and, the n-th correcting term αn corresponding to vector A are added to the n-th element Φ′n, of the data response matrix on the right side of Eq. (49).
Assuming that the corrected value is Φ″n
ΦΔn=Φ′n+αn(n=0,1,2, . . . M) (69)
where Φ′M is given by Eq. (46) and γM(i) obtained from Eq. (44). (an=0 may be set in the above equation, with βMD=0.)
From these equations, the following equation in which the size of Eq. (49) is enlarged is obtained.
The above Eq. is a system of linear Eq.s in which the additional column vector A and additional row vector P′M are added. In this case, P′+, β″+ and Φ″+ are the enlarged pilot-response-matrix, the enlarged unknown matrix, and the corrected data response matrix, respectively, and A is selected by the receiver. Specifically, with |an|=1, select the polarity (or complex number value) of αn so that the value (determinant) of matrix P′+ or the singular value may be increased.
For a case where the rank of P′ reduces by more than two, the processing may be made by increasing the number of dummy sequences h to more than two. In this way, the best characteristics may be provided to P′+. Consequently, with this method not much affected by the effect of additive white Gaussian noise,
highly reliable transmitted information estimate {β′} can be obtained.
Note that Eq. (49) can be solved in most cases under a condition of M≦L. If not, the regularity of these Eq.s can be improved by preparing the means of selecting the above additional column vectors A so that the absolute values of determinants D[P′+] of pilot-response-matrices P′ and P′+ [or the normalized singular values (ratios the minimum to the maximum singular values)] may increase.
Eq. (49) can be solved using a known method in which the elements of the equations contained in Eq. (49) are separated into real and imaginary parts to convert them to equations with doubled sized matrix. However, by producing fabricated pilot-frame r*pm(i), synchronously received data-frame r*D(i), and analyzing sequence ym(i) each of which is a real value frame with double length of 2L chips, and using the previously described complex number developed sequence extending method, Eq. (49) can be solved inmost cases under a condition of M≦2L. If not, the above-described additional column vector A may be added to solved it. Therefore, it is more advantageous to use the additional vector method together with the sequence extending method.
With h>1, the effect of enhancing the regularity of matrix P′ increases. This additional vector method can be applied to solve Eq. (55), etc. With the direct analysis method, column vector P′M in Eq. (70) cannot be obtained. When a system of linear equations with M unknowns is to be solved by setting L=M in Eq. (29), it can be solved by developing further the technology of using the unknown dummy number, the dummy spreading sequence, and the additional row vector stated above, even if the rank of pilot matrix ρ′ were to reduce.
By adding an additional column vector A similarly to the case in Eq. (70), and by introducing further an additional row vector C=(c0, c1, . . . cM−1) and a dummy-information βMD, Eq. (29) is converted to a system of linear equations with (M+1) unknowns as follows
where ρ′+, β′+ and d′+ are respectively the enlarged pilot matrix, enlarged unknown matrix, and compensated data-matrix. Note that dM is an additional constant while Δdm(m=0,1,2, . . . M−1) is the compensating term defined by column vector A and βMD as (Δdm=αmβMD).
For simplicity, let us assume
|αn|=|cm|=1 (n,m=0,1,2, . . . M−1). (72)
Besides, the following relationship is established, because of restriction of transmitted information
Virtual data element dM is given by the following equation
If βMD=0 is assumed for simplicity, the second term in the above equation and Δdm(m=0,1,2, . . . M−1) in Eq.(71) take 0. Considering cm(ε±=1) and βm(ε±1,0), there exists a row vector C so that dM=0 holds good for γ of even number and dM=1 (or −1) holds good for γ of odd number.
Assume, for example, to set γ=2, β0=β1=1 and co=−c1=1, and to provide an optional values to cm(m≠0,1), then dM=0 may be obtained. There exists at least one row vector C that gives dM=0 for each combination of β′+. Therefore, by preparing in advance multiple row vector sets Cl(l=0,1,2, . . . N′c−1), dM=0 for respective unknown sets can be found, using one of vectors in sets Cl.
