WO2011015802A1

WO2011015802A1 - Iterative successive interference cancellation for asynchronous bandwidth efficient distributed fdma systems

Info

Publication number: WO2011015802A1
Application number: PCT/GB2010/000586
Authority: WO
Inventors: Mohammud Zubeir Bocus; Yue Wang; Justin Coon
Original assignee: Toshiba Research Europe Limited; Kabushiki Kaisha Toshiba
Priority date: 2009-08-04
Filing date: 2010-03-26
Publication date: 2011-02-10
Also published as: GB0913586D0; GB2472569A; GB2472569B

Abstract

A block spread code division multiple access transmission system and a method of and receiver for cancelling interference in such a transmission system A block spread code division multiple access transmission system (100) comprises a plurality of user terminal transmitters (200) which are configured to: (i) apply precoding to each block of symbols by progressively rotating the phase of each block of symbols identically by a user-dependent amount; (ii) spread the precoded blocks in time or frequency using discrete Fourier transform spreading codes; and (iii) apply a guard interval in the form of a cyclic prefix. A receiver (300) uses a successive interference cancellation scheme to process the received signals. The successive interference cancellation scheme may be iterative.

Description

Iterative Successive Interference Cancellation for Asynchronous Bandwidth Efficient Distributed FDMA Systems

This invention relates to a block spread code division multiple access transmission system and a method of and receiver for cancelling interference in such a transmission system.

CDMA is a popular multiple access technique that is used to support multiple users simultaneously in a network. In synchronous CDMA each user is allocated a code to modulate their signal such that the codes of each user are mutually orthogonal. This means that the dot product of vectors representing each code is zero. This is referred to as the cross-correlation of the orthogonal codes being equal to zero. Since the codes are orthogonal, signals constructed using different codes do not interfere with each other and can be separated from each other at the receiver.

Many variants of CDMA exist, including direct sequence (DS) CDMA, multi-carrier (MC) CDMA, cyclic prefixed (CP) CDMA, and chip interleaved block spread (CIBS) CDMA. In addition to these variants, many receiver architectures are often available for implementation in CDMA systems, such as the well-known RAKE receiver, interference cancellation receivers, and receivers that rely on channel equalisation.

Some CDMA schemes are interference limited; that is to say, as the number of users in the network increases, residual interference caused by each user eventually cripples the network, thus rendering simultaneous multiple access nearly impossible. This residual interference generally results from the loss of orthogonality amongst users, which primarily occurs when the channel is temporally dispersive. The channel may be temporally dispersive, for example, due to multipath which occurs when a signal propagates from a transmitter to a receiver over a plurality of routes due to, for example, reflection or refraction, which means that the length of each path of the signal varies resulting in different propagation times for the different paths of the signal.

Several recent developments in block CDMA systems have led to multi-user interference (MUI) free transmission techniques.

Zhengdao Wang and G. B. Giannakis, "Wireless Multicarrier Communications", IEEE Signal Processing Magazine, Vol. 17, May 2000, pp 29-48, the contents of which are incorporated herein by reference, describes so-called 'generalised MC-CDMA¹ (GMC- CDMA).

Shengli Zhou and G. B. Giannakis, "Chip-interleaved block-spread code division multiple access", IEEE Transaction on Communications, Vol. 50, Feb. 2002, pp. 235- 248, the contents of which are incorporated herein by reference, describes chip- interleaved block-spread CDMA

"Performance comparison of distributed FDMA and localised FDMA with frequency hopping for EUTRA uplink," NEC Group and NTT DoCoMo, TSG RAN WG1 Meeting 42 R1-050791, Aug. 2005 and D. Galda and H. Rohling, "A low complexity transmitter structure for OFDM-FDMA uplink systems," in Proc. of the IEEE Vehicular Technology Conference (VTC), vol. 4, May 2002, pp. 1737-1741 , the contents of which are incorporated herein by reference, describe a single-carrier frequency division multiple access (SC-FDMA) (DFT-spread OFDM).

S. Tomasin and F. Tosato, "Throughput Efficient Block-Spreading CDMA: Sequence Design and Performance Comparison," in Proc. of the IEEE Global Telecommunications Conference (Globecom), Nov.-Dec. 2005, the contents of which are incorporated herein by reference, describes a throughput-efficient block CDMA system.

J. P. Coon, "Precoded Block-Spread CDMA with Maximum User Support and Frequency-Domain Equalization", in Proc. of the IEEE International Conference on Communications (ICC), Glasgow, 2007, and GB-A-2433397, the contents of which are incorporated herein by reference, describe a configurable scheme which has led to MUI free transmission techniques. The scheme achieves higher bandwidth efficiency compared to conventional CDMA systems such as CIBS-CDMA.

In these systems, any number of users - up to a given maximum number - can theoretically transmit simultaneously without causing any degradation in system performance. Beyond this maximum number of allowable users, the system becomes interference limited in a similar manner to other CDMA systems.

Below, reference is made to the system model based on J. P. Coon, "Precoded Block- Spread ... Equalization" and GB-A-2433397 as the bandwidth efficient BS-CDMA system. Such systems include the orthogonal frequency-division multiple access (OFDMA) systems using discrete fourier transform (DFT) precoding, and single-carrier based distributed frequency division multiplexing access (FDMA) systems.

Although prior art methods can provide MUI free and inter-block interference (IBI) free transmission using the block CDMA techniques, this can generally only be achieved with ideal assumptions. One of these assumptions is that the receiver receives signals from user terminals with perfect synchronization. However, in practice MUI and IBI occur when the signal reception among different users cannot be perfectly synchronised due to delays in transmission or due to delays from the channel. In these cases the system is not synchronous but is asynchronous, which means that the MUI free and IBI free models described in the prior art are not practically implementable.

In other cases, MUI and IBI could occur due to the insufficient cyclic prefix as described in X. Peng, T. S. Dharma, F. Chin and A. S. Madhukumar, "Novel interference cancellation methods for BS-CDMA in uplink broadband mobile communication systems", in IEEE Communication Letters, vol. 12, No. 8, August 2008, the contents of which are incorporated herein by reference. This document considers interference cancellation for BS-CDMA and proposes a scheme where the mobility condition of the mobile subscribers is used for specifying the order of detection. High mobility subscribers are considered as more severe interferers and are detected first such that their effect on others may be cancelled at an early stage. A multistage successive multi-user interference cancellation (SMC) is also proposed that demonstrates high performance gains when the number of stages is large.

In the case of asynchronous reception, transmitted blocks are no longer mutually shift orthogonal, and MUI and IBI occur, as described in the applicant's patent application no. GB 0817521.8, the contents of which are incorporated herein by reference. Depending on the order of signal reception, very high error floors can be observed. The MUI and IBI are suppressed in the referenced patent application by using a minimum mean squared error (MMSE) equaliser that takes the interference into account. Although the interference can be effectively suppressed by the proposed MMSE equaliser, to achieve optimal performance, the MMSE equaliser can be complex to implement. According to a first aspect of the invention there is provided a method of cancelling interference in a block spread code division multiple access transmission system comprising a plurality of user terminal transmitters which are configured to (i) apply precoding to each block of symbols by progressively rotating the phase of each block of symbols identically by a user-dependent amount; (ii) spread the precoded blocks in time or frequency using discrete Fourier transform spreading codes; and (iii) apply a guard interval in the form of a cyclic prefix, the method comprising:

(a) receiving asynchronously at a receiver signals from a plurality of user terminals;

(b) processing the signal from a first user terminal;

(c) reconstructing the interference caused by the signal transmitted by the first user terminal;

(d) using the reconstructed interference to process signals received from the other user terminals; and

(e) repeating steps (b) to (d) for the signals received from the other user terminals,

wherein the signals are processed in the order of arrival at the receiver.

