The present invention relates to a method, system and computer program for improving a communication system; and in particular, a communication system comprising a transmitter and receiver communicating over a plurality of sub-bands of differing power spectral densities.
Broadband powerline communication (PLC) is a technology that modulates a radio signal with data and transmits the signal on existing electricity powerlines in a band of frequencies that are not used for supplying electricity. In particular, current PLC systems transmit signals at relatively low frequencies (i.e. <30 MHz). High maximum throughput is achieved by employing modulations with a large number of bits per second per carrier per hertz (i.e. bpc/Hz) in the PLC modulation schemes. These systems typically define maximum and useable bps/Hz for each carrier. The transmitter and receiver are often capable of negotiating the used bpc/Hz for each carrier according to a received Signal to Noise Ratio (SNR) so as to optimize the channel capacity between any two nodes (within the limits of the defined modulation parameters). The available channel capacity between any two nodes of the powerline differs with the frequency of a transmitted signal (because of different attenuations, effects of multi-path delays and clock accuracy and noise). The available dynamic range of the implementation and the defined parameters of the modulation in the communication scheme also limit the achieved throughput.
The Federal Communications Commission (FCC) establishes limitations on the conducted and radiated emissions from electronic devices. Conducted emissions are currents that are passed out through a power cord and placed on a common power net, where they may radiate more efficiently because of the larger expanse of this antenna, thereby interfering with other devices. The frequency range for conducting emissions is 150 KHz to 30 MHz. Radiated emissions are electric and magnetic fields radiated by a device, wherein these emissions may be received by other electronic devices causing interference therewith. The frequency range for radiated emissions extends from 30 MHz to 40 GHz. There are other regulations in different regions, some specifically for powerline and some for more general applications and some set by different regulatory bodies.
In practice, power lines are neither shielded nor well-balanced. Thus, some of the RF energy they carry is radiated there from. This RF “leakage” can interfere with licensed radio services. Thus, PLC operators are often required to attenuate or “notch” PLC signals in frequency bands where licensed services are in nearby use. Furthermore, some of the regulations and standards require the injected power spectral density (PSD) of the radio signal in a PLC system to be below approximately −80 dbm/Hz in these notches and above 30 MHz, even through a significantly higher (e.g. up to −50 dBm/Hz) PSD can be injected outside the notches and/or below 30 MHz.
For simplicity, the term “notch” refers to a frequency band where the energy level of a PLC signal has been deliberately reduced to prevent interference with other users of the spectrum. The term “sub-band” refers to a frequency band where a PLC signal characteristic differs (e.g. in power level or directionality) from those in the rest of the PLC signal's bandwidth. The term “coverage” refers to the maximum distance between two nodes at which data transmitted there between is still detectable. In PLC it also refers to the percentage of node pairs that can communicate, to a given minimum performance. Similarly, the term “throughput” refers to the rate at which nodes send or receive data on a network. Coverage in a network is generally dominated by a maximum injected power, as node pairs that have difficulty communicating, do so primarily because of channel attenuation and receiver noise. However there are many node pairs where the throughput between them is not limited by the channel, but by the implementation of the communication system.
In future PLC systems (such as those being defined by IEEE P1901 and ITU-T Ghn), there is a desire to increase throughput, whilst maintaining or improving coverage, keeping a reasonable implementation cost and meeting regulatory requirements. One approach to increase throughput, is to increase the bandwidth of the single band starting below 30 MHz to make it go above 30 MHz. However, the severe stepped PSD (difference of ˜30 dB) imposed by the above regulations, makes it difficult to use a communication band extending above and below 30 MHz, at a reasonable implementation cost, because the dynamic range of both the transmitter and receiver must be capable of handling the step in signal power.
The PSD step between the largest and smallest sub-bands within a band and the defined maximum bpc/Hz for each of the carriers largely defines the dynamic range requirements of the transmitter and receiver. The PSD step further drives the level of quantization, noise and linearity required to maintain signal integrity. Linearity, noise and quantization requirements significantly affect the implementation costs of the analog and digital sections of a modem, and lead to practical limits which may make it too expensive or impossible to implement the same maximum bps/Hz in the sub-bands with a lower PSD.