In Eq. (73), by assuming γ=2 and setting dM=0, and every time of providing one of NC′ kinds of row vectors Cl sequentially, let us set column vectors Al for original pilot matrix ρ′ and Cl so that the regularity of ρ′+ may be enhanced. Using matrix ρ′+, thus obtained, an unknown estimate-set {β′}l is obtained. Using a method similar to Eqs. (34) and (35), let us determine the maximum likelihood set among N′c sets stated above. Making the maximum likelihood unknown set on the hard decision, the transmitted information detected value {{circumflex over (β)}s} may be obtained.
For received pilot matrix ρ′ whose rank reduces by more than two, it can be solved by using additional column vectors A0, A1, . . . Ah−1, row vectors C0, C1, . . . Ch−1 with h (≧2), dummy-information-set {βD}=(βMD, βM+1D, , , , βM−h−1D), and virtual data elements {d}=(dM, dM+1, . . . dM+h−1) to convert it to a system of linear Equations with (M+h) unknowns. An effect of regularity improvement will be demonstrated, when the above method is applied to an ordinary system of linear equations.
[Scrambled Sequence Analysis Method]
In CDMA transmission systems, the service area is divided into multiple cells, each of which has a base station, and users within a cell communicate with the base station in the cell. By allocating the same frequency band to respective cells, enhancing the frequency utilization efficiency is attempted. As a result, the communication is affected by interference not only from other users within the cell, but also from users in the neighboring cells. As described above, according to the present invention, the k-th user in each cell occupies the comb-tooth shaped frequency-slots in
In the previous method in which the k-th user uk uses data carrier-wave fk, each user can use the same spreading sequence set (M kinds of sequences), even if considering inter-station-interference within a cell. If the same sequence set can be allocated to users in a neighboring cell, the system design becomes more advantageous. As a means to realize such same sequence method, let us allocate a scrambling code σc(i)(c=0,1,2, . . . Nc−1) inherent to each cell. Assume here that the number of cells causing intercell interference is Nc=(7). A user in the c-th cell performs chip base multiplication to both spreading sequences gp and gm, producing the following scrambled sequence sets, referring to Eq. (9)
gpc(i) is also produced in the similar manner.
Conventional CDMA systems often use scrambling system. In this case, the transmitter transmits the above scrambled sequences without converting them to the enveloped frames having the guard sequences previously stated, while the receiver performs descrambling with the scrambling code (used by the transmitter) synchronized to the principal wave of the received signal for the desired user. Such conventional practice resultantly performs asynchronous descrambling to the delayed waves coming from the desired station, and performs correlative demodulation to produce odd-correlation-components, so that it may be difficult to utilize the enough energy of the delayed waves for signal detection process.
According to the present invention, the transmitter produces an enveloped transmit-frame by using the scrambled sequence set as the spreading sequences, while the receiver produces analyzing sequence z(i) in Eq. (25) based on pilot and data scrambled sequence sets gpc(i) and gmc(i) similar to those used by the transmitter of the desired station. By using this analyzing sequence, channel-response {μ} is produced. Based on this response and sequence gmc(i), is produced fabricated pilot-frame r*pm(i) in Eq. (27), with which the synchronously received data-frame is analyzed. In this way, demodulation detection capable of using all the energy of delayed waves coming from the desired station can be achieved, by eliminating the most of the inter-cell or inter-sector interference. Therefore, the present invention can provide the methods far more advantageous in terms of the receiving error rate characteristics.
[Pilot-Frame Transmission Method]
The first to fourth embodiment-examples described with reference to
a) shows the k-th transmit signal sk(t) of two users uk(k=0,1) in a data-sharing-channel-type time-division-pilot-transmission-system. The figure shows seven extended frames (with period TE) which include the frame-sequence ordinal number n,n′(n≠n′) at the end of the subscript in sk(t). In this case, pilot-frames spmk(t)(n=0,3, . . . ) and data-frames sDn′k(t)(n′=−1,1,2,4,5 . . . ) are arranged on the time axis as shown in the figure. This figure indicates the case in which insertion ratio of pilot and data-frames is 1:2 (equivalent to the pilot overhead αp=½ to be described later). sk(t) modulates the carrier-wave with orthogonal frequency fk in Eq. (18) and the modulated output is transmitted from transmitter TX(uk).
