Patent Application
20090207853
The invention generally relates to a communication system such as a mobile communication system. Particularly, the invention relates to a technique for determining, or detecting, the configuration of a sub-frame of a data frame transmitted within the communication system.
The 3rd generation (3G) technologies are becoming the global solution to mobile communications. Universal Mobile Telecommunication System (UMTS) is a major 3G technology. In order to ensure the competitiveness of the UMTS technology, the 3rd Generation Partnership Project (3GPP) has initiated a project named Long Term Evolution (LTE). LTE aims at improving UMTS to cope with future requirements. Such requirements are, among others, higher data rates, better spectrum utilisation, lower latency, improved services, lower costs, and better integration with open standards.
3GPP technical specification TS 36.211, version 8.0.0, describes the air interface recommended by 3GPP for the LTE technology. The air interface is termed as Evolved Universal Terrestrial Radio Access (E-UTRA). LTE and E-UTRA are sometimes used interchangeably.
The proposed E-UTRA air interface uses Orthogonal Frequency Division Multiplex (OFDM) for the downlink transmission. In an OFDM system, the available spectrum is divided into multiple closely-spaced carriers, called sub-carriers, which are orthogonal to each other. Each sub-carrier is independently modulated by a stream of low rate data symbols. Since data symbols are independently modulated and transmitted over a large number of orthogonal sub-carriers, the OFDM system exhibits numerous benefits including robustness against multi-path fading and efficient receiver architecture.
OFDM symbols may be generated using Inverse Fast Fourier Transform (IFFT) digital signal processing as, for example, described in 3GPP technical report TS 25.892. The IFFT converts a number N of complex-valued data symbols used as frequency domain bins into a time domain signal.
In any wireless communication system, data transmission is usually organised into a series of radio frames having a certain period, or length. The LTE system supports a number of frame configurations, an example of which is shown in
The current LTE technology features a number of different configurations of sub-frames comprised in LTE radio frames, particularly for use in downlink OFDM transmission in a Frequency Division Duplex (FDD) mode. The sub-frame configuration of an LTE FDD cell, or service area of an LTE base station, may determine, among other things, the number of OFDM symbols per sub-frame and the duration of individual OFDM symbols.
In a wireless communication system using LTE, the user equipment (UE) needs, during the initial cell search, to detect the sub-frame configuration specific to the LTE cell. This is one of the first steps to achieve synchronisation with the LTE base station thereby to be able to receive the downlink data therefrom. Usually, during the cell search, other parameters such as frequency-offset estimation and/or symbol and radio frame timing may also be detected. Detection, or determination, of the sub-frame configuration is a new problem in wireless communication, which does not exist in prior standards such as the GSM (the 2nd generation) or UMTS standard.
Therefore, it is one object to propose a solution to effectively determine the sub-frame configuration associated with a data signal.
According to a first aspect, a method of determining a sub-frame configuration is provided, the sub-frame being a sub-unit or sub-section of a data frame. The sub-frame comprises a set of data symbols, and each data symbol comprises a cyclic prefix. The basic steps of the method include receiving a data signal comprising at least one sub-frame having a specific sub-frame configuration, wherein there exists a plurality of different possible sub-frame configurations with each sub-frame configuration being identifiable by an associated cyclic prefix length, analyzing the data signal or a portion thereof to identify a cyclic prefix length associated therewith, and determining the configuration of the sub-frame included in the data signal based on the identified cyclic prefix length.
A sub-frame may take different “formats” (i.e. there may exist a plurality of sub-frame configurations). The sub-frame configuration of a received data signal may a priori not be known. Thus, the sub-frame configuration may need to be determined first before initiating further processing operations with respect to the received data signal. In the framework of the disclosure, a sub-frame configuration is associated with a certain cyclic prefix length. Effectively speaking, the sub-frame configuration is identifiable by an associated cyclic prefix length.
The analysis step may be performed in various ways. For example, the analysis step may comprise evaluating one or more hypotheses (i.e. possibilities) of the cyclic prefix length with respect to the data signal or a portion thereof to produce one or more corresponding evaluation results. Based on the one or more evaluation results, a most-likely cyclic prefix length may then be decided, according to which the sub-frame configuration is determined.
When evaluating each hypothesis of the cyclic prefix length, an autocorrelation of a data sample including the sub-frame may be performed to produce an autocorrelation result. The autocorrelation may be performed over a pre-determined period. The autocorrelation may be of a complex-valued nature; for example, a complex congregate of the sample delayed by a pre-determined length is multiplied with a current value of the sample.
