WO2005032022A1 - Method and system for multipath interference cancellation for a receiver in a multicode division multiple access system - Google Patents

Abstract

A multipath interference cancellation method in a multicode code division multiple access (cdma) system comprising a transmitter and a receiver is disclosed. The method is for determining, at the receiver, a set of data symbols of a multicode signal transmitted from the transmitter. The transmitted multicode signal propagates along multiple paths as respective multipath components. These multipath components are received by the receiver as a composite signal from which the set of data symbols of the multicode signal is obtained. The method includes estimating each multipath component, path by path, based on a modified composite signal. The modified composite signal is the result of subtracting a sum of previously estimated multipath components, if any, from the composite signal. The method further includes estimating the set of data symbols of the multicode signal based on the estimated multipath components. A receiver wherein the method is implemented is also disclosed.

Description

Method and System for Multipath Interference Cancellation for a Receiver in a Multicode Code Division Multiple Access System

Background of the Invention

Code Division Multiple Access (CDMA) is a multiple access communication technique that enables multiple users, such as mobile phone users, to share a common communication channel, i.e., the same frequency band.

In order to share such a common communication channel, each user in a conventional CDMA system is assigned a unique spreading code, such as a pseudorandom noise (PN) code or the product of a common PN code and an individual orthogonal sequence (Walsh code) . This unique spreading code is a high bandwidth signal which is modulated by the data signal to be . transmitted for one user.

The signal of a particular user together with the signals of other users are transmitted from a transmitter along the common communication channel to a receiver. The receiver is able to extract the transmitted signal of the particular user by using a unique spreading code for that particular user. Accordingly, users of a CDMA system share the same frequency bandwidth and time slots of the communication channel, and are distinguishable by the different spreading codes that are to be modulated by each user data.

CDMA has become very popular in commercial applications due to its inherent flexibility to achieve a higher capacity when compared to other multiple access techniques like FDMA (Frequency Division Multiple Access) and TDMA (Time Division Multiple Access) . Several commercial mobile systems such as IS-95, WCDMA (Wideband CDMA) and LCR TDD (Low Chip Rate Time Division Duplex) CDMA use CDMA as a multiple access technique.

Mutlicode CDMA has been introduced as a new transmission scheme for higher and more flexible data rate communication over wireless channels compared to the conventional CDMA transmission scheme, and has also been adopted in WCDMA, cdma2000 and LCR TDD CDMA systems. Unlike in the conventional CDMA transmission scheme wherein data to be transmitted for a user is used to modulate a single spreading code, data to be transmitted for a user using the multicode CDMA transmission scheme is used to modulate multiple spreading codes that are orthogonal to each other. In the multicode CDMA transmission scheme, data for a user is split into a number of data symbol streams or channels.' Each data symbol stream is made up of data symbols. These data symbol streams are then used to modulate respective parallel orthogonal channel (or spreading) codes'. . A data rate of up to 2 Mbps can be supported by such a multicode CDMA transmission scheme.

For example, in a LCR TDD CDMA system, a user subscribing to a service of 144 kbps information data rate is allocated 8 spreading codes in one time slot, and another user subscribing to a service of 384 kbps information data rate is allocated 10 spreading codes in one time slot. Accordingly, each user in the multicode CDMA transmission scheme is allocated more than one, for example K, orthogonal spreading codes. Specifically, each data symbol stream of the user is allocated one spreading code. The data symbol streams are used to modulate each orthogonal spreading code to define respective mutually orthogonal chip streams. Each data symbol stream has symbol duration T, and each spreading code has chip duration T_c. The spreading factor, SF, of the spreading code with respect to the data symbol stream is given by:

T SF = ~ T.

A superposition of the mutually orthogonal chip streams forms a multicode signal for the user. The multicode signal, which is a baseband signal, is used to modulate a carrier signal to result in a radio frequency signal that is transmitted. For ease of description, this transmitted radio frequency signal is referred to as a modulated multicode signal. An example of the transmission process.:is illustrated in' [1] .

However, a major problem of the multicode CDMA system is the receiving of the modulated multicode signal from different transmission paths. The modulated multicode signal propagates along the different transmission paths, with possibly different transmission delays, as respective transmitted signals to reach the receiver. The received radio frequency signal is the superposition of transmitted signals from the different transmission paths. The received radio frequency signal is demodulated from radio frequency to baseband frequency in the receiver to form a baseband composite signal. The baseband composite signal comprises respective multipath components which correspond to the different paths in which the modulated multicode signal propagates .

The chip streams within a first multipath component corresponding to a first transmitted signal reach the receiver at the same time, and hence, are synchronized with one another. Therefore, the orthogonality of the chip streams of the first multipath component are maintained and the chip streams do not interference with one another. The modulated multicode signal of the same user that propagates along a different path to reach the receiver as a second transmitted signal may be received by the receiver at a different (delayed) time. A second multipath component corresponding to the second transmitted signal is not synchronized with the first multipath component.' As a result, the first and second multipath components become interference to each other. Such interference is referred ' to as multipath interference.

Such multipath interference caused by the multipath components needs to be removed in order to recover the data symbols for a user accurately. Certain characteristics of the intra-user multipath components may be exploited for improving the performance of the receiver in a multipath communication channel. Such characteristics include (a) the chip streams of a multicode signal that propagates along a single path (as modulated multicode signal) to the receiver are synchronized and thus orthogonal; (b) the chip streams are propagated along the same multipath communication channel, i.e. each chip stream of the transmitted multicode signal travels along the same plurality of paths to reach the receiver; and *

(c) the spreading codes of all the chip streams of the user are known in both uplink (from a mobile station to a base station) and downlink (from a base station to a mobile station) .

A method for removing multipath interference in a multicode CDMA system is disclosed in [1] . A receiver of [1] uses a rake receiver having parallel rake fingers for estimating the multipath components of a composite signal, with each component corresponding to each path of the multipath channel, for multipath interference cancellation. ^■ Specifically, a vector demodulator is provided in- each rake finger assigned to each path.. In an initial stage, the ' ^• composite signal is demodulated simultaneously, i.e. in parallel in the rake fingers, to estimate data symbols of the multicode signal. The estimated data symbols are used to regenerate or estimate the multipath components corresponding to the multicode signal transmitted in each respective path. In a next stage, the regenerated multipath component of each path is used as multipath interference to be cancelled from the composite signal, and further estimation of the multipath component for each path is regenerated. Such further estimated multipath component for each path may be again cancelled from the composite signal in a third stage, and possibly even further stages, until a desired estimation accuracy is achieved.