According to a second aspect of the invention there is provided a method of cancelling interference in a block spread code division multiple access transmission system comprising a plurality of user terminal transmitters which are configured to (i) apply precoding to each block of symbols by progressively rotating the phase of each block of symbols identically by a user-dependent amount; (ii) spread the precoded blocks in time or frequency using discrete Fourier transform spreading codes; and (iii) apply a guard interval in the form of a cyclic prefix, the method comprising:

(a) receiving asynchronously at a receiver a plurality of blocks from each of a plurality of user terminals Ma;

(b) processing a first block ni from a first user terminal mi;

(c) processing a second block (n+1)i received from the first user terminal mi immediately after receipt of the first block ni,;

(d) reconstructing the interference caused by the first block ni and the second block (n+1)i received from the first user terminal mi;

(e) using the reconstructed interference to process a third block ni+1, the third block ni+1 being transmitted by a second user terminal mi+1 and being the next block received at the receiver immediately after receipt of the first block ni received from the first user terminal mi; and (f) repeating steps (c) to (e) for all received blocks, in each case using the reconstructed interference to process subsequently received blocks, wherein for each third block ni+1 no blocks subsequent to the second block (n+1)i from the previous user terminal mi are processed before the third block ni+1 is processed.

One aspect of the current invention is concerned with the reconstruction of the MUI from the main interference contributors and subtracting it successively from the received signal. The process of interference reconstructing can be employed iteratively The detection order (in the first iteration, if iteration is employed) may be in ascending order of delay, i.e., processing blocks of data based on the order of arrival of the blocks and/or the order of arrival of signals from the user terminal. In the case that iteration is used, the detection order in a second iteration may or may not be the same as in the first iteration. For example, the detection order could be the reverse of the detection order for the first iteration.

Interference reconstruction depends on knowledge of delay and channel state information (the channel taps), which can be calculated using known techniques at the receiver/base station (uplink). Note that asynchronous reception is not a significant problem in the downlink, in transmission from typically the base station to user terminals.

The order of processing of blocks may be given by:

Initialization: a = 1 , n(0) =1 and m(0) =1, t=0

for t =1 to Ma * T-1

if n(t-1) ==T

a + =1;

end if

if n(t-1) ==1 or m(t-1) = Ma;

n(t) = n(t-1)+m(t-1)-a+1;

m(t) = a;

else

m(t) = (m(t-1) % Ma) +1;

n(t) = m(t-1) + n(t-1) - m(t);

end if

end for

where: Ma is the number of active user terminals;

T is the number of blocks transmitted by each user terminal;

t is a time index;

a is an index; and

n(t) and m(t) denote the nth block of the mth user terminal respectively decoded at time t.

The blocks may be transmitted using a low-complexity single-carrier frequency division multiple access scheme or a low complexity orthogonal frequency division multiple access scheme.

The step of processing a block may comprise: despreading the block; decoding the despread block; and equalising the decoded block.

The step of processing a block further may comprise: soft bit mapping the equalised block; de-interleaving the mapped block; and channel decoding the de-interleaved block.

The equalising step may comprise equalising the block with a conventional equaliser used for synchronous reception.

According to a further aspect of the invention there is provided a block spread code division multiple access transmission system comprising:

a plurality of user terminal transmitters which are configured to:

(i) apply precoding to each block of symbols by progressively rotating the phase of each block of symbols identically by a user-dependent amount;

(ii) spread the precoded blocks in time or frequency using discrete Fourier transform spreading codes; and

(iii) apply a guard interval in the form of a cyclic prefix; and

a receiver for receiving asynchronously signals from the plurality of user terminals, the receiver being configured to:

(a) process the signal from a first user terminal;

(b) reconstruct the interference caused by the signal transmitted by the first user terminal;

(c) use the reconstructed interference to process signals received from the other user terminals; and (d) repeat steps (a) to (c) for the signals received from the other user terminals, wherein the receiver is configured to process received signals in the order of arrival at the receiver.

a plurality of user terminal transmitters which are configured to:

(iii) apply a guard interval in the form of a cyclic prefix; and

(a) receive asynchronously at a receiver a plurality of blocks from each of a plurality of user terminals Ma;

(b) process a first block ni from a first user terminal mi;

(c) process a second block (n+1)i received from the first user terminal mi immediately after receipt of the first block ni,;

(d) reconstruct the interference caused by the first block ni and the second block (n+1)i received from the first user terminal mi;

(e) use the reconstructed interference to process a third block ni+1, the third block ni+1 being transmitted by a second user terminal mi+1 and being the next block received at the receiver immediately after receipt of the first block ni received from the first user terminal mi; and

(f) repeat steps (c) to (e) for all received blocks, in each case using the reconstructed interference to process subsequently received blocks, wherein for each third block ni+1 no blocks subsequent to the second block (n+1)i from the previous user terminal mi are processed before the third block ni+1 is processed.

According to a further aspect of the invention there is provided a receiver for cancelling interference in block spread code division multiple access transmission system signals transmitted by a plurality of user terminal transmitters, the signals having been transmitted by: (i) applying precoding to each block of symbols by progressively rotating the phase of each block of symbols identically by a user-dependent amount; (ii) spreading the precoded blocks in time or frequency using discrete Fourier transform spreading codes; and (iii) applying a guard interval in the form of a cyclic prefix; the receiver being configured to:

(b) process the signal from a first user terminal;

(c) reconstruct the interference caused by the signal transmitted by the first user terminal;

(d) use the reconstructed interference to process signals received from the other user terminals; and

(e) repeat steps (b) to (d) for the signals received from the other user terminals,

wherein the receiver is configured to process received signals in the order of arrival at the receiver.

(b) process a first block ni from a first user terminal mi;

(e) use the reconstructed interference to process a third block ni+1 , the third block ni+1 being transmitted by a second user terminal mi+1 and being the next block received at the receiver immediately after receipt of the first block ni received from the first user terminal mi; and

(T) repeat steps (c) to (e) for all received blocks, in each case using the reconstructed interference to process subsequently received blocks, wherein for each third block ni+1 no blocks subsequent to the second block (n+1)i from the previous user terminal mi are processed before the third block ni+1 is processed.

According to a further aspect of the invention there is provided a carrier medium carrying computer readable code for controlling a microprocessor to carry out the method described above.

