GB2404533A

GB2404533A - Multicarrier transmission using overlapping frequency channels.

Info

Publication number: GB2404533A
Application number: GB0316540A
Authority: GB
Inventors: Panagiotis Fines; Ekaterini Christofylaki
Original assignee: Individual
Current assignee: Individual
Priority date: 2003-07-15
Filing date: 2003-07-15
Publication date: 2005-02-02
Also published as: GB0316540D0

Abstract

Multicarrier transmission to a receiver from multiple antennas arranges transmission pairs to overlap in time and frequency, and sub-receivers recover the data using narrowband embedded pilot sequences. A stream (fig. 3) of pilot (P) and data (Q) symbols are spread onto successive frequency channels (fig. 2). The data channels are allowed to partially overlap in frequency (figs. 10-13) by spacing them on the order of the signal symbol rate, and the data may include parity bits for Forward Error Correction. The pilot symbols can be repeated so as to result in a narrowband spectrum (fig. 7) and used in adaptive channel estimation at the receiver.

Description

1 2404533 A communication system, signals and apparatus This invention

relates to a communication method, transmitted signal design and communication apparatus. The invention aims to increase the power and bandwidth efficiency of communication systems by specifying a particular air interface design configuration with its associated transmitter and receiver.

Various telecommunication systems exist to provide one or two way communications, all aiming to use efficiently the available power and bandwidth resources and share them fairly between the users. There is a fundamental trade-off between the power and bandwidth resources explored by the numerous signal design techniques: modulation, coding and multiple access methods together with their associated communication equipment.

-

The current invention consists of a novel transmission signal design and a novel receiver aiming to increase the number of users sharing a limited bandwidth and at the same time use efficiently the power resource of each user.

In order to extend the communication session the transmitter must efficiently use its power source such as a battery. The current invention achieves this by using one of the modulation schemes: m-ary PSK, offset (or staggered) m-ary PSK, pi/4 QPSK, m- ary QAM, offset (or staggered) m-ary QAM, pi/4 m-ary QAM, MSK, GMSK and CPM. Any of these modulation schemes can be used at the transmitter with channel symbols either unfiltered or filtered at different levels.

Figure 1 shows a possible configuration where Tx', Tx2, ..., Txn denote n transmitters sending user defined independent binary data Me, M2, ..., Mn' In Figure 2 a single user bit stream denoted by M is split by a demultiplexer denoted as De-MUX into Ma, M2, ..., Mn parallel bit streams for transmission. In both cases, at each transmitter end, the user data M, are combined with framing and forward error correction (FEC) data and are modulated and organized as bursts or segments within continuous frames. The n transmitted signal waveforms in space are denoted by So, S2,

., Sn. The transmitted signals travel through free space or transponder satellite or regenerative satellite or any other transmission medium and suffer distortion and noise due to the channel imperfections. The timing of the transmissions So, S2,..., Sn is aligned so that all or a large part of the transmitted signals overlap in time. At the receive end of both configurations shown in Figure 1 and Figure 2, a single receiver exists denoted as Rx, receiving all signals S', i=1,2,...,n through antenna Al and optionally from a second or more antennas denoted as A2,... ,Am which provide some reception improvement due to space diversity. The receiver Rx processes all the incoming signals and outputs the detected user data Ma, M2, ..., Mn. In Figure 2 the received parallel bit streams Ma, M2, ..., Mn are multiplexed together by the multiplexer denoted as MUX, to give a single user bit stream M. One particular case is presented as an example, in which each user transmitter is using offset QPSK (OQPSK) with square root raised cosine filtering and roll-off factor alpha=1. All other modulation schemes mentioned above and other types and levels of filtering, are also applicable. An unwanted effect of using a power efficient modulation scheme is that the occupied bandwidth by each transmitter can be quite extensive. Using the particular example parameters, the transmitted signal utilizes 100% more bandwidth than the theoretical minimum using the classic TDMA/FDMA multiple access schemes. The classic TDMA/FDMA multiple access schemes require that the transmissions could either overlap in time or in frequency but not in both simultaneously. The current invention reduces the spectrum occupancy of a number of such transmitters by allowing them to overlap in frequency as well as in time...DTD: Figure 3, shows the signal design generated by each transmitter. The transmitted bits are mapped to channel symbols, each carrying two bits in this example since OQPSK is used. Each transmitter generates bursts or frame segments within continuous transmissions. The transmitted symbol stream consists of interleaved symbol groups of user related data symbolized by Q., as well as groups of symbols symbolized by P., i=1,2,... ,N that are not related to the user data but are pilots to assist the receiver in the demodulation and detection process. The two types of symbol groups before multiplexing, are shown in Figure 4 and Figure 5. The data carrying symbol groups Q. are shown in Figure 4 where '0' represents a symbol period without signal. The pilot carrying groups P. are shown in Figure 5 and '0' represents the no signal symbol periods. The transmitted symbol stream shown in Figure 3 consists of the summation of the symbol streams of Figure 4 and Figure 5.