The data frame (and its one or more sub-frames) may be divided into a number of smaller data units, or “slots”. One slot may include a certain number of the data symbols comprised in one sub-frame and, in turn, in the entire data frame. Assuming that the number of slots comprised in one data frame equals Cslot and that the duration period of a slot equals Pslot, then the above mentioned evaluation step may further comprise, with respect to each hypothesis of the cyclic prefix length, performing Cslot cyclic accumulations of the autocorrelation result with the period Pslot. As a result of the Cslot cyclic accumulations, an output signal may be produced for the respective hypothesis. It is to be understood that the number of cyclic accumulations performed to obtain the output signal is a design choice. In other words, the output signal could alternatively be obtained from performing more or less than Cslot cyclic accumulations using a suitable period.
This output signal may undergo further signal processing. For instance, a further is number of cyclic accumulations of the output signal may be performed with a periodicity vector so as to produce an accumulation result. The number of such further cyclic accumulations, denoted as C, may equal the number of data symbols in the slot corresponding to the hypothesis under evaluation. Further, for the accumulation result obtained, a maximum magnitude thereof may be computed and may be regarded as the evaluation result for the hypothesis. The C cyclic accumulations with respect to each hypothesis may be performed either serially or in parallel.
As to the periodicity vector mentioned above, it may comprise a set of elements which can form different sequences, with each sequence corresponding to a possible start of the sub-frame. In this case, the C cyclic accumulations with the periodicity vector are for example performed with respect to each possible sequence. Alternatively, the C cyclic accumulation with the periodicity factor may be performed with respect to a certain possible sequence only. Further, the C cyclic accumulation with respect to different sequences of the periodicity vector may be performed either serially or in parallel.
With respect to each different sequence of the periodicity vector, a maximum magnitude of the accumulation result of the C cyclic accumulations may be computed. The largest one of the maximum magnitudes may be selected as the evaluation result for the respective hypothesis.
Following the evaluation of the hypotheses, a most-likely cyclic prefix length can be decided based on the evaluation results. This deciding step may comprise, with respect to each hypothesis, weighting the evaluation result by a factor that accounts for the number of data symbols collected during the respective cyclic accumulations. This weighting computation may be performed either during or after the last pass of the cyclic accumulations. As a result, a weighted evaluation result is produced with respect to each hypothesis.
The values of the weighted evaluation results corresponding to all the hypotheses may be different. Thus, the deciding step may further comprise a step of searching for a maximum value among the weighted evaluation results to thereby decide on the most-likely cyclic prefix length.
The maximum searching may be masked with an a priori sub-frame configuration. This means, for example, that if it has been decided a priori that a certain configuration is impossible, the corresponding evaluation result will be ignored in the maximum search, even though it may yield the maximum value.
According to a second aspect, a computer program product is provided for performing the steps mentioned above when the computer program product is executed on a computing device or a component thereof. The invention can be practised by means of hardware, software, or hardware/software combined. As for a software aspect, the computer program product may be stored on a computer readable recording medium.
According to a further aspect, an apparatus adapted to determine a sub-frame configuration is provided. The sub-frame is part of a data frame and includes a set of data symbols each comprising a cyclic prefix. As said above, there may exist different sub-frame configurations, each of which being identifiable by an associated cyclic prefix length. Particularly, the apparatus comprises the following components: an input unit adapted to receive a data signal comprising at least one sub-frame having a specific sub-frame configuration; an evaluation unit adapted to analyze the data signal or a portion thereof to identify a cyclic prefix length associated with the data signal; and a determination unit adapted to determine the configuration of the sub-frame included in the data signal based on the identified cyclic prefix length; and an output unit adapted to output the determined sub-frame configuration.
In one variation, the evaluation unit is adapted to evaluate one or more hypotheses of the cyclic prefix length with respect to the data signal or a portion thereof to produce one or more evaluation results. In this variation, the determination unit may be adapted to decide on a most-likely cyclic prefix length based on the evaluation results produced by the evaluation unit. The determination unit may further be adapted to determine the sub-frame configuration according to the most-likely cyclic prefix length thus decided.
The evaluation unit of the apparatus may be adapted to perform an autocorrelation of a sample including the sub-frame over a pre-determined period so as to produce an autocorrelation result. Particularly, the evaluation unit may be adapted to perform a complex-valued autocorrelation, e.g. by multiplying a complex conjugate of the sample delayed by a pre-determined length with a current value of the sample.