The method described according to [1] however is not effective in canceling multipath interference. This is because in the initial stage when estimating each multipath component, the interference caused by the multipath components from the other paths are not taken into account. The initial estimation of the multipath components is, therefore, of a high bit error rate and is thus not accurate. This inaccuracy in the initial multipath component estimation will further affect the accuracy of subsequent estimations of the multipath components, and hence, may even cause the interference to increase after every cancellation stage.

[2] discloses a conventional receiver for canceling multiuser interference successively. The interference contributed by each user. is detected and regenerated, and subsequently subtracted from the received composite signal to: form a resultant signal. The interference;.contributed by a next user is detected and regenerated in a next stage, and is subtracted from the resultant signal, and so forth, until the interference contributed by the last user is detected and regenerated. However, the receiver according to [2] is used in the conventional single-code single user CDMA system, and not in the multicode CDMA system.

Therefore, a more accurate and effective method for canceling multipath interference for a multicode CDMA system is desired. General Description of the Invention

According to the invention, there is provided a multipath interference cancellation method in a multicode code division multiple access (cdma) system comprising a transmitter and a receiver. The method is for determining, at the receiver, a set of data symbols of a multicode signal transmitted from the transmitter. The multicode signal, when translated to a suitable frequency, for example a radio frequency, propagates along multiple paths, for example L paths, in a multipath communication channel as respective transmitted signals. These transmitted signals are received by the receiver and are demodulated to generate a baseband composite signal which comprises respective multipath components. The set of data symbols of the multicode signal is obtained from the baseband composite signal.

The method includes, in a first stage, estimating each multipath component, path by path, based on a modified composite signal. The modified composite signal is the <■ result of subtracting a sum of previously estimated multipath components, if any, from the composite signal. The method further includes, in a second stage, estimating the set of data symbols of the transmitted multicode signal based on the estimated multipath components.

It should be understood by any person skilled in the art that the multicode signal is a baseband frequency signal. Therefore, in order for the multicode signal to be transmitted from the transmitter to the receiver via a multipath communication channel, the baseband frequency multicode signal needs to be translated into a radio frequency signal by a suitable modulation technique, such as amplitude modulation. Similarly, received radio frequency signal which is the superposition of several transmitted signals corresponding to the propagation of the modulated multicode signal along the multiple paths need to be demodulated by the receiver into a baseband frequency signal as a baseband composite signal in order to be further processed. This baseband composite signal comprises multipath components corresponding to the propagation of the modulated multicode signal in different paths.

For ease of description, the invention shall be described based on baseband frequency signals. In other words, when in the following description, to the expression the transmission of a baseband frequency signal is referred to, this should be understood as also to include the modulation of the baseband frequency signal to a radio frequency signal in the transmitter side and the corresponding demodulation of the received radio frequency signal to the baseband frequency signal in the receiver side.

The multipath components are estimated, in the first stage, in a successive manner. In other words, a first multipath component that propagates along a first path of the multipath channel is first estimated based on the composite signal. This first estimated multipath component is then subtracted from the composite signal before a second multipath component that propagates along a second path is estimated. The process continues until all the L multipath components corresponding to the L paths are estimated.

The successive multipath interference cancellation method according to the invention, detects the multipath component corresponding to each path, and subtracts the estimated component of each path from the composite signal as the instantaneous power of the multipath component that propagates along each path is different. Thus, the successive method of the invention is able to produce good results .

Accordingly, the method according to the invention is able to provide a very good initial data estimation in the multicode cdma system compared to existing techniques. The second stage of the invention will further improve the estimation of the set of data symbols of the multicode : signal _'based-- on the estimated multipath components. The set of data symbols are representative of the transmitted data of the user.

The multipath interference cancellation method according to the invention may further include re-estimating each previously estimated multipath component when estimating the multipath component that propagates along a currently considered path. In such a case, each of the previously estimated multipath components, which is substracted from the composite signal to result in the modified composite signal, is taken to be the most recently estimated multipath component for the respective path. The re-estimating of each previously estimated multipath component results in an even more accurate estimation of the multipath component due to multipath diversity. In other words, all information relating to the previous paths are used for re-estimating each previously estimated multipath components .

The multipath components are preferably estimated, path by path, according to a decreasing order of path gains associated with the plurality of paths.

This sequence of decreasing order of path gains, or the strength of the path, has the advantage of higher accuracy in the estimation of each multipath component. This is due to the fact that the multipath component of the path with the highest strength or power results in the highest interference to other paths. Therefore, when the component associated with the highest interference is detected and subtracted first, the subsequent modified composite signal has significantly lower interference. This results in the subsequent estimation of the components to have less error, and hence, higher accuracy.

However, it should be noted that in other embodiments, the multipath components may be estimated in any other order of the path gain or power.

Estimating each multipath component in the first stage of the method includes demodulating the modified composite signal based on at least one characteristic feature of the respective path along which the multipath component propagates to generate a demodulated signal. Subsequently, an estimated preliminary set of data symbols based on at least the demodulated signal is determined, and the multipath component based on the estimated preliminary set of data symbols is estimated.

It should be pointed out that the estimated preliminary set of data symbols is not the final set of data symbols to be determined by the method according to the invention. This estimated preliminary set of data symbols is only a tentative estimation, which is used for estimating the multipath components . The estimated multipath components are subsequently used for determining the final set of data symbols in the second stage of the method according to the invention.

Estimating each multipath component may further include combining the demodulated signal with a sum of previously demodulated signals, if any. In such a case, the estimated preliminary set of data symbols is determined based on t e- combined signal. ' ^■ •

The combination of the demodulated signal with the previously demodulated signals allows the estimation of the current multipath component to be based, not only on information associated with the current path, but also the previously demodulated paths. Thus, a higher accuracy of the components can be achieved, as all information of the multicode signal in the different paths of the channel are taken into consideration. In other words, multipath diversity, as previously discussed, is again taken into account.