In Figure 3, Figure 4 and Figure 5, each individual data symbol is symbolized by q and carries bits of data or Forward Error Correction (FEC) parity bits or both. Each group of data symbols is denoted as Q., i=1,2,.

,N-1 and consists of any number of data symbols q...DTD: In Figure 3, Figure 4 and Figure 5, each group of pilot symbols is denoted as P., i=l, 2,..., N and consists of R symbols each named as pa where i=1, 2,..., N denotes the group index and j=1, 2,..., R denotes the position index within a group. Also the number R of pilot symbols per group could be zero or any other number and can vary in each P. group. The P. pilot symbol group can carry any bit pattern of the modulation scheme used. In this example each symbol carries two bits with possible values '00', '01', '10', '11', but in general, there are no restrictions and could carry a value beyond the specific modulation type used on the data sections Q.. Consecutive pilot symbols within a group are organized as one of the following: a) R times repetition of the same bit pattern, e.g. p,=P,2=. .. =P,R b) Alternating bit pattern for odd and even indexed symbols, e.g. Pn=-P,2= P,3=-pi4=- À À=p'R-l À À À=-PIR c) Any other bit pattern carried by the symbol sequence P,LP,2. P,R that could result in narrowband spectrum In the described example we will assume case a) above in which the P. group consists of R times repetition of the same symbol and, hence, most of the energy of each P. group is concentrated within approximately R times narrower spectrum than the bandwidth of the data section. The P. pilot group sequence can be a known sequence at the receiver or, optionally, could carry some information for the receiver to be used during the demodulation and detection process, such as signalling of the type and parameters of the modulation and the FEC for the data sections Q. . - 3,

Figure 6 shows in the frequency domain the spectrum occupancy of the data section groups Q. only and ignores the pilot sections P.. The transmitted signals are practically synchronized and partially overlap in frequency. In this example spectral overlap occurs, by tuning the frequency of each transmitter apart from each other to be approximately equal to the symbol rate frequency but, in general, other frequency spacing is also possible. The spectrum of the data sections Q. transmitted from transmitter TO are received as A, the data sections Q. transmitted from transmitter Tx2 are received as X2, and so on until the data sections Q. transmitted from transmitter Txn are received as Xn. Hence, A, X2,..., Xn partially overlap in frequency as shown in Figure 6. The invention does not require precision on the transmission start instant and frequency tuning for each transmitter.

Figure 7 shows the spectrum occupancy considering the P. sections only and ignores the data sections Q.. The spectrum of the pilot sections P. transmitted from transmitter Tx, are received as Ye, the pilot sections P. transmitted from transmitter Tx2 are received as Y2, and so on until the pilot sections P. transmitted from transmitter Txn are received as Yn. By reducing the spectrum occupancy of each P. by R times, the interference stemming from the adjacent channels is reduced to such a level that the P. sections can be regarded as non-overlapping in frequency and Ye, Y2,..., Yn can be processed and detected by a conventional TDMA/FDMA receiver. Furthermore, the energy of each P. group is R times above the energy of a single pi, pilot or data symbol improving the performance of the receiver operating on the P. groups only. The main function of such a receiver is not only to detect the P. but also to detect the propagation channel characteristics of each received signal such as the number of multipath components, relative propagation delays, carrier frequency, phase, amplitude and burst epoch, interpolate these parameters over the data sections and use them during demodulation and detection. Hence, the P. symbol group sequence provides the carrier reference that each receiver needs to derive and use during demodulation and detection. Furthermore, the carrier parameter estimates from each P. are applied together with the detected data in order to cancel out the adjacent channel interference over the data sections.