The present invention can be implemented in any convenient form, for example using dedicated hardware, or a mixture of dedicated hardware and software. The present invention may be implemented as computer software implemented by one or more networked processing apparatuses. The network can comprise any conventional terrestrial or wireless communications network, such as the Internet. The processing apparatuses can comprise any suitably programmed apparatuses such as a general purpose computer, personal digital assistant, mobile telephone (such as a WAP or 3G- compliant phone) and so on. Since the present invention can be implemented as software, each and every aspect of the present invention thus encompasses computer software implementable on a programmable device. The computer software can be provided to the programmable device using any conventional carrier medium. The carrier medium can comprise a transient carrier medium such as an electrical, optical, microwave, acoustic or radio frequency signal carrying the computer code. An example of such a transient medium is a TCP/IP signal carrying computer code over an IP network, such as the Internet. The carrier medium can also comprise a storage medium for storing processor readable code such as a floppy disk, hard disk, CD ROM, magnetic tape device or solid state memory device.

The invention will be described by way of example with reference to the accompanying drawings in which:

Figure 1 is a schematic representation of the arrival timing delays from user terminals of a code division multiple access (CDMA) transmission system;

Figure 2 is a schematic representation of CDMA transmission system according to a first embodiment of the invention;

Figure 3 is a block diagram of a transmitter of the system of Figure 2; Figures 4a to 4d are block diagrams of receivers that can be used as the receiver of the system of Figure 2;

Figure 5 is a flow chart showing a method for determining the order in which blocks received at the receiver are to be processed;

Figure 6 is a schematic representation of the order in which user signals are considered at the receiver of Figure 2;

Figures 7a and 7b are tables showing the order in which blocks are processed for a 5 block per user 3 user system and a 4 block per user 5 user system respectively;

Figures 8a and 8b are lookup tables corresponding to Figures 7a and 7b respectively;

Figures 9 and 10 are graphs showing numerical simulations of the performance of the method of the first specific embodiment of the invention;

Figure 11 is a block diagram of a receiver in accordance with a second embodiment of the invention; and

Figures 12 to 14 are graphs showing numerical simulations of the performance of the method of the second specific embodiment of the invention.

A first embodiment of the invention provides a successive multi-user interference (MUI) cancellation (SMC) architecture for block spread code division multiple access (BS- CDMA) transmissions (including low-complexity single-carrier frequency division multiple access (SC-FDMA) and orthogonal frequency-division multiple access (OFDMA)) where DFT spreading codes are used.

Figure 1 illustrates the general system model where interfering signals from other user terminals arrive before and after a signal from a specified user terminal. Compared to the signals from other user terminals, two distinct scenarios arise, namely 1) signals from the other user terminals arrive prior to the arrival of signals from the specified user terminal; and 2) signals from the other user terminals signals after the arrival of the signal from the specified user terminal. For example, when considering the symbols transmitted by user terminal a it can be seen that corresponding symbols from user terminal b arrive before the symbols transmitted by user terminal a and corresponding symbols transmitted by user terminal c arrive after the symbols transmitted by user terminal a. The delay between the ith user and the specified user is denoted as ^T| , where ^T/ can be positive or negative, corresponding to the case where the ith user terminal's signal arrives after or before the specified user terminal's signal, respectively. In this embodiment, for mathematical simplicity the case is considered where the delay is a multiple of the symbol duration, i.e., ^{T/ ~} ' '"' symbols and it will be evident to the person skilled in the art that this invention also applies to cases where the delay is not an integer multiple of the symbol duration.

Figure 2 is a schematic representation of a CDMA transmission system 100 according to an embodiment of the invention. The system 100 comprises a plurality of transmitters 200 and a receiver 300 such as a base station. The transmitters 200 communicate with the receiver 300 wirelessly through a transmission channel 400 with blocks from different transmitters 200 arriving at the receiver 300 asynchronously.

Each transmitter 200 has a structure as illustrated in Figure 3. Data coming from a source 210 is encoded in an encoder 220 before it is interleaved in an interleaver 230 and subsequently mapped in a mapping module 240. The mapped symbols are then precoded in a precoding module 250 by applying precoding to each block of symbols by progressively rotating the phase of each block of symbols identically by a user- dependent amount. The precoded blocks are then spread in time or frequency using discrete Fourier transform spreading codes in a spreading module 260. Alternatively, the precoding module 250 and the spreading module 260 can be interchanged, i.e., blocks of symbols are spread in time or frequency using discrete Fourier transform spreading codes and precoding is applied to each sub-block of data in the spread blocks by progressively rotating the phase of each sub-block identically by a user- dependent amount (as described in page 8 and the first two paragraphs of page 9 and as shown in Figure 1 of GB-A-2433397). Before transmission, a guard interval is applied in the form of a cyclic prefix (CP) in a CP module 270 to separate different spread blocks. The signal is then amplified and transmitted by means of an antenna 280. In an alternative embodiment, one or both of the encoding 220 and interleaving modules 230 may be omitted. In greater detail, the transmitted message is constructed at the transmitters 200 in the form of a stream of bits that may be encoded and/or interleaved. These bits are mapped to complex baseband symbols that are members of a given constellation, such as m-phase shift keying (PSK) or m-quadrature amplitude modulation (QAM), where m is the number of elements in the constellation. The resulting complex constellation symbols are arranged into blocks of K symbols, the 'th block being denoted by s _μu ^,ι -^ { \s _μ J^\iK) »,... ,^>s _μn(^\(^\i + I) >K - I) ^J)Y | ._{p t}._n. _e proposed . f .ramework . , each . symb _uol i b _uliock i i■s precoded 250 with a P* K user-specific (possibly time-dependent) precoding matrix μ-¹ , where P≥K _ and subsequently spread 260 on a block-by-block basis by a length-M spreading code ^c"-<

where ^μ-¹ with C denoting the set of complex n-tuples and ® is the

Kronecker product operator. The skilled person will recognise that this 'block spreading' and precoding can be interchanged. Also, a skilled person will recognise that this 'block spreading' and precoding can be achieved by simple symbol spreading as in a conventional direct sequence (DS) CDMA system followed by an appropriate chip interleaver and precoding module. Note that if L is the memory order of the channel impulse response, it is important that K≥L + l holds true in order to perform low-complexity de-spreading and detection at the receiver. Following the spreading operation, a cyclic prefix is appended to the user's message in the CP module 270. It is assumed that the CP consists of ** symbols so that inter-block interference is mitigated. It is further assumed that the channel remains static for the duration of each individual block.

The blocks transmitted by the transmitters 200 are received at the receiver 300. The receiver 300 of Figure 2 has the structure of one of the receivers 300a, 300b, 30Oc₁ 30Od respectively shown in Figures 4a, 4b, 4c and 4d.

In a case where the blocks of all user terminals are received synchronously, the blocks would be superimposed one on the other and the receiver would just change the spreading code it uses for the despreading operation to obtain the data from the desired user. In practice this would be implemented with a single received path and a bank of spreading codes, where one spreading code is chosen at a time. Alternatively, decoding of each of the synchronous user terminals' blocks could be performed simultaneously with the received signal, which comprises the blocks of all user terminals without delays between them, being replicated the desired number of times and each copy decoded with the desired spreading code.