For example, when a first multipath component is demodulated, a first preliminary set of data symbols is estimated based on the demodulated output, which is a soft output, from a first path only. Based on the first estimated preliminary set of data symbols, the first multipath component is regenerated. Next, a second multipath component from a second path is demodulated. The demodulated output from the second path is combined with the demodulated output from the first path. A second preliminary set of data symbols is estimated based on the combined demodulated output. Since this combined demodulated output is the sum of the demodulated outputs from the first and second paths, it exploits the multipath diversity. Therefore, the second estimated preliminary set of data symbols has higher accuracy than the first estimated preliminary set of data symbols. Based on the second estimated preliminary, set of data symbols, the second multipath component is- .generated, and the first multipath component is regenerated.

It should be pointed out that the demodulated output is a soft output. A soft output is also known as a soft decision, which is contrasted to a hard decision. A hard decision is a decision output of either 0 or 1 (alternatively -1 or 1) based on soft decision. In other words, soft output is an output value based on which a hard decision is made. As mentioned above, the demodulating of the modified composite signal is based on at least one characteristic feature of the respective path along which the multipath components propagate. The at least one characteristic may be determined using so called channel estimation.

Channel estimation may be performed using a midamble signal as disclosed in [6] when the invention is implemented in a multicode CDMA system according to the LCR TDD CDMA specification [6]. In such a case, the midamble signal is time-multiplexed with the multicode signal before the multicode signal is modulated by a carrier signal to a radio frequency signal to be transmitted. The midamble signal is extracted by the receiver by demultiplexing the demodulated baseband signal received by the receiver, and analyzed to determine the at least one characteristic feature of each of the paths in the communication channel .

Channel estimation may also be performed using a code- multiplexed pilot channel as disclosed in [7] when the invention is implemented in a multicode CDMA system according to the WCDMA specification [7] . In this case, a pilot symbol channel is used to modulate an additional spreading code. This additional pilot symbol chip stream is also superimposed with the other chip streams to form the multicode signal. In the receiver, the pilot symbol stream will be extracted from the composite signal by despreading for channel estimation.

In order to further improve the accuracy of the estimation of the multipath components in the first stage of the method according to the invention, the sum of previously estimated multipath components may be weighted by a factor λ_n before being subtracted from the composite signal.

The factor may be defined according to the following:

0 < λ₂ ≤ λ₃ ≤ .. ≤ λ_n ≤ .. < λ_z__j ≤ λ_L ≤ 1

wherein λ_n is the factor for weighting the sum of previously estimated multipath components to be subtracted from the composite signal in n^th Rake finger in the first stage of the method, wherein the n^th Rake finger demodulates the multipath component of the n^th path in the plurality [ ) of paths. The n^th path preferably corresponds to the path in the plurality of paths carrying the n^th strongest multipath component .

The weighting of the components increases the accuracy of estimation of the components. .:,.'^".

As the estimation of the multipath components may still include estimation errors, especially during the estimation of the first few multipath components, the estimated components is weighted by the factor λ_n to reduce the effect these estimation errors have on subsequent estimation of multipath components. The weighting factor should preferably be low when estimating the first component, and be high (e.g. 1) when estimating the last component. This is because the estimation errors when estimating the first component is higher than the estimation errors when estimating the last component.

From the above described first stage of the method according to the invention, the multipath components are estimated. In the second stage of the method according to the invention, the set of data symbols of the multicode signal is estimated based on the estimated multipath components from the first stage.

The second stage of the method according to the invention may include, for each path, subtracting from the composite signal a sum of estimated multipath components of the other paths and demodulating the subtracted composite signal based on at least one characteristic feature of the path. After demodulating the subtracted composite signal, the demodulated signals of all the paths are combined and the set of data symbols of the multicode signal is estimated based on the combined demodulated signal. ^•

The sum of estimated multipath components of the other :.paths may be weighted by a factor β before being subtracted from the composite signal. The factor β should preferably has a value equal to or greater than λ_L and less than or equal to 1.

As mentioned above, the weighting of the other estimated components before they are subtracted from the composite signal before demodulating a multipath component by the rake finger improves the estimation accuracy of the set of data symbols of the multicode signal. Also, since this is performed after the multipath components of all L paths are estimated, the weighting factor at this stage should be higher than the weighting factor used when estimating the multipath component of the L^th path, λ_L .

In a further embodiment of the invention, the estimating of the set of data symbols based on the estimated multipath components may be repeated a predetermined number of times to increase the accuracy of the estimation of the set of data symbols. In this case, the estimated set of data symbols is used to regenerate the multipath components, and a sum of the re-generated multipath components (from other paths) are subtracted from the composite signal in each rake finger assinged to each path as described above. The subtracted composite signals in each path are again demodulated and combined, and a further estimated set of data symbols is determined.

The sum of regenerated multipath components are preferably weighted by a factor of β_r before being subtracted from the composite signal. The factor β_r should preferably have a value greater than the factor β .

According to another embodiment of the invention, there is provided a receiver having a first estimation module and a second estimation module that together implements the above- described method. Accordingly, the first estimation module is adapted to estimate each multipath component, path by path, based on a modified composite signal. The modified composite signal is the result of subtracting a sum of previously estimated multipath components, if any, from the composite signal. The second estimation module is adapted to estimate at least a set of data symbols of the multicode signal based on the estimated multipath components.

The first estimation module preferably performs the path by path estimation of each multipath component according to a decreasing order of path gains associated with the plurality of paths. The advantage of following the decreasing order of path gains for the path by path estimation is the higher estimation accuracy compared with any other path order, as already described above.

The first estimation module may include multiple estimation sub-modules, each of which corresponds to one of the paths. Each estimation sub-module may include a demodulation unit adapted to demodulate the modified composite signal based on at least one characteristic feature of the respective path to generate a demodulated signal. The estimation sub-module may also include a decision unit adapted to determine an estimated preliminary set of data symbols based on at least the demodulated signal and an estimation unit adapted to estimate a multipath component that propagates along the respective path based on the estimated preliminary set of data symbols.

The estimation sub-module may further include a combining unit adapted to combine the demodulated signal with a sum of previously demodulated signals, if any, to generate a combined signal. In such a case, the decision unit is adapted to determine the estimated preliminary set of data symbols based on the combined signal.