Following the frequency planning of Figure 6 for a large number of transmitters, approximately half bandwidth is used when compared to the classic TDMA/FDMA multiple access technique. However, this gain in bandwidth cannot be explored by conventional TDMA/FDMA receivers who will fail due to the high levels of adjacent channel interference caused by the overlapping regions. It should be noted that a conventional receiver could only operate successfully in the presence of transmitters that are either odd indexed (e.g. A, X3, ...) or even indexed (e.g. X2, X4, ...) of Figure 6, so that the transmitted spectrums do not overlap. Instead, the novel signal design should be processed by a novel receiver able to receive all carriers as follows.

As it is shown in Figure 1, one receiver unit is required to receive and recover the user data sent by all the transmitters. This receiver unit consists of a number of identical sub-receivers of Figure 8, one for each transmitter, and interconnected together as shown in Figure 9.

Each sub-receiver consists of various subsystems shown in Figure 8, represented by the numbered boxes 1, 2,..., 8 and the signal flow represented by letters A, B...., H. Subsystem 1 accepts through Al the incoming signal from the antenna(s) or from a lower intermediate frequency version of that signal, and splits it by means of a demultiplexer into two signal streams: B represents the Q. data sections route and G represents the route of the P. pilot groups.

Subsystem 8 accepts through G the pilot sections P. and demodulates the bit values that each Pi group is carrying. While detecting the P. information, other parameters are also estimated such as the channel parameters: number of multipath components, relative delays, carrier frequency, phase, amplitude, transmission epoch (start of the burst or frame) and exact symbol timing. The carrier estimates are taken from the P. time instants and through interpolation give a series of estimates at the symbol rate frequency throughout the data sections. The interpolated carrier and synchronization parameters are output through H and pass to subsystems 3 and 7. Subsystem 8 could be implemented as a classic TDMA/FDMA receiver which operates at the P. group rate and not at the symbol rate and, hence, its performance shall be aided by the higher signal to noise ratio.

Subsystem 2 accepts three inputs: the data sections Q. through B. the regenerated interference of the lower in frequency adjacent channel through C, the regenerated interference of the higher in frequency adjacent channel through D. The function of subsystem 2 is to subtract C and D from B. In the case where the transmitter frequency spacing is less than the symbol rate and, hence, more transmitters are interfering then this subsystem 2 accepts more inputs one for each interfering transmitter.

Subsystem 3 accepts at the input the interference reduced signal sections Q. from subsystem 2 and using the carrier and synchronization parameters from subsystem 8 via H. demodulates the incoming signal. The demodulation process consists of carrier frequency, phase and amplitude correction, matched filtering and symbol timing recovery and outputs soft values for each symbol. The output of subsystem 3 consists of soft data and/or FEC bit values.

Subsystem 4 is a soft-input soft-output FEC detection device which accepts the demodulated soft symbol values and outputs the user data bit values together with reliability information for each bit. There are many ways to construct soft-input soft- output FEC decoders, examples are soft output Viterbi algorithm (SOVA), Maximum A-posteriory (MAP) decoders and iterative turbo decoders. In the case of partial response modulation used by the transmitters (e.g. GMSK) the detection unit should also perform equalization in addition to the detection process. The receiver output bits M, shown in Figure 1, are obtained at point E of Figure 8 and they are available at any stage during the process and can be checked for their integrity using, for example, checksums embedded within the user data.

Subsystem 5 converts the incoming data bit values and reliability information from subsystem 4 to symbols and reconstructs the transmitted burst or frame data sections based on the same framing parameters used by the transmitter counterpart. If the reliability of the soft-outputs of subsystem 4 is very high, then the generated symbols are nearly equal to the symbols used at the transmitter modulator before data filtering. - st

Subsystem 6 filters the reconstructed symbol stream from subsystem 5 using the same modulation and filtering characteristics as in the corresponding transmitter. The re- modulated signal is then passed to subsystem 7.