In the present embodiment where reception is asynchronous, multiple paths are present which arrive asynchronously at the receiver 300. To decode a block from a later user terminal it is necessary to recover the data from the earlier user terminals since they contribute the most to the interference. This recovered data is then processed (using interference reconstruction) and subtracted from the received signal so that the receiver views its version of the signal received from the later user terminal as one in which interference from previous user terminals is minimal. For example, to decode data from user terminal a in Figure 1, it is required that data from the user terminal b (which arrives before data from user terminal a) is recovered. This recovered data is then subtracted from a replica of the received signal before the despreading and decoding operation of user terminal a, and a similar procedure is followed for subsequent user terminals.

It is assumed that the receiver 300 has knowledge of the delays between the different signals originating from the different user terminals. The receiver firstly synchronises to the start of the received signal, which corresponds to the start of the earliest (first) user terminal's first block. There are two possible approaches to the order in which received blocks from different user terminals are processed.

According to an embodiment of this invention, a first approach is to decode all the blocks of the first user terminal and reconstruct the interference when processing signals from the later user terminals. The receiver then moves its synchronisation point to the start of the second user terminal's signal, decodes all its blocks, reconstructs the interference on the third and later user terminals and so on.

According to another embodiment of this invention, a second method is to decode the blocks transmitted by later user terminals as soon as possible and this is the method used in the present invention as described below with reference to Figure 4, which shows the structure of the receiver 300 which processes the asynchronously received blocks of the Ma user terminals. The wireless signals transmitted by users a to Ma of Figure 1 are received at an antenna 310. The first block to be processed by the receiver 300 is the first block received by the receiver 300 and the receiver 300 must synchronise to the start of that block before processing that block. As shown in Figure 1 , the signals from the various user terminals experience different delays before arriving at the receiver 300 such that the blocks arrive asynchronously. To process the signals the users are renumbered according to the order in which they arrive at the receiver 300 - user terminal a as transmitted becomes user terminal 2 as received and user b as transmitted becomes user terminal 1 as received, and this new ordering of the user terminals is shown in Figure 6. Following this renumbering, each block of renumbered user terminal 1 begins to arrive before the corresponding block of renumbered user terminal 2 and each block of renumbered user terminal 2 begins to arrive before the corresponding block of user terminal 3 etc.

Referring to Figure 4a, at the receiver 300a, a wireless signal transmitted by user terminal 1 from transmitter TXb is received at an antenna 310 and the output of the antenna 310 is passed to a CP module 320 which removes the guard interval. After this point the asynchronously received signals are shown in Figure 4 as being processed by different paths.

In relation to the signal received from renumbered user terminal 1 which has had its guard interval removed, the signal is de-spread in a de-spreading module 330. Each of the de-spreading modules 330 in the receiver 300 uses a different de-spreading code, chosen from a bank of de-spreading codes, which corresponds to the spreading code used for that signal in the respective transmitter 200. The de-spread signal is decoded using a linear decoder 340 that is suitable for use with the respective linear precoders 250. The decoded signal is processed with an equalising module 350 and the subsequent signal is de-mapped in a soft bit mapping module 360. Note that the equalizer 350 used in the present invention can be any conventional equaliser used for synchronous reception, such as a linear minimum mean-square error (MMSE) frequency domain equaliser (FDE) or decision feedback equaliser (DFE), as opposed to the optimal MMSE equaliser used for asynchronous reception as described in the applicant's patent application no. GB 0817521.8. The subsequent signal, which corresponds to the interleaved and encoded signal produced by user terminal b is deinterleaved in a deinterleaving module 380 and then decoded by a channel decoder 390.

In an alternative embodiment where there is one of or no encoder 220 and interleaver 230 in the transmitter 200, one or both of the de-interleaving 380 and channel decoding 390 modules may be omitted.

Note that the signal output from the equalising module 350 is also passed to an interference reconstruction module 370. Once the interference to which the block was subjected is reconstructed in the interference reconstruction module 380 the estimated interference from user terminal 1 is passed to a second chain of de-spreading to soft bit mapping modules 330-360, as shown in Figure 4, and this reconstructed interference is used to cancel the estimated interference contributed to by the processed block of the user terminal 1 on blocks received from the second user terminal 2. This process is repeated for all Ma user terminals of the system 100.

In an alternative embodiment as shown in Figure 4b, the output from the equalising module 350 is passed to an interference reconstruction module 370, and the reconstructed interference is removed from subsequent receive chains prior to the decoding module 340.

In an alternative embodiment as shown in Figure 4c, the output from the equalising module 350 is passed to an interference reconstruction module 370, and the reconstructed interference is removed from subsequent receive chains prior to the equalising module 350.

In an alternative embodiment as shown in Figure 4d, the output from the equalising module 350 is passed to an interference reconstruction module 370, and the reconstructed interference is removed from subsequent receive chains prior to the soft bit mapping module 360. Note that in this case, the equaliser modules 350 must be linear, such as zero forcing (ZF) or linear MMSE equalisers, for all receive chains except for the last chain, which can use any equaliser.

In case of perfect synchronisation between the different users in the system, IBI and MUI can be easily mitigated by the careful use of spreading codes, equalisers and decoders. In practical cases, residual interference prevents the use of simple techniques at the receiver. More complex equalisers are required at the receiver to suppress IBI and MUI. Examples of such equalisers can be found in the applicant's patent application no. GB 0817521.8. The present invention allows for quasi- synchronous reception for each user terminal in an asynchronous system such that low complexity equalisation designed for the synchronous reception can be used to achieve performance that is close to that of synchronous reception.

For a block-spread CDMA system with M users, where the users are ordered and indexed according to the times of arrival of their signals, i.e., the user whose signals arrive the first is the first user, and whose signals arrive the last is the M th user. The composite signals at the base station after the removal of the CP and block despreading and decoding can be written as:

ro-1 M

_>(i} = MH_ms_m(i) +∑θ_f^m +∑ ^_β__m + v(i)

fc=l α=*n+l

(1)

where AfH*n^smW _{jg the} ^ _recej_ved) despread block for the mth user before equalization, ^VW is the equivalent noise term, ^⁰-"¹* is the interference term from the ath user whose signals arrive later than the synchronization instant, ^h→m is the interference term from the ύth user whose signals arrive earlier than the synchronization instant. In addition,

(2)

and θ_b→m = D* [Δt_m ^χ6(0 - At_me^L _d ^cpMi + 1)]

(3)

where • a-*m j_{s an} MPxMP upper triangular Toeplitz matrix with its first row being [0, . .. _Λh_a(L - ^] ), . . . ML - Q] _ _where I_a = L - L_CP + τ_a→m ~ l _{L js the} channel memory order, and ^cp is the length of cyclic prefix.

• ^0→m is an MPxMP lower triangular Toeplitz matrix with its first column being

[0,••• ,0, -fn,{0),••• , -h_b{τ^_m - I)F

pLcp

• is a circulant matrix obtained by circulantly shifting the identity matrix down by ^Lcp.

x U)

• ^μ denotes the ith block of signals from the /"th user after block spreading and precoding, i.e.,

x_μ{i) = (c_μ ® A_μ)s_μ(ϊ)

• (5)

• where ^s^ is the ith block of the signals from the /'th user before block spreading and precoding, ^c^is the length-M spreading codes, ^μ is the precoding matrix for the M^'th user, and ^ denotes the Kronecker's product; and

• ^{m m m} is the despreading and decoding matrix for the mth user, and the decoding matrix ^O, js identical to the precoding matrix **-^m.