Receiver RX(u0) extracts the synchronously received frame corresponding to pilot-frame spm0(t) in time-division-manner using the frame pulse. This frame can be used to produce channel-response {μ0} with a circuit approximately similar to pilot-signal generating circuit Dp0 in
λ=fd/fl. (76)
The pilot-frame insertion frequency (or αp to be described later) can be reduced, for λ<<1. Note that receiver RX can identify the order of pilot-frames and data-frames using known synchronization technology.
It is also possible to allocate common comb-tooth shaped frequency-slots to all the users in a cell for pilot-transmission and to allow each user to use the frequency-slots in the time-division-manner. The composition of baseband received frames, received by base station BS using such common pilot-channel-type time-division-transmission-system, is shown in
On the other hand, the data-frames received by BS is represented as rDnk(t)(n=−1,0,1, . . . ). This is a continuous waveform on the k-th carrier-wave fk (or a common carrier-wave fD), which has a composition approximately similar to the data-flock-frame rDf(t) in
The time slot of synchronizing frame rF(t) is supplied by a timing circuit in the receiver of base station BS. The transmit-timing of user uk is controlled so that the synchronously received data- and pilot-frames, rDk(t) and rpk(t) may arrive at the center of or near by slot rpk(t), that is the k-th slot away from this synchronizing frame-slot. Therefore, a certain synchronously-received frame does not contain neighboring flock-frames coming from different users.
For frequencies fp and fk of carrier-waves carrying rp(t) and rDk(t) in
If this time-division pilot-frame-transmission-method is applied to the common data-channel-method of the third embodiment-example, by replacing fk with fD=f00+fG, the number of the carrier-waves of N=2 is obtained as given by
By designing the required parameter using this principle, RX can receive the isolated pilot-frames without being subjected to any disturbance due to data-frames or pilot-frames coming from other users.
[Description of the Partial Circuits]
a) shows the continuous time correlator (matched filter) Cor1(q). r(t) is a continuous time received frame waveform [L chip sequence waveform obtained by multiplying g(t) in Eq. (2) by information β is assumed], while q(t) is a continuous time chip-waveform in Eq. (3). Both waveforms are multiplied at the multiplier MOD and the resultant output is applied to the integrator I. If the output of I is applied to the AND-gate A the integrated value r(i) is obtained sequentially with the trigger input d(t) designating the end of integration, as the 0-shift cross-correlation output. Integrator I is reset at every trigger timing point. This circuit is used as Cor(q) in
b) shows a circuit to produce the 0-shift cross-correlation output between discrete time waveform r(i) obtained in FIG. (a) and data analyzing sequence y(i). All the processing is performed on the discrete time axis iTc. Therefore, when compared with FIG. (a), the integrator I becomes an adder. The trigger point of trigger input d(t) becomes [t=n(LTc), n=0,1,2, . . . L] with the frame number n, and the output Φn is obtained at every LTc. This circuit is used as Cor(yn) in
c) shows a matched filter which produces the cross correlation output in FIG. (b) for all the shift positions (τ=0,1,2, . . . L−1). Here is explained an example circuit which obtains cyclic cross-correlation-function between received frame r(i) and pilot analyzing sequence z(i) defined in Eq. (25).