As already mentioned above, the frame (and its one or more sub-frames) may comprise a set of slots. Assuming that the entire data frame comprises Cslot slots and one slot having a duration period of Pslot, the evaluation unit of the apparatus may be adapted to, with respect to each hypothesis under evaluation, perform Cslot cyclic accumulations of an autocorrelation result with a period Pslot to thereby produce an output signal. Of course, the output signal could alternatively be obtained from more or less than Cslot cyclic accumulations.
The evaluation unit may be further adapted to, with respect to each hypothesis under evaluation, perform C cyclic accumulations of the output signal with a periodicity vector, thereby to produce an accumulation result for the hypothesis. C stands for the number of data symbols comprised in the slot corresponding to the hypothesis under evaluation.
Moreover, the evaluation unit may be adapted to, with respect to each hypothesis under evaluation, compute a maximum magnitude of the accumulation result. The maximum magnitude is regarded as the evaluation result for this hypothesis.
The determination unit of the apparatus may be further adapted to, with respect to each hypothesis of the cyclic prefix length, weight the evaluation result (produced by the evaluation unit) by a factor that accounts for the number of data symbols selected during the respective cyclic accumulations. The determination unit thus produces a weighted evaluation result.
Further, the determination unit may be adapted to search for a maximum value from all the weighted evaluation results corresponding to all the hypotheses evaluated. The search result thus gives the most-likely cyclic prefix length.
As a yet further aspect of the invention, an application-specific integrated circuit (ASIC) comprising the apparatus mentioned above is provided. The ASIC may be adapted for base-band data transmission. The ASIC may be integrated in a data card or a user terminal of a data communication system. Typical examples of such user terminals are mobile telephones, laptops, and Personal Digital Assistants (PDAs), just to name a few. The ASIC, or the data cards or user terminals, may be used in a wireless communication system in accordance with the LTE standard or any other standard.
In the following, the invention will be described in more detail with reference to exemplary embodiments illustrated in the drawings, wherein:
In the following, for purposes of explanation and not limitation, specific details are set forth to provide a thorough understanding of the present invention. It will be apparent to one skilled in the art that the invention may be practised in other embodiments that depart from these specific details. In particular, the person skilled in the art will appreciate that one or more of the functions explained herein below may be implemented using individual hardware components, using software functioning in conjunction with a programmed microprocessor or general purpose computer, using an integrated circuit such as an application specific integrated circuit (ASIC) and/or using one or more digital signal processors (DSP).
The LTE technology that will be considered in the context of the following embodiments supports a number of frame structures for downlink transmission, an example of which is shown in
Although the LTE radio frame shown in
A data frame may comprise data symbols such as, but not limited to, OFDM symbols. Typically, a sub-frame (as well as a slot within the data frame) comprises a set of such data symbols. To each data symbol, a cyclic prefix (CP) is appended as guard period as is well known in the art. One technique of generating the cyclic prefix is to copy the end portion of an (original) OFDM symbol and append the copy to the front of the symbol, thus slightly extending the (original) symbol duration Tu by Tcp. This copying and appending is illustratively shown in
Depending on the number of data symbols comprised in a slot and the lengths of the cyclic prefixes appended to the symbols within the slots, different sub-frame configurations can be designed. For example, three sub-frame configurations may be defined with respect to the LTE radio frame structure of
The “Normal cyclic prefix” configuration is defined as comprising seven data symbols per slot, the first symbol having a cyclic prefix length of Tcp=5.2 μs (i.e. equivalent to the duration of 160 data symbols), and the remaining six symbols having a cyclic prefix length of Tf=4.7 μs each (i.e. equivalent to the duration of 144 data symbols). In case of OFDM data symbols, the sub-carrier spacing is set at 15 kHz.
The first “Extended cyclic prefix” configuration (assuming sub-carrier spacing at 15 kHz in case of OFDM data symbols) can be defined as comprising six data symbols per slot, each symbol having a cyclic prefix length of Tcp-e=16.7 μs (i.e. 512 data symbols). The second “Extended cyclic prefix” configuration (assuming sub-carrier spacing at 7.5 kHz in case of OFDM data symbols) may be defined to comprise 3 data symbols per slot with a cyclic prefix length of Tcp-e2=33.3 μs.
Since a sub-frame configuration is associated with, and thus identifiable by, a particular cyclic prefix length as illustrated in
Referring to
DCPA 12 is a pre-processing and analysis function which performs, among other things, correlation of data samples, cyclic accumulations of the correlation results, and search for maximum magnitude within the results of the cyclic accumulations. As to be described in detail below, all this is done with different parameters which characterize the different sub-frame configurations.