Subsystem 7 accepts the estimated carrier frequency, phase and amplitude, symbol timing and start of transmission from subsystem 8 via H. and applies them to the re- modulated signal from subsystem 6. The carrier estimates from subsystem 8 are used to adjust the frequency, phase and amplitude of the re-modulated signal. The synchronization parameters are used to adjust the epoch of the burst/frame and the symbol timing. After these adjustments, the signal is output at F. The signal at point F is the re-generated signal that replicates the corresponding transmitted waveform, based upon the estimates of the propagation channel distortion (from subsystem 8) and estimates of the data bits (from subsystem 4). The re-generated signal at point F is noise free and any mismatch to the actual transmitted signal is caused by errors during the estimation and detection processes.

Figure 9 shows how all the sub-receivers are interconnected. The received signal A' is applied to all sub-receivers R. where i=1,2,...,n simultaneously via their input signal path A. The Rk sub-receiver, regenerates the signal at F corresponding to its transmitter counterpart, and sends this signal to the D input of the lower in frequency subreceiver Rk and to the C input of the higher in frequency sub-receiver Rk+ . The Rk sub-receiver also accepts via C the regenerated signal from the path F of the lower in frequency sub-receiver Rk ', and via D accepts the regenerated signal from the path F of the higher in frequency adjacent sub-receiver Rk+. Sub-receiver R' has zero input at C and sub-receiver Rn has zero input at D. If some transmissions have not taken place the corresponding sub-receivers can be inactive and their output F and E can be zero. In cases where the transmitter spacing is less than the symbol rate, each sub-receiver accepts more signals, one from each sub-receiver output F related to interfering transmitters.

Once the incoming received signal A' has been applied to all subreceivers, they activate their subsystems 1 and 8 of Figure 8. By reviewing all subsystem 8 outputs, a supervisory circuit can determine the number of operating transmitters, their exact frequency, signal power and epoch. Based on this information the interconnection between the subreceivers can be reconfigured leaving active only those that their corresponding transmission took place. Also the sub-receiver accepting the highest power transmission can be prioritized. While subsystems 1 and 8 of all sub-receivers can be active at least once, subsystems 2 to 7 of all sub-receivers shall iterate and re- exchange their information improving the interference cancellation at the output of subsystem 2 and hence, improving the data bit reliability at E of all sub-receivers.

After several iterations or when the reliability does not improve any more, the data bits from each sub-receiver can be read through E. This terminates the detection process of the receiver.

The invention increases the spectral efficiency of TDMA/FDMA transmission systems by allowing the transmissions to partially overlap in the frequency domain and keeping the interference levels under control and localized. The performance of the receiver depends on: firstly the received signal strength of the individual transmitters, secondly on the amount of distortion suffered by the transmission medium and thirdly on the amount of adjacent channel interference caused by the - 6 frequency overlap. If the loading of an allocated band is not full, then the frequency allocation and interference level can be incremental. A frequency channel allocation system instructs each transmitter at what frequency to transmit. The frequency allocation system can use the following algorithm aiming to minimize the interference between transmitters: When the number of transmitters is small, the frequency grid of the transmitters can be arranged so that they will not overlap in frequency as shown in an example in Figure 10 with four transmitters TO to Tx4. Using the current invention up to seven transmitters can occupy the same bandwidth as shown in Figure 13. Extra transmitters can be allocated between the non-overlapping transmitters and hence, the receiver will suffer interference from two adjacent channels.

A more graceful increase of interference levels is shown in Figure 11, where the frequency grid is rearranged in order to accommodate Tx5. In this case Tx and Tx2 suffer half the interference since they overlap on one side only. The same applies to Tx4 and TXs. Also the interference is localised since there are pairs of mutually interfering transmitters (Tx, Tx2 and Tx4, Tx5) but each pair does not interfere with any other transmitter. Tx3 suffers no interference.