The second estimation module may include multiple demodulation sub-modules, each of which corresponds to one of the plurality of paths. Each demodulation sub-module may include a subtraction unit adapted to subtract from the composite signal a sum of estimated multipath components of the other paths and a demodulation unit adapted to demodulate the subtracted composite signal based on at least one characteristic feature of the path. The second estimation module may further include a combining unit adapted to combine the demodulated signal from each path and a decision unit adapted to estimate the set of data symbols of the multicode signal based on the combined demodulated signal.

Brief Description of. the Figures Figure 1. is a schematic drawing showing a modulated multicode signal from a base station propagating along 3 separate paths, as three respective multipath components, to -reach a receiver in a mobile station. Figure 2 is a schematic drawing showing a transmitter structure in the base station and a receiver front end structure in the mobile station in Fig.l that uses a midamble signal for channel estimation. Figure 3 is a schematic drawing showing a transmitter structure in the base station and a receiver front end structure in the mobile station in Fig.l that uses pilot symbols for channel estimation.

Figure 4 is a schematic drawing showing a composite signal r(t) constituted by the multipath components in Fig.l.

Figure 5 is a block diagram of a 2-part multipath interference cancellation system having an initial data estimation module and a parallel multipath interference cancellation module according to an embodiment of the invention, wherein the system is shown coupled to a channel estimation module.

Figure 6 is a detailed block diagram of the initial data estimation module of Fig.5, showing multiple rake fingers.

Figure 7 is a detailed block diagram of a rake finger as shown in Fig.6.

Figure 8 is a detailed block diagram of the parallel multipath interference cancellation module as shown in Fig.5 according to an embodiment of the invention.

Figure 9 is a graphical representation of the Bit Error Rate (BER) performance of an example for estimating the set of data symbols of multicode signal propagated along a 3-path rayleigh fading channel using a method as implemented by the system of Fig.5 and a method according to the state of the art. Figure 10 is a graphical representation of the Bit Error Rate (BER) performance of an example for estimating set of data symbols of the multicode signal propagated along a 4- path rayleigh fading channel using a method as implemented by the system of Fig.5 and a method according to the state of the art.

Figure 11 is a graphical representation of the Bit Error Rate (BER) performance of an example of the initial data estimation and the parallel interference cancellation stage for estimating set of data symbols of the multicode signal propagated along a 3-path rayleigh fading channel using a method as implemented by the system of Fig.5.

Figure 12 is a graphical representation of the Bit Error

Rate (BER) performance of an example for estimating the set of data symbols of multicode signal propagated along a 3- path rayleigh fading channel using the method as implemented by the system of Fig.5 without a weighting factor and the method as implemented by the system of Fig.5 with a weighting factor. ■ . ■^■ •

Detailed Description of the Invention

Figure 1 shows an illustrative multicode CDMA communication system 100. The communication system 100 includes a base station 101 and a mobile station 102. Although only a single mobile station 102 is illustrated in Fig.l, it should be understood that the communication system 100 may include more than one mobile station. The mobile station 102 communicates with the base station 101 via signals transmitted over a multipath fading wireless communication channel .

The invention henceforth is described in the context of a downlink transmission according to the LCR TDD CDMA communication system 100. However, it should be noted that the invention may also be used in other communication systems like Wideband-CDMA and cdma2000. The invention can also be implemented in an uplink transmission system.

Figure 2 shows a transmitter structure in the base station 101 and a receiver front end structure in the mobile station 102.

In the multicode CDMA system 100, data bits 130 for a single user is split into multiple data bit streams 131 for parallel transmission. Data modulation, such as Quadrant- Phase Shift Key (QPSK) [8], is performed for each data bit stream 131 to generate data symbol streams 132 comprising data symbols. The data symbols are complex values and have duration T. Each data symbol stream 132 is used to modulate a spreading code 133 of a group of orthogonal spreading codes allocated to the user. After modulation of the spreading codes, the data symbol streams 132 are transformed or spread into chip streams 134. All the chip streams 134 are superimposed to form a multicode signal 135.

The multicode signal 135 is a baseband frequency signal, and hence, is not suitable to be transmitted over the communication channel. Therefore, the baseband multicode signal 135 needs to be translated to a radio frequency signal using modulation techniques, such as amplitude modulation. In the modulation process, the baseband multicode signal 135 is used to modulate a carrier signal, resulting in the radio frequency signal.

Before modulating the carrier signal, the multicode signal 135 may be multiplexed with a midamble signal 136 (when the invention is implemented in a multicode CDMA system according to the LCR TDD CDMA specification) by a multiplexer 137 to form a multiplexed signal 138. The midamble signal 136 is a special chip sequence used for channel estimation for the paths in the communication channel. The chip sequence of the midamble signal 136, is known by both the base station 101 and the mobile station 102.

The complex multiplexed signal 138 is split into its respective real part 139 and imaginary part 140. The real part 139 and the imaginary part 140 are each shaped by a pulse shaping filter 141,142 to limit the bandwidth of the real and imaginary parts 139,140 of the multiplexed signal 138, resulting in corresponding bandwidth limited signals 143,144. The bandwidth limited signals 143,144 are for modulating the respective carrier signals 145,146, and the corresponding modulated signals are subsequently combined to form a modulated multicode signal 147. The modulated multicode signal 147 is transmitted from the base station 101 to the mobile station 102 along the multipath communication channel. The carrier signal 145 which is being modulated by the real part bandwidth limited signal 143 is cos(ω_ct) and the carrier signal 146 which is being modulated by the imaginary part bandwidth limited signal 144 is -sin(ω_ct), wherein ω_c is the carrier frequency of the carrier signals 145,146, and t is the time variation.

The orthogonal spreading codes 133 used for spreading the data symbol streams 132 may be generated, for example, by multiplying Walsh sequences with common PN sequences.

When the invention is implemented in a multicde CDMA system according to the WCDMA specification, a known pilot symbol 150 is used for channel estimation for the paths in the communication channel instead of the midamble signal 136.

The known pilot symbol 150 is used to modulate an additional spreading code 151 to form a pilot symbol chip stream 152 as illustrated in Fig.3. The additional spreading code 151 may have a different spreading factor from the other spreading codes 133. However, the additional spreading code 151 should also be orthogonal to the other spreading codes 133. The piiot symbol chip stream 152 is superimposed with the other chip streams 134 to form the multicode signal 135. In this case, there is no multiplexing of the midamble signal 136 with the multicode signal 135. The multicode signal 135 is split directly into its respective real part 139 and imaginary part 140.