Assuming the receiver has knowledge of the channel state information and the relative delays of each user, it is possible to decode the received signal for each subscriber in ascending order of delay. In equation (1), the interference that comes from a user terminal that arrives later than the reference user terminal is:

This term can be ignored without incurring major performance penalties, as described in the applicant's patent application no. GB 0817521.8, because it can be shown that the interference from a later user to an earlier user is usually very small. Especially, if the length of the cyclic prefix is greater than the memory of the channel added to the delay, ^α =0, interference is actually zero. The procedure for recovering the data for the earliest of the users is identical to the receiver structure in GB-A-2433397. For the signal from user terminal 2 that arrive second, the effects of all later user terminals can be ignored. The earlier user terminal contributes to the term θ_b→m = D* [ΔL_ra ^χδ(t) - At_me^L _d ^cpχΦ + 1)]

, where b=1, which accounts for interference to signals from this later user terminal.

The principle idea behind this invention is to reconstruct the interference from all earlier user terminals and then subtract it from the signal received from the currently synchronised user terminal. For example, to detect the data of the user terminal 2 that arrives second, the value of βu» = D£ [At_mMi) - ΔL_meJ^cpχ6(i + 1)] (where b=1) is calculated, since the data from the first user terminal 1 has already been recovered. Recovering the data from the second user terminal 2 allows for the reconstruction of interference from the first and the second user terminals onto all later users. This sequential subtraction process is carried out for all user terminals.

As stated above, there are two possible implementation techniques for this interference cancellation system. According to a first embodiment of the invention, all the blocks from the earliest user terminal 1 are decoded and recovered and then signals arriving from the second terminal 2 are decoded and recovered, and so on. The downside of this method however is that since it is necessary to decode the entire signals from all earlier user terminals there might be a considerable delay in decoding signals from the later user terminals. In a second embodiment of the invention, the interference reconstruction and decoding is carried out on a block-by-block basis, or an array decoding process. Referring to the terms in θ_b→m = D* [Δt_mX6(i) - At_me_d ^LCPx_b(i + I)]

, it is observed that interference on the ith block is caused by blocks i and i +1 from earlier user terminals. An illustration of the order of decoding the blocks is given in Figures 7 and 8. In each case, the first two blocks transmitted by the first user terminal whose signal arrives as the receiver 300 are first decoded, followed by block 1 arriving from user terminal 2 (which corresponds to user terminal a of Figure 1). To obtain interference on the second block of the second user terminal 2, block 3 of the first user terminal 1 must be known. Obtaining the second block of user terminal 2 allows for recovering the first block of user terminal 3 (which corresponds to user terminal c of Figure 1). Each time a block for a later user terminal is decoded, it is necessary for an extra block for the previous user terminal to be decoded. The process continues until an extra block of the first user terminal 1 is decoded. For a three user terminal system where each user terminal transmits 5 blocks, the block processing order is shown in Figures 7a and 8a and for a five user terminal system where each user terminal transmits 4 blocks, the block processing order is shown in Figures 7b and 8b.

Though the second technique involves more control on the decoding order, it allows fairness in decoding the data among all the user terminals. There are two possible implementation techniques for the array decoding. The first approach is to use a lookup table, for example as shown in Figures 7 and 8, to determine the order of decoding. Alternately, blocks of data from all user terminals can be decoded in an order determined by the algorithm as illustrated in Figure 5.

It is observed that for decoding the ith block of user terminal m, it is necessary that block i+f-m of user terminal f has been decoded, where f < m. After decoding the ith block received from one user terminal, the receiver checks if the condition mentioned is met for the following user terminal. If not, the block of earlier user terminals needs to be decoded.

Assume that there are Ma active user terminals, each of which has transmitted a total of T blocks. Let t denote the time index. At time t, the nth block of the mth user terminal is decoded, denoted as n(t) and m(t), respectively. For example, n(3) =3 and m(3) =1 means the third block of the first user terminal is decoded at t=3. The order of decoding is described below with reference to Figure 5.

In step S405 the algorithm is initialized such that: a = 1 , n(0) =1 and m(0) =1 , t=0.

In step S410, t is incremented and a "for" loop is begun, which runs from t =1 to Ma ^* T- 1

In step S415, if the block at the previous time t-1 , is equal to the number of blocks transmitted T by each user terminal, i.e. if n(t-1) ==T, then the algorithm proceeds to step S420, where a is incremented, i.e., a + =1 ; else the algorithm proceeds to step S425. In step S425, if the block at the previous time t-1 , is equal to unity, i.e. if n(t-1) ==1 then the algorithm proceeds to step S430; else the algorithm proceeds to step S435.

In step S435, if the user terminal at the previous time t-1 , is equal to the total number of user terminal, i.e. if m(t-1) = Ma; then the algorithm proceeds to step S430; else the algorithm proceeds to step S440.

In step S430 the value of the block number n(t)at that time t is set as:

n(t) = n(t-1)+m(t-1)-a+1;

and the value of the user number m(t) at that time is set as

m(t) = a;

Alternatively, in step S440 the value of the user number m(t)at that time t is set as:

m(t) = (m(t-1) % Ma) +1;

and the value of the block number n(t) at that time is set as

n(t) = m(t-1) + n(t-1) - m(t);

At step S445 the algorithm checks if all n blocks of all Ma user terminals have been decoded. If not all n blocks of all Ma user terminals have been decoded then the algorithm repeats steps S410 to S445. If all n blocks of all Ma user terminals have been decoded then the algorithm ends at step S450.

The pseudocode for the algorithm shown in Figure 5 is given by:

Initialization: a = 1 , n(0) =1 and m(0) =1 , t=0

for t =1 to Ma * T-1

if n(t-1) ==T

a + =1 ;

end if

if n(t-1) ==1 or m(t-1) = Ma;

n(t) = n(t-1)+m(t-1)-a+1;

m(t) = a;

else

m(t) = (m(t-1) % Ma) +1;

n(t) = m(t-1) + n(t-1) - m(t);

end if

end for Following the algorithm shown in Figure 5 and as described above for a three user terminal system where each user terminal transmits five blocks results in the block processing order shown in Figure 8a and following the algorithm for a five user terminal system where each user terminal transmits four blocks results in the block processing order shown in Figure 8b.

Figure 9 and 10 give the plot when the proposed successive interference cancellation method is used, with 16 active users and QPSK modulation. In the figures, user terminal 1 is assumed to be the synchronous user with delay of it being 0. The delays of the rest of the users terminals are 1 ,1 ,1 ,2,2,2,3,3,3,4,4,4,5,5,5 relative to the first user terminal. Moreover, all of the simulations consider an exponentially decaying channel with 6 channel taps. The length of cyclic prefix is 8, and the size of each data block is 16. The performance shown in Figure 9 is that with perfect channel state information at the receiver 300, while that in Figure 10 shows a more realistic case where the channels are estimated by the least-squared (LS) estimator by transmitting two Chu sequences

In the examples shown in Figures 1 and 6 to 10 and as described in the algorithm shown in Figure 5, there are Ma user terminals where each user terminal transmits the same number, n, of blocks and the first block transmitted by each user terminal arrives at the receiver 300 asynchronously but in a time frame less than the length of each block, i.e., all the first blocks transmitted by each of the user terminals 200 begins to arrive at the receiver 300 during the time that the first block of the first received user arrive at the receiver 300.