In the figure, D is a delay circuit by Tc seconds; Sh is a circuit having such a function as to delay by Tc-seconds and to shift sequence z(i) cyclically by one chip; and Cor−j(j=0,1,2, . . . J−1) is a correlator similar to FIG. (b). In the figure, the inputs to the j-th correlator Cor−j can be viewed as r(i) and z(i−j), when considering them by neglecting the delay time. Accordingly, the output is the j-shift-correlation-value μj, occurring at the time position delayed by jTc. Thus, the sum {μ} of all the correlator outputs is obtained. If all of elements are to be obtained simultaneously, J components of μj are obtained as parallel outputs by eliminating D and delay function in S′h. (As the input to the correlator is a complex number, the complex operation processing is performed at each correlator.) This circuit is also used as matched filter MF(z) in
[Power Bandwidth Characteristics]
Let the number of chips v required to transmit one bit, as an evaluation measure indicating the frequency utilization efficiency of this system. v is represented by the following equation
where,
Converting the value of the above equation to the information quantity that can be transmitted over the occupied bandwidth of 1 Hz, approximately (2/v) bits/Hz is obtained. Note that αp<1 can be achieved, when the data-sharing-channel-type- or common pilot-channel-type-time-division-transmission-method as described above is used. When the third embodiment-example as described above is used, αD<1 can be also achieved. When the cell size is relatively small, αh can be reduced. For a large information quaintly I per symbol, a long extended frame-period can be used, consequently αh can be reduced.
Now let us consider the transmission power of a user located at an average distance from the base station. As the degradation of the error rate characteristics due to additive white Gaussian noise on the pilot-frames is more considerably lager than that due to noise on the data-frames, the pilot-frame power PP must be L-times larger than, according to the basic spreading sequence length, the unit power PD0 of data-frame for a case of γ=1. On the other hand, for a low-speed moving body, λ=fd/fI<<1 is established in Eq. (76), therefore, PP can be chosen to be a sufficiently small value by using values obtained by integrating a number of neighboring pilot frame-responses as the pilot response. [Assuming that the number of frames to be integrated is Nα, the noise power contained in the pilot-response becomes 1/Nα] Here, let the total transmission power of each user be as follows
PU=Pp+PD=(η+γ)PD0 (80)
If λ<<1 and η=PP/PD0<<1 are established, then the power can be reduced to PU≅γPD0. Alternatively, instead of reducing η, v can be further reduced by setting αp<<1. With γ=around 2–3, it is not required to consume excessive power for a transmitter of a mobile station which favors low power dissipation.
Consider applying this invention to the down-link transmission. In this case, a common pilot-frame can be used, so that the number of carrier-waves to be allocated to the pilot-frame transmission may become one, resulting in dramtic reduction in αp=1/K in Eq. (79). The transmit-power of the base-station-transmitter for the down-link, when all the users are located in an average distance, is given by
i PB=(η+Kγ)PD0 (81)
Since the base-station can practically increase the transmit-power, it is possible to increase γ and resultantly to decrease v. When the maximum value of γm is used, γ of the above equation should take an average less than γm. The product of the above transmit-power and the chip/bit is a power-bandwidth-product that is required for transmission of one-bit
information, and given by the following equation
PB=vP (82)
where PU or PB is used as P. It is considered that the less PB is the more advantageous the system becomes.
Now, consider the design-example shown in Table 1.
X, Y, and W are example-systems for up-link-transmission, while Z is one for down-link-transmission. X is also a system shown in
When compared with the systems of using a fixed value for γ, the systems using the maximum value γm demonstrates further advantageous characteristics. Thus, the present invention can realize systems having extremely higher frequency-utilization-efficiency than that of conventional systems.
The extended frame-period and header sequence-period are given as follows, when the information-transmission-rate of the system is assumed to be R(bits/sec).
Th=10.8 μsec is obtained in the above design-example (X) for R=100 kbps. When the cell radius is assumed to be 1 km, the maximum delay time is τM≅1.8 μsec. Therefore, τM<<Th indicates that the quasi-synchronous condition in Eq. (19) is sufficiently satisfied.
Note that the present invention has an ideal Rake-reception-function. This function is also called path-diversity, which utilizes the power of not only the direct wave but also multiple delayed waves, resulting in improvement of the error rate characteristics. Moreover, by adding the two kinds of sequence extending methods to the systems shown in Table 1, the number of sequences, M, or the number of users, K can be increased, resultantly materializing extremely superior systems which are capable of reducing value PB to a fraction of the values shown in the table.