In principle, any PLC system may support multiple carriers operating in separate sub-bands operating with different injected power levels. It may also use only a subset of the carriers to communicate with nodes using only a part of the communication band (e.g. legacy systems extending over 1.8 Mhz to 30 MHz). It may also be part of a system where another independent band is used in parallel in the same medium or across another medium.
According to a first aspect of the invention, there is provided a communication system comprising a transmitter and receiver in communication for transmitting signals through one or more communications channels, wherein at least one of the transmitter and the receiver each comprise a quality measure, which contributes to an overall quality measure for the system and wherein each communications channel comprises a plurality of sub-bands between at least two nodes, wherein the plurality of sub-bands each having a specific maximum power spectral density (PSD) and wherein one or more sub-bands between the transmitter and the receiver are identified as having a higher maximum PSD than other sub-bands and wherein the PSD of the one or more identified sub-bands is reduced in a predetermined manner to thereby increase the overall quality measure for the system.
According to a second aspect of the invention, there is provided a method for a communication system comprising a transmitter and receiver in a communication channel, wherein at least one of the transmitter and the receiver comprise a quality measure, which contributes to an overall quality measure for the system and wherein each communications channel comprises a plurality of sub-bands between at least two nodes, wherein the plurality of sub-bands each has a specific maximum power spectral density (PSD), the method comprises the steps of identifying one or more sub-bands having a higher PSD than other sub-bands between the transmitter and the receiver and reducing the PSD of the one or more identified sub-bands in a predetermined manner to thereby increase the overall quality measure for the system.
According to a third aspect of the invention, there is provided a computer program product stored in a medium readable by a computer machine, the computer program product tangibly embodying readable program means by causing the computer to perform the method of the second aspect.
According to an alternative aspect of the invention, there is provided a method of improving a communication system comprising a transmitter and receiver communicating over a plurality of sub-bands of differing maximum power spectral densities, the method comprising the steps of checking a quality of one or more signals received by the receiver in a one or more of the sub-bands, identifying the one or more signals within the one or more sub-bands which are received best by the receiver, communicating this information to the transmitter, reducing the power at which at least some of the identified signal(s) in the identified sub-band(s) is transmitted by the transmitter, checking whether the receiver is still receiving the identified signal(s) in the identified sub-band(s) without substantial degradation, checking whether the quality of the other signals in at least some of the remaining sub-bands have substantially improved, altering a bit loading of the signals in substantially all of the sub-bands according to the quality of the received signals and continuing to monitor and alter the power at which the identified signal(s) in the identified sub-band(s) are transmitted by the transmitter until an overall quality of the signals received by the receiver in substantially all of the sub-bands, is optimized to increase an available throughput.
The present invention can be applied to only some of the communications between two nodes. The invention can be applied to communications between some nodes of a network and not others. The invention can be applied to communications between some nodes of the network, while other communications to other nodes use a subset of the available signals. The invention can be applied to communications between some nodes of the network, while other communications use a different band to communicate in parallel. The invention can be applied to communications between some nodes of the network, while other communications use a different medium to communicate in parallel.
An embodiment of the invention is herein described, by way of example, with reference to the accompanying Figures, in which:
Referring to
In particular, the signal to noise ratio of a receiver (Rx) can be expressed by the following equation, wherein Power (Tx) refers to the power at which a signal is transmitted from the transmitter (Tx), Attenuation refers to the attenuation of the communications channel 12; and Noise (Rx) and Noise (Tx) respectively refer to the noise of the receiver (Rx) and transmitter (Tx).
This is a first order expression, which demonstrates that in poor channels the injected power (Power(Tx)), the channel attenuation and the noise at the receiver (Noise(Rx)) will largely determine the signal to noise ratio of the receiver (Rx). This same first order expression also demonstrates that in good channels where the attenuation is low and the receiver noise (Noise (Rx)) is not dominant, the signal to noise ratio of the receiver (Rx) is effectively limited by the signal to noise ratio of the transmitter (Tx). Thus, transmitter noise can effectively impose a noise floor on a receiver, thereby limiting the dynamic range of the receiver. Transmitter noise (e.g. from clipping, quantization errors, analogue non-linearities and distortion) is predominantly white noise, and is measurable across all of the transmission sub-bands.