In practice, there is not necessarily such synchronisation between the transmittal timing of the user terminals and when each user terminal sends a burst of ni blocks the number of blocks in each burst will vary and the overlap between the arrival time of the blocks will vary.

Hence a more general description of the order in which blocks arriving from multiple users can be given as follows.

When a plurality of blocks are received at a receiver from each of a plurality of user terminals Ma; a first block ni received from a first user terminal mi is processed to determine the transmitted symbols. After that block has been processed, a second block (n+1)i received from the same first user terminal mi immediately after receipt of the first block ni is processed next, to determine the transmitted symbols in that block. Once these two blocks have been processed, the interference produced by the first block ni and the second block (n+1)i received from the first user terminal mi is reconstructed.

Next the symbols transmitted in a third block ni+1 , are derived. The third block ni+1 is a block which is transmitted by a second user terminal mi+1 and is the next block received at the receiver 300 immediately after receipt of the first block ni received from the first user terminal mi. The symbol derivation comprises using the reconstructed interference to cancel the interference caused by the first block ni and the second block (n+1)i to the third block ni+1 and then processing the third block ni+1 to determine the transmitted symbols.

Once the first and second blocks received from the first user terminal and the third block received from the second user terminal are processed, the above steps are repeated for all received blocks. According to the present invention, for each third block ni+1 no blocks subsequent to the second block (n+1)i from the previous user terminal mi are processed before the symbols are derived for the third block ni+1.

It is observed that the performances of all user terminals are very close to the case of synchronous reception, which corresponds to the plot of user terminal 1 with or without perfect channel state information at the receiver 300. Furthermore, error propagation is limited to a single block rather than the whole transmission data. The larger the size of a transmitted block is with respect to the order of the channel, the smaller is the degradation due to error propagation. Note that the performance degradation due to error propagation and the size of each transmitted block are inversely proportional, i.e., a long transmitted block facilitates system operation with very little error propagation.

The principle advantage of the receiver employing the successive interference cancellation method is that the system can achieve performance close to the synchronous reception. In the applicant's patent application no. GB 0817521.8, the optimal MMSE equaliser demonstrated error floor reduction in asynchronous reception at the price of high computational complexity where a matrix inversion operation may be involved. This present invention uses the low complexity conventional equalisers designed for synchronous reception while still effectively suppressing the MUI and IBI. A second class of embodiments of the invention will now be described, in the context of implementing iteration in conjunction with some of the technical concepts indicated above. The reader's attention is directed to figure 11 of the drawings. Figure 11 illustrates a receiver incorporating elements of the receivers described above, but also incorporating a facility to iterate the SIC process exemplified so far. It will be appreciated that figures 4a to 4d demonstrate that interference reconstruction can be carried out in various different ways, in accordance with an embodiment of the invention, and that the same concept can be carried across here to the iterative use of SIC.

In this embodiment, for a block-spread CDMA system with M_a active users, the Λh received block after despreading can be expressed as:

Z_m(ϊ) = MR_mS_m(i) +

where ^^^W is the /Yh received, despread block for the mth user before equalization, ^VW is the equivalent noise term, Φ^α→m is the interference term from the ath user whose signals arrive later than the synchronization instant, ^b^^m is the interference term from the ύth user whose signals arrive earlier than the synchronization instant. In addition, φ_α→m

(7)

and θ_b→m = D*

- Δt_meJ^σ"x6(i + I)]

(8)

where • A α^u-^→rn j_s an MPxMP upper triangular Toeplitz matrix with its first row being

[0, ... , 0Λ_a(L - I), - - - A(L - Q] _ _where Z₀ = L - LOP + τ_α→m - 1 _ _{L js the} channel memory order, and ^Lcp is the length of cyclic prefix.

• is an MPxMP lower triangular Toeplitz matrix with its first column being

[0,•■• , 0, -Zi₆(O),■•• , -h_b(τ^_m - I)F

x (i)

• denotes the ith block of signals from the /"th user after block spreading and precoding, i.e.,

*_μ(i) = (c_μ ® A_μ)s_μ(ϊ)

• (9)

• where ^s^ is the ith block of the signals from the Mth user before block spreading and precoding, ^C/MS the length-M spreading codes, ^μ is the precoding matrix for the Mth user, and ^denotes the Kronecker's product; and

• ^{m m m} is the despreading and decoding matrix for the mth user, and the decoding matrix ^rm is identical to the precoding matrix ^-^m.

It can be confirmed that for the received signal of a particular user, interference power from users whose signals arrive earlier dominate the multiuser interference power, and interference power from users whose signals arrive later is usually much smaller than that from users whose signals arrive earlier. I.e., the interference power due to the ^°-*^m terms can be neglected when both the interference power due to the interference terms ^^α-^^m and ^→m exist.

At the receiver, after despreading and decoding, a fast Fourier transform (FFT), a frequency domain MMSE equalizer, and an inverse FFT (IFFT) can be used to recover the message from the mth user. When quasi-synchronous reception is considered, an iterative successive interference cancellation can be employed before FFT (the processing of iterative successive interference cancellation will be detailed later). The signal before FFT is denoted ^WmW. The estimated Λh transmitted block of the mth user is given by

s_rn(i) = F»G_mFw_m(i) where F is the FFT matrix, and ^m is the frequency domain equalizer for the mth user. The equalized time domain signals are then detected according to the log-likelihood criterion, given by

M*⁾ = ^^g S pmⁱWⁿ IM*⁾ - £ll² (11) where ^¹W is the detected Λh block of symbols for the mth user, "^*" represents the 1-2 norm operation, and £ is a column vector with each entry of which belongs to a set ^§ containing the normalized constellation symbols for a given modulation. For example,

for QPSK modulation.

The detected symbols are then demapped, deinterleaved, and decoded to recover the transmitted bits of the desired user.

Iterative Successive Interference Cancellation:

The proposed receiver of the second class of embodiments of the invention iteratively employs successive interference cancellation in a block-wise manner.

First iteration

In the first iteration, the signals are detected according to the times of arrival of the signals from different users. For example, when the users are ordered and indexed according to their times of arrival, where the first user represents the user whose signals arrive the earliest, the receiver detects the signals of the first user first.

When the first user is considered, there is no interference from users whose signals arrive earlier (cf. ( 6) where m=1), ^Zl^ only consists signals of the first user plus noise term and a small amount of interference from users whose signals arrive later (the later M-1 users). Neglecting the interference from users whose signals arrive later for now,

where the notations ^Sm '*' and ^Wm W denote the signals of the mth user after and before equalization in the qrth iteration. The detected transmitted signal of the first user in the first iteration, denoted as ^{1 κ }}\ are obtained by using (6), where ^smW is

*(l) / -\

substituted by ^si W.