DCPL 14, as a post-processing of the evaluation results from DCPA 12, decides on the most-likely cyclic prefix length. According to the most-likely cyclic prefix length thus decided, the (most-likely) sub-frame configuration is determined.
Apart from the basic functional blocks DCPA 12 and DCPL 14, an input, or receiving function, 16 can be added as an option. Via the input 16, a data signal (or a series of data signals) which includes the sub-frame under question can be received, which is then passed to the DCPA 12 for evaluation. The input 16 may be coupled to receiver components of the mobile terminal.
Likewise, an output or transmitting function 18 can be connected to the DCPL functional block 14. The most-likely cyclic prefix length and/or most likely sub-frame configuration finally determined at the DCPL 14 can be transmitted to other signal processing components of the mobile terminal via the output 18.
Step 1:
The length of the sliding window, w, is a pre-determined compromise parameter, which can be calculated as will be described below.
Boundaries of consecutive data symbols, e.g. OFDM symbols, feature a separation defined by the cyclic prefix length ncp plus the IFFT window length n0. The parameters for the three different sub-frame configurations are listed in the table of
Using the basic time unit Ts=1/(15000×2048) seconds, the full sample rate may be defined as rs,0=1/Ts=30.72 MHz. The receiver can operate at a reduced sampling rate rs=rs,0/D (D being an integer and usually chosen to be a power of 2). Obviously, the parameters in
The autocorrelation result in respect of each hypothesis of the cyclic prefix length is passed to a second step 2 of the DCPA 12, which performs a cyclic accumulation of the autocorrelation result. Particularly, the second step consists of Cslot cyclic accumulations with a time period Pslot, where Cslot is the number of slots in one data frame, e.g. equal to the number of slots in a LTE radio frame, and Pslot corresponds to the duration period of a slot within the data frame. Mathematically, this is:
Step 2:
As a particular example, a period of Pslot=15360/D can be used, which corresponds to one slot (i.e., 0.5 ms duration) in the data frame structure shown in
The result of the Cslot cyclic accumulation with respect to each hypothesis is output from the second step 2 to the third step 3. At the third step, C cyclic accumulations of the output signal (from step 2) are performed with a periodicity vector P to produce an accumulation result, where C is the number of data symbols in the slot corresponding to the hypothesis under evaluation and the periodicity vector P may be expressed as P=[P1, . . . , Pc−1]. Mathematically, step 3 reads
Step 3:
where the sum in the argument is defined to equate to zero for upper limit c=0. The different Pi results from the different cyclic prefix lengths, e.g. in the sub-frame configuration 1 in
In the simplest case with identical CP lengths, Pi=P, the equation simplifies to
In the fourth step 4, the maximum squared magnitude of the accumulated correlation result shall be computed.
Λmax=maxn|Λcyc[n]|2 Step: 4
The computation can be done during or after the last pass of cyclic accumulation.
All the four steps above need to be performed for each of the hypotheses to be evaluated. An exemplary realization is depicted in
n
0
(1)=128, C(1)=7, Pi(1)ε{ncp(1a)+n0(1), ncp(1b)+n0(1)}={137,138} 1.
n
0
(2)=128, C(2)=6, Pi(2)=P(2)=ncp(2)+n0(2)=160 2.
n
0
(3)=256, C(3)=3, Pi(3)=P(3)=ncp(3)+n0(3)=320 3.
Since sub-frame configurations 1 and 2 use the same value for n0, only two parallel processes of step 1 are required. For cyclic accumulation, it is important to note that P(2) and P(3) are fixed, each, while P(1) may vary. This gives the opportunity to distinguish between two modes.
Mode-1: Suppose the sub-frame start is unknown, then there are 7 seven different sequences possible for Pi(1) (at least six times 137 and up to one 138), so that step 3 shall use 7+2=9 parallel cyclic accumulations. In the latter case, a pre-selection (step 4b) may be optionally performed to select, among those 7 hypotheses for sub-frame configuration 1, the maximum one to obtain Λmax(1). Note that the pre-selection is not really necessary in DCPA, since it could equivalently be handled in DCPL with slightly larger interface between these two blocks.
Mode-2: If the sub-frame start is known or can be ignored, one single sequence Pi(1) may be defined. Therefore, it is sufficient to simply compute only one (e.g., the first) instead of all seven periodicity hypotheses. This means that step 3 may use only 1+2=3 parallel cyclic accumulations. This is an interesting alternative, if there is a bottleneck in terms of computational complexity.