According to the present invention as described for the first embodiment-example, in order to transmit transmit-information of I bit per symbol, a transmitter creates M kinds of enveloped sequences in which guard sequences are added to a repeated spreading sequence, from which γ components are selected to produce an original information set {β} consisting of (M−γ) components of 0 and γ components of non-0. Using this set, a selected information-set {βs} consisting of γ components is produced. By modulating the respective selected enveloped sequences with {βs}, a baseband data-frame is produced. In the similar manner, using one pilot basic sequence, an isolated pilot-frame is produced. Each data-frame modulates an orthogonal individual carrier-wave so that data-frames transmitted by respective users may not be interfered one another, generating a frequency-division type radio-band transmit-data-frame to be transmitted. A receiver receives data-flock-frame and pilot-flock-frame corresponding to the previously-described transmitted frames, analyzes the former on the basis of accurate channel responses produced from the demodulated output of the latter, and estimates the influence of interference-wave-components from the analytical result. As a consequence, the receiver eliminates the influence of interference-components from the received demodulated components, so that transmit-information of I (for example, I=13 with M=65 and γ=2) per symbol can be detected without being subjected to interfering disturbance. In consequence, a high frequency-utilization-efficiency is achieved. Moreover, the transmit-power can be kept low, unless an excessively large value is selected for γ so that this method may be applied to an up-link-transmission.
The invention shows a direct analysis method used by a receiver which operates for the transmit-functions of the invention according to the first embodiment. In the receiver, a pilot-matrix is produced from both channel responses produced from synchronously received pilot-frames and the respective spreading sequences, and a received data-matrix is produced based on a synchronously received data-frame. With the direct analysis method solving a system of linear equations in M unknowns, produced using of these matrices and an unknown matrix, respective transmitted information can be detected without being subjected to the disturbance due to the other spreading sequence components transmitted simultaneously. Therefore, even when the number of sequences of spreading sequence population M or the number of selected sequences γ is excessively large, for example, an estimate value β0s′ of information β0s carried by the 0-th spreading sequence g0s can be obtained without being subjected to the disturbance due to the other spreading sequence gms(m≠0) components transmitted simultaneously. Moreover, this analysis is performed on the basis of the channel response obtained with an isolated pilot, and it utilizes the same received signal power as that of an ideal Rake-reception. Therefore, this method is effective in achieving excellent error rate versus SN (signal/noise) ratio characteristics.
The invention is a method to transmit the respective user data-frames with a common carrier-wave fD, over the common comb-tooth shaped frequency-slots, where the k-th user transmitter transmits the pilot-frame using frequency-division-transmission with the carrier-wave f′k over the k-th comb-tooth shaped frequency-slots. Namely, this method is effective in saving the bandwidth to be occupied, because all the users share the same bandwidth for data-transmission.
The invention is methods related to the transmission of accurate pilot-frames. The transmitter provides an enveloped sequence, in which guard sequences are added to a core sequence made by repeating a basic spreading sequence for pilot. When transmitting the baseband transmit-pilot-frame made by multiplying the enveloped sequence by the pilot information, such synchronizing technology is provided that the received timing deviation of flock-frames which the receiver received from respective user stations may be kept within the range of guard sequences. In addition, respective users transmit the transmit-pilot-frames over the respective comb-tooth shaped frequency slots in frequency-division-manner by using carrier-waves with mutually-orthogonal frequencies (belong to a carrier-wave set with the frequency interval which is the reciprocal of the synchronously received frame-period). As a consequence, the pilot-frame concerned may have the orthogonal relationship with the data-frames transmitted by all the stations or the pilot-frames transmitted by the other stations. Alternatively, such technology is provided, that the pilot-frame uses a carrier-wave with the same orthogonal frequency as the data-frame and it is transmitted in the time-division manner with the data-frame. Alternatively, such frequency-division, time-division combined technology is provided, that with all the pilot-frames are transmitted on the time axis of common pilot-frame in time-division-manner by using a carrier-wave common to user stations.