After the signals from the first user are detected, the base station moves on to detect the signals of the second user. When the mth user is considered, the signals of the first m-1 users in the first iteration have been obtained. These signals are used to reconstruct m-1 interference terms by using (8), where ^XbW j_S replaced by *& (*)-, which is the spread and precoded signal of ^s& W= i.e., *& (*) ⁼ (^Cfe ® ^ΛfcK (*).

The reconstructed interference terms from users whose signals arrive earlier in the first iteration are denoted as "*→mW. The recovered signals for the mth user in the first

τn-1

w<i»(t) = »_m(t) - χ; ^β2_m(i) iteration before FFT, equalizer, and IFFT are given by &=i

After is obtained, the transmitted symbols for the mth user can be detected by following the same approach as for the first user.

More iterations

When only one iteration is used (as in the first class of embodiments), it is assumed that interference from users whose signals arrive later than the synchronization instant can be neglected. This assumption does not cause large performance degradations for the reference user when the interference power from users whose signals arrive later is much smaller than that from the users whose signals arrive earlier. In some cases when asynchronization is severe, interference accumulated from the signals of the users that arrive later may also cause unreliable detection of the reference user's message. After the first iteration when all the users' signals are detected, these signals can be used to reconstruct all interference affecting a given reference user. The interference from users whose signals arrive later can be reconstructed by using ( 7), where ³^W x^ (i)

(a=m+1, ...,M) is replaced by ^{α v} ', which is the spread and precoded signal of

W from the previous iteration. The reconstructed signals from users whose signals arrive later in the qrth iteration are

1(9-1)

ψα→m j_{ne sum o}f these reconstructed interference from the later users is then subtracted to update the signal before equalization, i.e.,

m-l M

w«>(i)

θ£_m(i) -

ΦΪ<ZS(Ϊ)

b=l o=m+l and the transmitted symbols are detected by updating ^m v > as s%(i) = F^HG_mFw$(i)

followed by a log-likelihood detector.

Receiver structure

A receiver structure of the iterative successive interference cancellation is illustrated in figure 11. In this figure, the notation Z~*^m jn the blocks 325 indicates a delay of τ_m on the input signal. For example, when the receiver detects the signal for the second user, it needs to synchronize to the second user, which experiences a delay of τa- In addition, ^b is used to denote an array with ^-*"* as its mth column, and

*^α is used to denote an array with ^ψα→m as its mth column. Finally, the thick and thin arrows are used to represent data flow in the form of an array or a vector, respectively.

DFT spreading codes and phase-rotation precoders and decoders are employed. The equalizer used in the current embodiment is the conventional linear MMSE equaliser designed for synchronous reception, as opposed to the optimal MMSE equaliser that was previously derived.

The SIC process can be used for an arbitary number of iterations. In a second iteration, the signals can be detected from a different order than in the first iteration. For example, the signals can be detected in a reverse order according to their times of arrival. In addition, the receiver may choose to update the signals of a given user or not in an iteration other than the first iteration.

Pseudocode for detecting the signals using the iterative successive interference cancellation receiver is given below (where Q is the number of iterations):

Given: Δ^_→m and Δ^_,_m VQ, 6, and TO.

Initialization: xj_n (i) = O for m = 1,■•■ , M and Vi.

Iteration: q = 1 to Q

For 6 = 1 to TO— 1

ø<f _m(i)

+ 1)]

End For

For a = m + 1 to M

#!7#(«) = D£ [ΔSL^-¹^ - 1) - Δ2L_meέ_op*$'-¹>(θ]

End For

S&> (I) = F* GmPwJl' (t)

iS?(i) = axgmin_δm{f) ||S_ra(t) - ^||², ^ e S

icSϊ⁾(t) = (cm ® r_m)iδ⁾ (i)

End For

End Iteration

Parallel detection to reduce latency

It is observed that interference on the rth block is caused by blocks / and / +1 from earlier users. This is the same as for previous examples and the order of decoding the blocks is given in figure 7. The first 2 blocks of the earlier user are first decoded, followed by block 1 of user 2. To obtain interference on the second block of the second user, block 3 of the first user must be known. Obtaining the second block of user 2 allows for recovery of the first block of user 3. Each time a block for a later user is decoded, it is necessary for an extra block for the previous user to be decoded. The process continues until an extra block of the first user is decoded.

Though the second technique involves more control on the decoding order, it allows fairness in decoding the data among all the users. There are two possible implementation techniques for the array decoding. The first approach is to use a lookup table to determine the order of decoding. Alternately, blocks of data from all users can be decoded in an order determined by the following algorithm. It is observed that for decoding the Λh block of user m, it is necessary that block i+b-m of user b has been decoded, where b < m. After decoding the rth block on one user, the receiver checks if the condition mentioned is met for the following user. If not, the block of earlier users needs to be decoded.

The model assumes that there are M_a active users, each of which has transmitted a total of T blocks, and t denotes the time index. At time t, the Λh block of the mth user is decoded, denoted as i(t) and m(t), respectively. For example, i(3) =3 and m(3) =1 means the third block of the first user is decoded at t=3. The pseudocode for the order of decoding is given below.

Initialization: a = 1 , /(0) =1 and m(0) =1 , /=0

fort =1 to Ma * 7-1

o if /(t-1) ==T

a + =1 ;

o end if

o if /(t-1) ==1 or m(t-1) =/Wa;

/^■(t) = /(t-1)+m(t-1)-a+1;

m(t) = a;

o else

m(t) = (m(t-1) % Ma) +1;

/(t) = m(t-1) + /(t-1) - m(t);

o end if

end for

The description of this embodiment now follows with an example involving a system with 8 users. The following 3 asynchronous scenarios should be considered: 1) signals from all users have the same times of arrival except for one, who has a delay of 7, Le., r = I⁰' ⁰' ⁰' ⁰' ⁰' ⁰' ⁰' ⁷I;

2) signals from all users except for the first user have one chip delay relative to the previous user, i.e., ^{r =} t⁰' *' ²' ³' ^{4> 5}' ⁶' ⁷I

3) signals from the last 7 users have the same delay of 7, i.e.,

T = [0,7, 7, 7, 7, 7, 7, 7]

For the 3 scenarios, a channel length of 9 and a CP length of 8 are considered. Therefore, the first scenario considers the worst case scenario for the 8^th user because it suffers from interference from the first 7 users with a delay close to the channel memory order. Similarly, the third scenario considers the worst case scenario for the first user, where the interference comes from the remaining 7 users with a relative delay close to the channel memory order.

The bit error rate (BER) performances of the systems considering different synchronization scenarios are given in Figures 12-14. In all the simulations, a Raleigh fading channel was used, with exponential decaying power delay profile. A BS-CDMA system was used, where the DFT spreading/despreading codes and the phase- ramping precoder and decoding matrices are used. QPSK was considered in each figure, using a block length of P- 16 for each user. As a benchmark, the performance of a synchronous BS-CDMA system is also plotted.

It is observed that in each figure, the BER performance of the asynchronous systems is close to that of a synchronous system after only two iterations of successive interference cancellation are employed.