Using the previous example and referring to
In principle it is possible to drop the noise floor in sub-band Sn+1 whilst maintaining the same high-powered transmission in sub-band S2. However, this would be very expensive because of the need for extensive gates and problems with the linearity of the receiver's analogue to digital converter.
The preferred embodiment takes a counter-intuitive approach to this problem by reducing the power level injected in the one or more highest-powered sub-bands, to improve the total SNR of the overall communication system (when the quality of the channel allows this improvement). In particular, reducing the maximum power spectral density of a band causes quantization noise to be reduced in approximately the same quantity. Similarly, reducing the maximum power spectral density of a band causes analogue distortion to be reduced by a factor greater or equal to the reduction in PSD.
Referring to
The approach adopted in the preferred embodiment is the inverse of a traditional power-shaping technique, which typically adds power to a signal. However, as discussed above, because of inter alia FCC regulations, the preferred embodiment operates in an environment wherein it is not possible to add power. Instead, the preferred embodiment operates by effectively reducing power in a first part of a band to reduce the noise floor and improve performance in another part of the band. However, this runs counter to the accepted teaching of increasing power to increase throughput.
Example Implementation
Referring to
Similarly, the maximum allowed power spectral density in the 30 to 50 MHz frequency range is −80 dBm/Hz.
Let the communication system use OFDM with a maximum constellation size of 1024 QAM or 4096 QAM. Furthermore, let the communication system comprise devices using the complete band from 0 to 50 (e.g. P1901 devices) and other devices using only the sub-band from 0 to 30 (e.g. HPAV devices). The large (i.e. 30 dB) difference between the allowable power spectral densities in the two sub-bands, makes digital quantization (and maintaining linearity of the analogue blocks) particularly difficult. The noise floor resulting from the higher power spectral density sub-band (i.e. sub-band S1) means that the signal to noise ratio in the sub-band from 30 to 50 MHz is lower than that needed for maximum performance (i.e. for 1024 QAM a signal to noise ratio of at least 27 dB is required).
The preferred embodiment finds an optimum point at which the signal to noise ratio in sub-band S2 is almost maximum but the signal to noise ratio in sub-band S1 is still good. The preferred embodiment finds this optimal point in a manner that maximizes the aggregated throughput of both sub-bands. For example, let quantization noise be 40 dB. Similarly, let analogue distortion be 35 dB. Thus, the total noise is approximately 34 dB.
However, it will be recalled that the second sub-band (S2) power is 30 dB below that of the first sub-band (S1). Further, let the channel provide flat noise at a PSD of −100 dBm/Hz, i.e. 20 dB under the second sub-band (S2) power level (i.e. at 50 dB in the first sub-band (S1) and 20 dB in the second sub-band (S2)). Thus, the signal to noise ratio in the first sub-band (S1) is 34 dB and in the second sub-band (S2) is 4 dB.
For the present example, let the required signal to noise ratio for maximum performance of a 1024 QAM (i.e. approximately 3 dB less is required each time we go to the previous constellation size, with one bit less) be 28 dB. Let the system also have all possible constellations from BPSK to 1024 QAM in one-bit steps.
In the first sub-band (S1) maximum throughput (10 bits per carrier from 0 to 30 MHz) is achieved. However, in the second sub-band (S2) no throughput (0 bits per carrier from 30 to 50 MHz) is achieved. Thus, the total average is 10*3/5+0*2/5=6 bits per carrier across the whole band. Now, as shown in
Further, let this power reduction lead to a reduction in distortion noise by 10 dB (because it is non-linear). In the first sub-band (S1), the quantization noise is 40 dB. Similarly, the analogue distortion is 45 dB. Thus, the total signal to noise ratio is now 37 dB. In the second sub-band (S2), the quantization noise is 40−30+10=20 dB below the signal.
Similarly, the analogue distortion is 45−30+10=25 dB below the signal. The channel noise power spectral density is still at −100 dBm/Hz which is still 20 dB below the injected power in this sub-band of −80 dBm/Hz. Thus, the total signal to noise ratio in (S2) is 17 dB. Thus, the throughput could be: 10 bits per carrier in the first sub-band (S1) and 6 bits per carrier in the second sub-band (S2). This leads to an average of 10*3/5+6*2/5=42/5=8.4 bits per carrier across the whole band which is an improvement.