According to the present invention, the receiver can receive an accurate isolated pilot of the desired station, without being disturbed by the interference due to the data-frame of the own station as well as pilot- and data-frames transmitted by the other users, and can produce an exact channel response based on the received pilot. The error rate versus SN ratio characteristics of the embodiment-examples according to the present invention can be remarkably improved due to high purification of this pilot. The time-division-transmission of pilot-frame is advantageous in that the pilot overhead can be reduced.
The invention uses the correlation analysis method, instead of the direct analysis method of the receiver of the first embodiment, Namely, the correlation-functions of synchronously received pilot- and data-frames with the same analyzing sequence y is obtained, and using its outputs, that are the pilot-response and data-response, a system of linear equations with size M×M is established, and by solving it the transmitted information is obtained. This method is free from a restriction of limiting the size to the sequence length L, and can produce a system with size M xM for M kinds of sequences prepared in the transmitter. Besides, this method enables the analysis focused on specific sequence components, which is effective in increasing the degree of freedom in solving the equations.
The invention is such technology is provided that the effective length of pilot (response)-matrix is doubled by means of the complex-number developed type sequence extended method, the dimensions of the system of linear equations is doubled. Namely, according to this technology, the sequence length of synchronously received data-frame r*D which the receiver has received is equal to the sequence length L of transmission basic spreading sequence, but the L components of chip elements with the r*D complex amplitude are separated into the real and imaginary parts. The imaginary part is assumed as a new real part, and both real parts are added to make up a double-length sequence-frame for analysis.
The invention is such technology that the dimensions may be increased by means of sequence extending with transmission-diversity and diversity-reception. Namely, Nv pieces of transmit-diversity-antennas and Nd pieces of receive-diversity-antennas are provided, respectively, on the output side of the transmitter and on the input side of the receiver, and plural (Nd) received inputs are applied to plural (NvNd) pieces of pilot- and data-receive-blocks which have such functions as pilot-signal-demodulation, channel-response-generation, fabricated-pilot-frame-generation, and data-frame-demodulation. The respective NvNd fabricated pilot-frames with L chip elements which are the outputs from these blocks are aligned on the time axis, generating an extended fabricated pilot-frame with effective sequence-length NvNdL. On the other hand, NvNd of data-frames with L chip elements, which are outputs from the above blocks, are similarly aligned to produce an extended data-frame. Thereby the dimensions of the system of linear equations can be increased to NvNdL.
The invention is such technology that the sequence extending technology may be combined. Thus, the invention provides such technology as to increase the effective length of spreading sequences to be processed by the receiver. The technology not only enhances the regularity of said receive-pilot-matrix and pilot-response-matrix, but also increases the number of sequences M or the number of available users K to be accommodated. Therefore, the technology is effective in enhancing the frequency-utilization-efficiency of the systems, while reducing the power-bandwidth-product.
The invention is such technology that problems taking place whenever the kinds M of basic spreading sequences prepared by the transmitter is excessively larger than the spreading sequence length (effective spreading sequence length) may be overcome. On the above-stated condition, the rank of pilot-matrix and pilot-response-matrix produced tends to reduce. By utilizing such character that the number γ of selected sequences to be transmitted is sufficiently smaller than M or the spreading sequence length, plural virtual solution sets are obtained. By introducing an evaluation function indicating the difference from the true values, one of these virtual solutions (sets) is determined as the maximum likelihood solution (set). Since this allows M to increase, there is an effect in improving the frequency-utilization-efficiency.
The invention enables an increase in the information quantity per symbol and thereby improves the frequency-utilization-efficiency, by allowing selection of γ under a condition of γ≦γm, instead of fixing γ, that is the number of the spreading sequences to be transmitted by the transmitter. This technology can be achieved only when used together with the high-level interference-wave-elimination technology of the present invention.