Figure 12 shows the frequency grid to accommodate Tx6. By doing so there are two independent triplets of interfering transmitters: TV, Tx2, Tx3 and Tx4, Tx5, Tx6. Only Tx2 and Tx5 suffer full two-sided interference leaving Tx, Tx3 and Tx4, Tx6 with half level one-sided interference. - 7

Claims

I A method of controlling the transmission of data in a communication system, including: a plurality of transmitters transmitting a plurality of signals and a receiver.
2 A method as claimed in claim 1, including independent data supplied to independent transmitters.
3 A method as claimed in claim 1, including data extracted from a single data steam and supplied to independent transmitters.
4 A method as claimed in claims 1 to 3, wherein said plurality of transmission signals comprises a plurality of frequency channels which are successive in frequency.
A method as claimed in claims I to 4, including a receiver receiving the said transmitted signals via a single antenna.
6 A method as claimed in claims I to 4, including a receiver using a plurality of antennas receiving the said transmitted signals.
7 A method as claimed in claims I to 4, wherein the said transmitted signals comprising data, include a sequence of pilot symbols and data symbols with reference to Figure 3.
8 A method as claimed in claims I to 7, wherein the said pilot symbols are spread over the transmission time as groups of any number of pilot symbols with values within a group consisting of: repetition of the same value; alternating values; any other symbol pattern with narrowband spectrum with reference to Figure 7.
9 A method as claimed in claims I to 7, wherein the said data symbols are spread over the transmission time as groups of any number of data symbols with values of: information data bits and parity bits for forward error correction.
A method as claimed in claims 1 to 4, wherein the said transmitted signals comprising a band limited transmission using any band limiting modulation method.
11 A method as claimed in claims 1 to 10, wherein the said transmitted signals are: synchronous or asynchronous and all or sections overlap in time.
12 A method as claimed in claims 1 to 10, wherein the said transmitted signals partially overlap in frequency as a result of transmission channel spacing approximately of the order of the signal symbol rate with reference to Figure 6 and Figures 10 to 13.
13 A method as claimed in claims 5 and 6, wherein the transmitted signal in claims 1 to 4 and claims 7 to 12, is received and is processed by sub-receivers with reference to Figure 8 and interconnected as with reference of Figure 9.
14 A method as claimed in claim 13 including a method to process the received signal in order to separate the data symbol stream with reference to Figure 4 and the pilot symbol stream with reference to Figure 5 in order to be processed separately. - 8
A method as claimed in claims 5 to 6 and 13 told, wherein the pilot symbol streams are processed by dedicated processors to extract the transmission channel state: dominant signal characteristics such as delay, amplitude, frequency and phase and number of multipath echoes and characteristics of each echo such as delay, amplitude, frequency and phase.
16 A method as claimed in claims 5 to 6 and 13 told, wherein the data symbol characteristics such as delay, amplitude, frequency and phase are corrected using the extracted parameters using the method in claim 15.
17 A method as claimed in claims 5 to 6 and 13 told, wherein the signal with corrected parameters is applied to the forward error correction decoder for correcting data errors.
18 A method as claimed in claims 5 to 6 and 13 tol7, further including: the extracted data by the forward error decoder are multiplexed with pilots, re-encoded, re modulated, band limited with the same methods as in claims 1 to 4 and 7 to 12.
19 The regenerated signal with the method in claim 18 is readjusted in delay, amplitude, frequency and phase according to the parameters extracted with the method in claim 15.
The methods in claims 13 to 19 are applied to each signal reception and the output signal using the method in claim 19, is subtracted from the adjacent frequency channel receptions: the next higher in frequency and the next lower in frequency with reference to Figure 9.
21 A method of further including at the said receiver in claims 5 and 6, one or more repetitions of methods in claims 13 to 20.
22 Apparatus arranged to perform the method of any one of claims 1 to 21.
23 A method substantially as herein described with reference to Figures 1 to 2 and Figures 8 to 9.
24 A signal substantially as herein described with reference to Figures 3 to 7 and Figures to 13.
A method of transmitting data and signals over wired or wireless links, comprising signals as claimed in any one claims 1 to 4 and claims 7 to 12.
26 A method of receiving data and signals over wired or wireless links, comprising receiving a signal as claimed in any one of claims 5 and 6 and 13 to 21.