It should be noted that when the data bit streams 131 are modulated using Binary-Phase Shift Key (BPSK) , the imaginary part 140 of the multicode signal 135 is zero. Therefore, the portion of the transmitter as shown in Fig.2 and Fig.3 for processing the imaginary part 140 may be omitted. Accordingly, the midamble signal 136 used for channel estimation shall also be a real value.

In the multipath communication channel, the modulated multicode signal 147 does not arrive at the mobile station 102 along just a single path but along multiple paths, for example along the 3 paths as shown in Fig.l. The modulated multicode signal 147 may propagate, as a first transmitted signal, along a first path 106 to the mobile station 102. The modulated mutlicode signal 147 may propagate, as a second transmitted signal, along a second path 105. As shown in Fig. 1, this second path 105 includes a reflection by a building 103. Similarly, the modulated multicode signal may propagate, as a third transmitted signal, along a third path 107. As shown in Fig. 1, this third path 107 includes a reflection by another building 104.

All the transmitted signals are received by the mobile station 102 as a. received radio frequency signal 160. The received radio frequency signal is then demodulated using a ' carrier signal 161 to extract a demodulated baseband frequency signal 162. The carrier signal 161 used for demodulating the received radio frequency signal 160 can be expressed as e^~}a>c' , with the carrier frequency ω_c having the same value as the carrier frequency used for modulating the baseband multicode signal 135. The demodulated baseband frequency signal 162 is passed through a low pass filter 163, and demultiplexed by a demultiplexer 164 into a composite signal 165 and a demultiplexed midamble signal 166, when the invention is implemented in the multicde CDMA system according to the LCR LDD CDMA specification. When the invention is implemented in the multicde CDMA system according to the WCDMA specification, the output of the low pass filter 163 is the composite signal 165 (see Fig.3) .

It should be noted that although only the downlink transmission (from the base station 101 to the mobile station 102) is illustrated in Fig.l, the modulated multicode signal 147 may also be transmitted in an uplink transmission (from the mobile station 102 to the base station 101) . In this case, the base station 101 receives the transmitted signals corresponding to the modulated multicode signal 147 arriving at the base station 101 along different paths of the multipath communication channel.

The structure of the composite signal 165 is shown in Fig.4. Accordingly, the composite signal 165 is constituted by the multipath components 111,112,113. Each multipath component 111,112,113 corresponds to the modulated multicode, signal 147' which propagates along each path of the multipath communication channel, i.e. the respective transmitted signals. Each multipath component 111, 112, 113' is the same information-bearing signal representative of the multicode signal 135 transmitted from the base station 101 to the mobile station 102. This information bearing signal is formed by superposition of the transmitted chip streams 114.

The transmitted chip streams 114 belonging to the same multipath component 111, 112, 113 are synchronized, and hence do not interfere with one another as their orthogonal properties are maintained. However, the transmitted chip streams 114 of the first multipath component 111 may not be synchronized with those of the second and the third multipath component 112, 113. The transmitted chip streams 114 of different multipath components 111, 112, 113 are thus no longer orthogonal with respect to one another. This non- orthogonality between transmitted chip streams 114 of different multipath components 111, 112, 113 results in interference between the multipath components 111, 112, 113. Such interference between the multipath components 111,112,113 is known as multipath interference.

Figure 5 shows a block diagram of a hardware or software system that implements a method according to an embodiment of the invention for multipath interference cancellation in the multicode CDMA communication system 100. The method and thus the system determines the estimated set of data symbols of the multicode signal from the composite signal r(t). The system includes two modules: an initial data estimation module 121 that .implements an initial data estimation stage of the method, and a parallel multipath interference cancellation module 122 that implements a parallel multipath* interference cancellation stage of the method. The initial data estimation stage and parallel multipath interference cancellation stage will be described in detail later.

The multicode signal 135 is represented as:

ira wherein x (t) is the multicode signal,

K is the number of data symbol streams of the multicode signal,

C_k (t) is the k^th spreading code, and b_k (t) is the k^th data symbol stream for modulating the k* spreading code.

b_k (t) can be expressed as :

wherein d_k(m) is the m^tb data symbol of k^th data symbol stream for modulating the k^th spreading code, N is number of symbols in k^th data stream b_k (t) , and g_τ(t) is a rectangular waveform of duration T: g_τ(t)=l , for 0 ≤ t <T, g_τ(t) = 0, otherwise.

The baseband composite signal 165, r (t) , received by the mobile station 102 is represented as:

r(t) ^α_n(t)e^{iΦ ,)}x(t -τ_n) ₊ n(t) (3)

wherein r (t) is the composite signal, _n(t)e'^φ"^(t) is the path coefficient of path n, _n(t) is the rayleigh distribution amplitude of path n, φn (t) is the associated phase of path n, τ_n is the path delay of path n, n (t) is the additive background noise of the multipath communication channel, and I is the number of resolvable paths in the multipath communication channel .

The composite signal, (t) , is provided as input to the initial data estimation module 121 and the parallel multipath interference cancellation module 122. The output of the initial data estimation module 121 is also provided as input to the parallel multipath interference cancellation module 122, which in turn generates an output 123 which is the estimated set of data symbols. This set of data symbols is representative of the transmitted data of the user.

The paths of the multipath communication channel are assumed to be slow fading with parameters that are invariant over at least one bit period.

Channel characteristics such as the path coefficient o _n(t)e^jΦΛ,) , path delay τ_n and path strength |α„(t) |² of all the paths in the multipath communication channel are determined using a channel estimation module 120 based on the received baseband midamble signal 166, m(t) . Channel estimation may also be performed using any known existing technique such as those disclosed in [3] , [4] and [5] .

It should be noted that when the invention is implemented in the multicode CDMA system according to the WCDMA specification, the composite signal, r (t) , is also provided as an input to the channel estimation module 120 for obtaining the characteristic features of the paths in the communication channel. This is because pilot symbol chip stream 152 used for channel estimation is part of the composite signal, r (t) . It should be highlighted in this case that the pilot symbol and the corresponding spreading code are known by both the transmitter and the receiver, and hence, need not be estimated like in the case for the data symbols. Therefore, the portion of the multipath components contributed by the pilot symbol chip stream can be regenerated directly by the receiver and subtracted from the composite signal before the estimation of the data symbols of mulitcode sginal.