Referring to
Furthermore, the method of the preferred embodiment is adaptable so that it need not be applied to all of the nodes in a network. In particular, the method of the preferred embodiment may be used to optimize communications between node A and node B, whereas, communications between node A and node C may be conducted in the traditional (maximal power) manner.
Furthermore, the preferred embodiment is adaptable to accommodate different nodes on the network communicating over different frequencies and different number of carriers. For example, node A may communicate with:
Furthermore, the method of the preferred embodiment may not be uniformly implemented in both communication directions between nodes in the network. In particular, it is possible that the noise floor may differ between the two communication directions between two nodes, even if the attenuation of the channel between the nodes remains the same.
It will also be appreciated that the preferred embodiment may not be applied uniformly across all of a given sub-band, because the sub-band may be divided into further sub-bands therein even down to the individual carriers. In this case, the carriers on these further sub-bands are negotiated separately. More generally, it will be appreciated that not all of the carriers in the ‘sub-band’ need to be have their power spectral densities reduced in order to improve the signal to noise ratios in the other sub-bands. In other words, some carriers could be transmitted with more power than others, based on their respective signal to noise ratios, in an effort to maximize the throughput on every carrier in all of the bands.
Thus, at a network level, the method of the preferred embodiment may be implemented in a local or co-ordinated approach. In the local approach, at each communication link, a transmitter decides the power that must be injected into the link (i.e. the power at which a message must be communicated across the communication link) depending on the type of frame being sent, the channel conditions of the communications link and the number of sub-bands the receiver uses. The local approach could also be implemented by the receiver requesting a certain power spectral density profile from the transmitter. In the co-ordinated approach, a master of the network decides the channel allocation for each node and at the same time the power spectral density that must be used by that node (to this end, the master needs knowledge about every type of device, every channel condition and the type of frames to be sent).
However, regardless of whether a local or co-ordinated approach is used at the network level, fundamentally, the method of the preferred embodiment is based on a negotiation process between the receiver and the transmitter. In particular, referring to
The receiver checks 28 again to determine if it is still receiving a good quality signal in the identified sub-band. The receiver also checks whether the quality of the signals received in the other sub-bands has improved. The receiver communicates 30 the quality information to the transmitter, which alters 32 the bit-loading of the signals in the sub-bands in accordance with the quality information from the receiver.
The receiver and transmitter continue to monitor and alter the power at which the identified signal(s) in the identified sub-band(s) are transmitted until the overall quality of the signals received by the receiver, is optimized to increase the available throughput.
On another note, and referring to
Current PLC systems (e.g. those defined within Homeplug, HDPLC and Universal Powerline Association (UPA)), use multi-carrier OFDM based modulations. However, the preferred embodiment does not preclude the use or combination with other modulation or access schemes.
Similarly, it will be appreciated that whilst the preferred embodiment has been described within the context of a powerline communication system, this particular context has been provided for explanatory purposes only. In particular, the skilled person will understand that the preferred embodiment is in no way limited to this particular context. Instead, the preferred embodiment is equally applicable to any other situation in which signals are communicated in sub-bands with widely differing power spectral densities.
Alterations and modifications may be made to the above without departing from the scope of the invention.
Number | Date | Country | Kind |
---|---|---|---|
0810855.7 | Jun 2008 | GB | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/EP2009/057284 | 6/12/2009 | WO | 00 | 3/16/2011 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2009/150224 | 12/17/2009 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
7203508 | Ohkubo et al. | Apr 2007 | B2 |
7440399 | Julian et al. | Oct 2008 | B2 |
8031686 | Li et al. | Oct 2011 | B2 |
20010021231 | Kroeger et al. | Sep 2001 | A1 |
Number | Date | Country |
---|---|---|
1770870 | Apr 2007 | EP |
WO-2008005507 | Jan 2008 | WO |
Entry |
---|
De Iulis, M., “International Search Report”, for PCT/EP2009/057284 as mailed Aug. 28, 2009, 3 pages. |
Number | Date | Country | |
---|---|---|---|
20110164670 A1 | Jul 2011 | US |