The principle advantage of the receiver employing the iterative SIC method is that the system can achieve performance close to the synchronous reception. In previous implementations, the optimal MMSE equaliser demonstrated error floor reduction in asynchronous reception at the price of high computational complexity where a matrix inversion operation may be involved. This example uses the low complexity conventional MMSE or ZF equalisers designed for synchronous reception while still effectively suppressing the MUI and IBI.

Various modifications will be apparent to those in the art and it is desired to include all such modifications as fall within the scope of the accompanying claims.

Claims

1. A method of cancelling interference in a block spread code division multiple access transmission system comprising a plurality of user terminal transmitters which are configured to (i) apply precoding to each block of symbols by progressively rotating the phase of each block of symbols identically by a user-dependent amount; (ii) spread the precoded blocks in time or frequency using discrete Fourier transform spreading codes; and (iii) apply a guard interval in the form of a cyclic prefix, the method comprising:

(b) processing the signal from a first user terminal;

wherein the signals are processed in the order of arrival at the receiver.

2. A method of cancelling interference in a block spread code division multiple access transmission system comprising a plurality of user terminal transmitters which are configured to (i) apply precoding to each block of symbols by progressively rotating the phase of each block of symbols identically by a user-dependent amount; (ii) spread the precoded blocks in time or frequency using discrete Fourier transform spreading codes; and (iii) apply a guard interval in the form of a cyclic prefix, the method comprising:

(b) processing a first block ni from a first user terminal mi;

(e) using the reconstructed interference to process a third block ni+1, the third block ni+1 being transmitted by a second user terminal mi+1 and being the next block received at the receiver immediately after receipt of the first block ni received from the first user terminal mi; and

(f) repeating steps (c) to (e) for all received blocks, in each case using the reconstructed interference to process subsequently received blocks, wherein for each third block ni+1 no blocks subsequent to the second block (n+1)i from the previous user terminal mi are processed before the third block ni+1 is processed.

3. A method as claimed in claim 2, wherein the order of processing of blocks is given by:

Initialization: a = 1, n(0) =1 and m(0) =1, t=0

for t =1 to Ma * T-1

if n(t-1) ==T

a + =1;

end if

if n(t-1) ==1 or m(t-1) = Ma;

n(t) = n(t-1)+m(t-1)-a+1;

m(t) = a;

else

m(t) = (m(t-1) % Ma) +1 ;

n(t) = m(t-1) + n(t-1) - m(t);

end if

end for

where:

Ma is the number of active user terminals;

T is the number of blocks transmitted by each user terminal;

t is a time index;

a is an index; and

4. A method as claimed in any one of the preceding claims, wherein the blocks are transmitted using a low-complexity single-carrier frequency division multiple access scheme.

5. A method as claimed in any one of claims 1 to 3, wherein the blocks are transmitted using a low complexity orthogonal frequency division multiple access scheme.

6. A method as claimed in any one of the preceding claims, wherein the step of processing a block comprises:

despreading the block;

decoding the despread block; and

equalising the decoded block.

7. A method as claimed in claim 6, wherein the step of processing a block further comprises:

soft bit mapping the equalised block;

de-interleaving the mapped block; and

channel decoding the de-interleaved block.

8. A method as claimed in claim 6 or 7, wherein the equalising step comprises equalising the block with a conventional equaliser used for synchronous reception.

9. A method in accordance with any preceding claim and including repeating the performance of the method for at least one further iteration.

10. A method in accordance with claim 9 wherein, in a second iteration of the method, the order in which blocks are processed is the reverse of the order in which they are processed in the first iteration.

11. A block spread code division multiple access transmission system comprising: a plurality of user terminal transmitters which are configured to:

(iii) apply a guard interval in the form of a cyclic prefix; and

(a) process the signal from a first user terminal; (b) reconstruct the interference caused by the signal transmitted by the first user terminal;

(c) use the reconstructed interference to process signals received from the other user terminals; and

(d) repeat steps (a) to (c) for the signals received from the other user terminals, wherein the receiver is configured to process received signals in the order of arrival at the receiver.

12. A block spread code division multiple access transmission system comprising: a plurality of user terminal transmitters which are configured to:

(ii) spread the preceded blocks in time or frequency using discrete Fourier transform spreading codes; and

(iii) apply a guard interval in the form of a cyclic prefix; and

(b) process a first block ni from a first user terminal mi;

13. A system as claimed in claim 12, wherein the order of processing of blocks is given by:

Initialization: a = 1 , n(0) =1 and m(0) =1 , t=0 for t =1 to Ma ^* T-1

if n(t-1) ==T

a + =1;

end if

if n(t-1) ==1 or m(t-1) = Ma;

n(t) = n(t-1)+m(t-1)-a+1;

m(t) = a;

else

m(t) = (m(t-1) % Ma) +1;

n(t) = m(t-1) + n(t-1) - m(t);

end if

end for

where:

Ma is the number of active user terminals;

T is the number of blocks transmitted by each user terminal;

t is a time index;

a is an index; and

14. A receiver for cancelling interference in block spread code division multiple access transmission system signals transmitted by a plurality of user terminal transmitters, the signals having been transmitted by: (i) applying precoding to each block of symbols by progressively rotating the phase of each block of symbols identically by a user-dependent amount; (ii) spreading the precoded blocks in time or frequency using discrete Fourier transform spreading codes; and (iii) applying a guard interval in the form of a cyclic prefix; the receiver being configured to:

(b) process the signal from a first user terminal;

(e) repeat steps (b) to (d) for the signals received from the other user terminals, wherein the receiver is configured to process received signals in the order of arrival at the receiver.

15. A receiver for cancelling interference in block spread code division multiple access transmission system signals transmitted by a plurality of user terminal transmitters, the signals having been transmitted by: (i) applying precoding to each block of symbols by progressively rotating the phase of each block of symbols identically by a user-dependent amount; (ii) spreading the precoded blocks in time or frequency using discrete Fourier transform spreading codes; and (iii) applying a guard interval in the form of a cyclic prefix; the receiver being configured to:

(b) process a first block ni from a first user terminal mi;

(f) repeat steps (c) to (e) for all received blocks, in each case using the reconstructed interference to process subsequently received blocks,

wherein for each third block ni+1 no blocks subsequent to the second block (n+1)i from the previous user terminal mi are processed before the third block ni+1 is processed.

16. A receiver as claimed in claim 15, wherein the order of processing of blocks is given by:

Initialization: a = Vn(O) =1 and m(0) =1 , t=0

for t =1 to Ma * T-1

if n(t-1) ==T

a + =1;

end if

if n(t-1) ==1 or m(t-1) = Ma;

n(t) = n(t-1)+m(t-1)-a+1; m(t) = a;

else

m(t) = (m(t-1) % Ma) +1;

n(t) = m(t-1) + n(t-1) - m(t);

end if

end for

where:

Ma is the number of active user terminals;

T is the number of blocks transmitted by each user terminal;

t is a time index;

a is an index; and

17. A receiver in accordance with any one of claims 11 to 16 and operable to iterate operations as stated for the cancellation of interference.

18. A receiver in accordance with claim 17 and operable to perform said operations on said blocks in a second iteration in a reverse order from that in which said operations are performed in a first iteration.

19. A carrier medium carrying computer readable code for controlling a microprocessor to carry out the method of any one of claims 1 to 8.