The detailed implementation of the initial data estimation module 121 is shown in Figure 6. The initial data estimation module 121 includes multiple estimation sub- modules. Each sub-module includes a rake finger (or demodulator) 200,204, a decision unit 201 and a regenerator unit 202. The structure of the rake finger 200,204 will be described in greater detail later.

A multipath component, which is representative of a corresponding transmitted signal that propagates along a path having a strongest path strength, as determined by the channel estimation module 120, is estimated first by demodulating the composite signal r (t) using the first rake finger 200. For convenience, the multipath component which is to be estimated first is referred as the first multipath component. The output of the first rake finger 200 for data symbol m is a soft demodulated output which can be expressed as :

wherein

Y⁽¹⁾ is the soft output vector of the first rake finger 200, a e^~i^ is the conjugate of path coefficient of the first (strongest) path in the multipath communication channel, c_k(i) is the conjugate of the spreading code for each data symbol stream k,

T_j is the path delay of the first (strongest) path, T is the data symbol duration, and y is a demodulated output of the rake finger 200 corresponding to data symbol stream k.

It should be understood that the invention is used to estimate each data symbol (m) until all the N data symbols of each data symbol stream are estimated.

The demodulated output is a soft output. A soft output is also known as soft decision, which is contrasted with hard decision. Hard decision is a decision output of either 0 or 1 (alternatively -1 or 1) based on soft decision. In other words, soft output is an output value based on which a hard decision is made. Based on the soft output vector Y⁽¹⁾ , a first estimated set of data symbols (corresponding to the m^th data symbol of each data symbol stream) is obtained by a first decision unit 201 using the following expression:

D«-[rf« ^®». d?> - . d$f = signrγ ) (5)

wherein sign(Y^ω) is the sign of Y⁽¹⁾ , d is the first estimated data symbol for data symbol stream k,

K is the number of data symbol streams, and D⁽¹⁾ is the first estimated set of data symbols.

The multicode signal can then be estimated, based on the first estimated set of data symbols obtained from equation (5) , to be:

*^w( -|**^w('to.( ^■ . (6)

bP(t) is estimated data symbol stream for data symbol stream k, based on the first estimated set of data symbol Z)⁽¹⁾ , determined according to equation (2) .

The first multipath component, or the multipath component corresponding to a first transmitted signal which is transmitted along a first path can be estimated since the spreading codes c_k (t) for all data streams of the multicode signal are known to the mobile station 102. The first multipath component is estimated by first delaying the estimated multicode signal x®(t) with the corresponding path delay τ_λ and multiplying the delayed estimated signal ⁽¹⁾(t-T_j) with the path coefficient _te^}φl of the first (strongest) path in a regeneration unit 202. The multipath component of the first path is regenerated as follows:

wherein s^ t is the regenerated multipath component of the first (strongest) path, τ₂ is the path delay of the first (strongest) path, and a_λe^iΦi is the path coefficient of the first (strongest) path.

The regenerated multipath component of the first path s^® t is then subtracted from the composite signal r (t) in a subtraction unit 203 to obtain a first resultant signal r⁽² (t) :

wherein r⁽²⁾(t) is the first resultant signal.

The first resultant signal r⁽²⁾(t) is received and demodulated by a second rake finger 204 in order to estimate a second multipath component which correspond to a second transmitted signal that is transmitted to the receiver along a second path. Preferably, the path along which the second transmitted signal propagates has a second strongest path strength, as determined by the channel estimation module 120.

The soft output vector from the second rake finger 204 is expressed as:

wherein y⁽²⁾ is the soft output of the second rake finger 204, and α₂e^"/fc is the conjugate of path coefficient of the second strongest path, τ₂ is the path delay of the second (strongest) path is a demodulated output of the rake finger 204 corresponding to data symbol stream k.

The demodulated output of the first rake finger 200, Y^(1> , and the demodulated output of the second rake finger 204, y⁽²⁾, are combined in a combining unit 205 to generate a combined output :

wherein Y r('12) ' is the combined output . Based on the combined output Y^<12> , a second estimated (updated) set of data symbols of the multicode signal can again be obtained by a second decision unit 206 using the following expression:

!><*> -[<*«,</«••• d -

(11)

wherein sign(Y^(12}) is the sign of Y^(12> , d is the second estimated data symbols for data symbol stream k, and D^<2) is the second estimated set of data symbols.

The multicode signal is further estimated to be:

K ⁽²⁾( = W^}(' ( (12) wherein bP t) is the estimated data stream for data symbol stream k, based on the second estimated set of data symbol D⁽²⁾ .

By delaying the estimated multicode signal x⁽ '(t) with the path delay τ₂ and multiplying the estimated signal x⁽²⁾(t-τ₂) with the path coefficient a₂e^jΦz of the second strongest path in a second regeneration unit 207, an estimated multipath component along the second strongest path is obtained as:

s (t) = ₂e^}φ>χW(t -τ₂) (13) wherein

5₂ ²⁾(t) is the regenerated multipath component of the second strongest path, τ₂ is the path delay of the second strongest path, and ₂e^lφ2 is the path coefficient of the second strongest path.

Similarly, a re-estimation of the first strongest multipath component is obtained as:

wherein sl²⁾(t is the re-estimated multipath component of the first path regenerated by the second regeneration unit.

Similarly, the regenerated multipath components s^ t and s₂ ²⁾(t) are subtracted from the composite signal r (t) to form a second resultant signal r⁽³⁾(t) for estimating a third multipath component that propagates along a further (third strongest) path of the multipath communication- channel:

r<³>(t)= r(t)-s<²>( - ²⁾( (15)

wherein r⁽³⁾(t) is the second resultant signal.

The above process for obtaining an estimated multipath component is repeated for each further multipath component, until the last multipath component, for example the component that propagates along a path having a weakest signal strength, is estimated. It should be noted that when the multipath component for a current path is estimated, the multipath components for all the previous paths are also re- estimated. The estimated multipath component of the current path and the most recently re-estimated multipath component of each of the previously processed paths are subtracted from the composite signal when estimating the multipath component of the next path.