The invention is such technology as to add h(≧1) dummy information and an additional column vector A into the system of M linear equations. In the case of , by adding an additional column vector A and an additional row vector C, it is converted into a system of (M+h) linear equations, thereby the regularity of the pilot (response) matrix may be enhanced by selecting the additional vector A, when solving a system of M linear equations the receiver has produced by based on pilot-response-matrix and data-frame-response. Note that the additional row vector C is selected to obtain the maximum likelihood solution. This technology is effective in minimizing the influence to the solution due to the noise and error components contained in the pilot-response-matrix by enhancing the regularity.
The invention is to provide such technology that respective users in each cell generate a scrambled spreading sequence set inherent to the cell based on the M kinds of common basic spreading sequences, as the M kinds of basic spreading sequences, by allocating a cell-inherent scrambling sequence to each cell. Despite users, mobile stations, move to and from in many cells, they can always use the same basic sequence set in the all cells, thereby of the system scale can be simplified. Moreover, in this invention, the receiver produces the above scrambled sequence set and performs demodulation analysis based on the said principle using this set and the channel response. Therefore, the superior interference (due to inter-cell and intra-cell) separation function can be achieved and effective utilization of all the energy of delayed waves coming from the desired station can be utilized. Therefore, this method realizes more advantageous transmission characteristics than that of conventional systems.
The invention provides M′N kinds of sequence population made by multiplying M′ kinds of basic spreading sequences by Hadamarl codes with size N×N. A transmitter produces data-frames by selecting γ components of the spreading sequences from the population, and transmits it. The present invention is effective in generating multiple kinds of spreading sequence sets with small cross-correlation based on a few basic spreading sequences for transmission.
The invention provides LM′ kinds of sequence population in correspondence to the spreading sequence length L, by producing the cyclically shifted sequences based on each spreading sequence of M′ kinds of basic spreading sequences, and then adding them to the basic spreading sequences. A transmitter produces data-frames by selecting γ components of the spreading sequences from the population, and transmits it. The receiver obtains a data analyzing sequence corresponding to the basic spreading sequence, and prepares M′ pieces of matched filters which are matched to the analyzing sequences, outputting M=LM′ pieces of correlation-outputs. By solving a system of linear equations in M unknowns based on such outputs, the transmitted information is detected. Accordingly, LM″ pieces of sequence populations can be used on the basis of M′ kinds of sequences, so that the information quantity per symbol may increase. Application of the present invention technology to down-link-transmission enables the use of a common pilot-frame. Besides, the down-link has the synchronous transmission characteristics, so that the number of guard-chips can also be reduced, thus reducing the overhead of pilot- and guard-sequences. Consequently, a higher frequency-utilization-efficiency for the down-link than that for the up-link can be achieved.
Consider the application of the present invention to radio LAN. Since LAN is a stationary system, the transmitting frequency of pilot-frames can be decreased. Therefore, the overhead for pilot-transmission can be reduced by the use of a common carrier-wave for pilot-frame-transmission. As a result, a high frequency-utilization-efficiency may be achieved.
For the reason described above, application of the transmission technology according to the present invention to cellular mobile communication systems or radio LAN systems will achieve remarkable performance. When the additional vector type high accuracy solution method according to the present invention is applied to all systems requiring the solution of a system of linear equations containing noise and error components in multiple unknowns, remarkable effects can be obtained.
Number | Date | Country | Kind |
---|---|---|---|
2001-117506 | Mar 2001 | JP | national |
2001-367937 | Oct 2001 | JP | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/JP02/02134 | 3/7/2002 | WO | 00 | 9/11/2003 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO02/073853 | 9/19/2002 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
6201802 | Dean | Mar 2001 | B1 |
6563856 | O'Shea et al. | May 2003 | B1 |
6721349 | Willenegger et al. | Apr 2004 | B1 |
Number | Date | Country |
---|---|---|
7-143110 | Jun 1995 | JP |
9-116462 | May 1997 | JP |
2000-354021 | Dec 2000 | JP |
2001-94466 | Apr 2001 | JP |
Number | Date | Country | |
---|---|---|---|
20040131030 A1 | Jul 2004 | US |