In principle, step 2 and step 3 may be exchanged; but this could be computationally expensive in Mode 1, since step 2 needs to be performed for each of the 9 output signals of step 3.
Returning to
Referring to the specific example described for
max(6*3*Λmax(1), 7*3*Λmax(2), 7*6*Λmax(3))
represents the detected sub-frame configuration. The different weighting of the maximum values in the above equation accounts for the different number of data symbols collected during cyclic accumulation, which may be 7 in cyclic accumulation buffer (1), 6 in buffer (2), or 3 in buffer (3). Factor three could be removed from all operands to maximum search, but is left in the equation for clearness of mutual relationship. The maximum search can be masked with a priori possible and impossible sub-frame configurations.
As a by-product of the sub-frame configuration detection, a rough estimate for data symbol timing can be derived from the time position which corresponds to the detected maximum hypothesis. For example, referring to step 4, it can initially be recorded, for each of the up to 9 Λcyc[n]'s, the particular n that yields the Λmax. Then (at after the DCPL 14) the specific n can be retrieved from the 9 candidates for n which corresponds to the overall maximum. This n then gives the desired time position.
Further, a rough estimate for carrier frequency offset can be obtained by evaluating arg(Λcyc[n]) corresponding to the detected maximum hypothesis at the time position of the previous subsection. This estimate exhibits an uncertainty in multiples of a sub-carrier spacing, which is integer multiples of ±15 kHz (for sub-frame configurations 1 and 2) or integer multiples of ±7.5 kHz (for sub-frame configuration 3).
The performance chart of
Solid lines show the Mode-1 algorithm as described above. The slightly degraded performance shown in dashed lines is achieved with the Mode-2 algorithm, which for C(1)=7 computes only the first instead of all seven periodicity hypotheses. This clearly leads to increased false detection of sub-frame configuration 1 and reduced false detection of sub-frame configurations 2 and 3. The false detection rates are much more homogeneous if the Mode-1 detection is employed. Hence, the Mode-1 of the proposed algorithm evaluates all seven hypotheses for sub-frame configuration 1.
When comparing window length w=2 in
For a technical realization, the DCPA 12 and/or DCPL 14 may be implemented in hardware and/or software. As a example, the DCPA functionality 12 may be implemented in hardware exposing a suitable interface to a programmable device, such as a processor or an ASIC, on which DCPL functionality 14 may be implemented in the form of software.
Alternatively, the DCPA 12 and DCPL 14 may both be integrated in an application-specific integrated circuit (ASIC). The ASIC may be adapted for base-band data transmission. The ASIC may be integrated in a data (or network) card or in a user terminal of a data communication system. Typical examples of such user terminals are mobile telephones, laptops, PDAs, just to name a few. The ASIC, or the data cards or user terminals, may be used in a wireless communication system such as LTE.
The techniques for determining sub-frame configuration described above is one ingredient to enable initial cell search in an LTE communication system. It allows using almost all available downlink signal energy for the desired detection task, which may also start at an arbitrary time instant within a radio frame. This is in contrast to Synchronization Signal (SSIG)—based cyclic prefix detection according to which the distance between directly adjacent synchronization signals is determined. In an exemplary LTE realization, the distance between the primary synchronization signal (P-SSIG) and the secondary synchronization signal (S-SSIG) could be detected to obtain an indication of the cyclic prefix length. Such SSIG pairs occur once every 5 ms in the downlink, which provides limited signal energy for detection. A further drawback is the fact that the detection can only be performed at two defined time instants within a radio frame.
What has been described above are various exemplary techniques for determining the sub-frame configuration of an LTE radio frame. While the invention has been described with respect to particular embodiments, those skilled in the art will recognize that the invention is not limited to the specifics described and illustrated herein.
As can be appreciated, the proposed solution according to the invention is not limited to the LTE downlink alone, but is equally applicable to other data communication scenarios involving the similar concept of data unit configuration. Therefore, while the invention has been described in relation to exemplary embodiments, it is to be understood that this disclosure is only illustrative. Accordingly, it is intended that the invention be limited only be the scope of the claims appended hereto.
| Number | Date | Country | Kind |
|---|---|---|---|
| 08002948.1 | Feb 2008 | EP | regional |
This application claims the benefit of U.S. Provisional Application No. 61/030,185, filed Feb. 20, 2008, which is hereby incorporated herein by reference in its entirety. This application also claims priority, under 35 U.S.C. §119, to European Patent Application No. EP 08002948.1, filed Feb. 18, 2008, which is hereby incorporated herein by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| 61030185 | Feb 2008 | US |