These estimated multipath components are subsequently fed to the parallel interference cancellation module 122 for canceling multipath interference and determining the estimated set of data symbols of the multicode signal.

Figure 7 shows the block diagram of the first rake finger

200. The composite signal r (t) received by the rake finger 200 is passed into a plurality of despreading units corresponding to each spreading code used in the multicode signal. The composite signal in each despreading unit is multiplied by the respective conjugate of spreading code- in the respective multiplication unit 220, and is subsequently integrated over a data symbol interval by an integration unit 222. The respective integrated output is further multiplied by the conjugate of the respective path coefficient in a multiplication unit 223 to produce the soft demodulated output 7⁽¹⁾= ( y > , ..., y )^τ of the rake finger 200.

The rake fingers for the other paths operate in the same manner as the first rake finger 200, except that the input to the other rake fingers are modified composite signals as described above, and the respective integrated output is multiplied by the conjugate of the respective path coefficient.

It should be highlighted that when determining a multipath component that propagates along a particular path of the multipath communication channel, the multipath components that propagate along the other paths are considered as interference to the multpath component of that particular path.

In the parallel multipath interference cancellation module 122, the input into a rake finger assigned to each particular path is obtained by subtracting from the composite signal all estimated multipath components that propagate along the other paths. A detailed block diagram of the parallel interference cancellation module 122 is shown in Figure 8.

A plurality of rake fingers 231, each assigned to a resolved path in the multipath communication channel, is provided.

Before the composite signal r (t) is received into each of ■ " the rake fingers 231, the interference to each path associated with each rake finger 231 is subtracted from the composite signal in the plurality of subtraction units 230 as follows:

Λ(')- - (- Jl,ι ≠nsrHt) (16)

wherein P_ni*) ^^{s the} iⁿP^{ut to a} rake finger assigned to path n, s '(t) is the regenerated multipath component for path i from the initial data estimation module 121, and is the interference signal to path n .

As mentioned earlier, the interference to each path is the multipath components of the other paths, which are already estimated by the initial data estimation module 121.

The input signal P_n (t) for each path n is demodulated by each of the rake fingers 231, respectively, as follow:

wherein Z_n is the soft output vector of the n t^τhⁿ rake finger, a_ne^~iΦ" is the conjugate of path coefficient for the n^th path, c_t ^*(t) is the conjugate of the spreading code for data symbol stream i, and zⁿ⁾ is demodulated output from n^th path for data symbol stream k, and τ_n is the path delay of the n^th path.

In the combiner 232, all the demodulated outputs corresponding to each data symbol stream are combined:

wherein

V is the combined demodulated output vector .

Based on V, a decision is made by a decision unit 233 to estimate the set of data symbols of the multicode signal which is represented as :

D = [d₁...d_k...d_κf = sign(V) ( 19)

wherein d_k is the estimated data symbol of the data symbol stream

It is possible in other embodiments to repeat the steps in the parallel multipath interference cancellation a plurality of times in order to obtain a more accurate estimate of the set of data symbols of the multicode signal, until a desired accuracy of the estimation is achieved. In such a case, further estimates of the multipath components along each path are regenerated based on D . The regenerated estimated multipath components are used as interference signals to be subtracted from the composite signal in further subtraction units corresponding to each path of the multipath channel. The subtracted composite signals are demodulated and all the demodulated signals are combined. Based on the combined demodulated signals, the set of data symbols are estimated yet again.

In another embodiment of the invention, the estimated multipath component of the current path and the most recently updated estimated multipath components of all previous paths are multiplied by a factor λ_k before they are subtracted from the composite signal r (t) for estimating the multipath component of the next path in the initial data estimation module 121.

The factor λ_n obeys the following expression:

0 < ≤ λ_j ≤ ... ≤ λ_n ≤ ... ≤ λ_L__λ ≤ λ_L ≤ 1 (20)

Therefore, equation (8) is modified as:

and equation (15) is modified as:

Similarly, it is also possible, in the parallel multipath interference cancellation module 122, to multiply the estimated interference signals to each path by a factor β before they are subtracted from the composite signal. Therefore, equation (16) becomes:

wherein β obeys the following expression:

λ_L ≤ β ≤ 1 (22)

When the steps in the parallel multipath interference cancellation are repeated a plurality of times in order to obtain a more accurate estimate of the data symbols of the multicode signal, the regenerated multipath components are preferably weighted by a factor of β_r before being subtracted from the composite signal. The factor β_r should preferably have a value greater than the factor β , or the factor of a previous parallel multipath interference cancellation stage.

Examples

Simulation has been performed to evaluate the performance of estimating the data symbols of multicode signal propagated along a multipath channel using the above-described method.

The method was implemented in a multicode CDMA communication system according to the LCR TDD CDMA specification. The multicode signal for a single user consists of 10 data streams that are used to modulate 10 corresponding spreading codes having equal power. The spreading factor (which is the ratio of the bandwidth of the spreading code to the bandwidth of the data stream) is 16.

The multicode signal is transmitted along two kinds of multipath communication channels: channel A and channel B. Channel A is a 3-path rayleigh fading channel with equal power. The relative path delay is 0, 4 and 15 chips (chip is the reciprocal of transmitted signal bandwidth) , respectively. Channel B is a 4-path rayleigh fading channel with relative power of 0, -3, -6 and -9 dB, respectively. The relative path delay is 0, 1, 2 and 3 chips, respectively. The Doppler spread for both channel A and channel B is 5 Hz. The channel estimation used is the Maximum Likelihood method. In Figure 9, 10, 11 and 12, E_b /N₀ is the Signal to Noise Ratio (SNR) per bit in dB.

The Bit Error Rate (BER) performance of estimating the data symbols of the multicode signal propagated along the 3-path rayleigh fading channel (channel A) using the method according to the invention and the method according to [1] is shown in Figure 9. The corresponding BER performance for estimating the data symbols of multicode signal propagated along the 4-path rayleigh fading channel (channel B) using the method according to the invention and the method according to [1] is shown in Figure 10.

The curves 301,302,304,305 represent the BER using the method of [1] after a first, second, third and fourth stage of interference cancellation. Curves 303,306 represent BER using the above-described method, after a one-stage and two -stage parallel interference cancellation. The one-stage and two-stage parallel interference cancellation refer to the parallel interference cancellation performed by one and two cascaded parallel multipath interference cancellation modules 122, respectively. It should be noted that the initial data estimation has been performed by the initial data estimation module 121 prior to performing parallel interference cancellation by the one or more parallel multipath interference cancellation modules 122.

It can be seen from Figure 9 that for a multipath communication channel having 3 paths of equal power, the method according to the invention implemented using one initial data estimation stage 121 and two parallel interference stage 122 has better BER performance than the 4-stage parallel interference using the method of [1] .

In Fig. 10, the curves 311,312 represent the BER using the method of [1] after a second and third stage of interference cancellation. Curves 313,314 represent BER using the above- described method, after a one-stage and two-stage parallel interference cancellation. '^• >^'

It can be seen from Fig.10 that for a multipath communication channel having 4 paths of unequal power, the method according to the invention implemented using one initial data estimation stage 121 and one parallel interference stage 122 has almost the same performance than a 3-stage parallel interference using the method of [1] . However, it can be seen that the method according to the invention implemented using one initial data estimation stage 121 and two parallel interference stage 122 has much better BER performance than the 3-stage parallel interference using the method of [1] .

It can be seen from both Fig.9 and Fig.10 that the BER of the method according to an embodiment of the invention is substantially lower than that obtained using the method of [1] . This means that the errors of the estimation using the above-described method is lower than [1] . Therefore, the performance of the interference cancellation of the above- described method is better than that of [1] .

Figure 11 shows the BER performance of estimating the set of data symbols of the multicode signal in Channel A using the above-described method. Curve 320 shows the BER performance after demodulating the first strongest path in the initial data estimation model 121. Curve 321 shows the BER performance after the demodulated output from the first strongest path and second strongest path are combined in the ^■ initial data estimation module 121. Curve 322 shows the BER performance after the, demodulated output from first strongest path, second- strongest path and third strongest path are combined in the initial data estimation module 121. Curve 323 shows the BER performance after the parallel interference cancellation module 122. It can be clearly seen that the BER performance using the above-described method is improved by the parallel interference cancellation module 122.

Figure 12 shows the comparison of the BER performance of estimating the set of data symbols of the multicode signal in Channel A using the method described above, with weighting factor and without weighting factor. The weighting factors used are λ₂ = 0.5 , λ₃ = 0.1 , β = 0.8 and βr - ⁰-⁹ "

Curve 330 is the BER performance after combining the demodulated output from the first and second strongest path, and curve 331 is the corresponding BER performance when the estimated interference from the first path is multiplied with a weighting factor λ₂ - 0.5 before it was subtracted from the composite signal to estimate the interference by the second path. Curve 332 is the BER performance after combining demodulated output from the first, second and third strongest path, and curve 333 is the corresponding BER performance when the first and second estimated interference are multiplied by a weighting factor λ₃=0.7 before being subtracted from the composite signal.

Curves 334,336 are the BER performance after a one-stage and two-stage parallel interference cancellation, respectively. Curves.335,337 are the BER performance corresponding to the respective curves 334,335 when each of the estimated interference for subsequent re-estimating the data symbols of the multicode signal is multiplied by a weighting factor (β=0.8, £.=0.9).

From Fig.12, it can be seen that the use of a weighting factor provides an improvement to the accuracy of the estimation of the data symbols of the multicode signal. The following references are cited in this document:

[1] J. Chen, J. Wang and M. Sawahashi, "MCI Cancellation for Multicode Wideband CDMA Systems", IEEE JSAC, vol. 20, No. 2, Feb 2002, pp 450-462.

[2] M. Ammar, T. Chonavel, S. Saoudi, "Multistage SIC Structure for Uplink UNTS Multiuser Receiver Over Multipath Rayleigh Channels", VTC 01, pp 2514-2518.

[3] B. Steiner, P. Jung, " Optimum and suboptimum channel estimation for the Uplink of CDMA Mobile Radio Systems with Joint Detection", European Transaction on Telecommunications and Related Techniques, Vol.5, No. 1, pp.39-50, January- February 1994.

[4] B. Steiner, P. W. Baier, " Low cost channel estimation in the uplink receiver of CDMA Mobile Radio Systems", ^■ Frequenz, vol.47, pp.292-298, Nov/Dec 1993.

[5]-'^';I. Held and A. Kerroum, "TD-SCDMA Mobile Station Receivers: Architecture, Performance, Impact of Channel Estimation", China Wireless Congress, HanZhou, China, 2002.

[6] 3GPP TS25.221 v4.6.0, "Physical Channels and Mapping of Transport Channels onto Physical Channels (TDD) (Release 4) ," Sep. 2002.

[7] 3GPP TS25.211 v4.6.0, "Physical Channels and Mapping of Transport Channels onto Physical Channels (FDD) (Release 4) ," Sep. 2002. [8] J. G. Prokis, "Digital communications", McGraw-Hill, Fourth Edition, 2000.

Claims

What is claimed is:

1. A multipath interference cancellation method for determining at least an estimated set of data symbols of a multicode signal transmitted from a transmitter of a code division multiple access (cdma) system, wherein the system further comprises a receiver, and wherein the transmitted multicode signal propagates along a plurality (L) of paths in a multipath communication channel as respective multipath components that are received by the receiver as a composite signal, the method comprising: estimating each multipath component, path by path, based on a modified composite signal, wherein the modified composite signal is the result of subtracting a sum of previously estimated multipath components, if any, from the composite signal; and estimating the set of data symbols of, the multicode signal based on the estimated multipath components .^'

2. The multipath .interference cancellation method according to claim 1, further comprising .. re-estimating each previously estimated multipath component when estimating the multipath component that propagates along a current path.

3. The multipath interference cancellation method according to claim 2, wherein each of the previously estimated multipath components is the most recently estimated multipath component for the respective path.

4. The multipath interference cancellation method according to any one of claims 1 to 3, wherein the multipath components are estimated, path by path, according to a decreasing order of path gains associated with the plurality of paths.

5. The multipath interference cancellation method according to any one of claims 1 to 4, wherein the estimation of each multipath component comprises: demodulating the modified composite signal based on at least one characteristic feature of the respective path along which the multipath component propagates to generate a demodulated signal; determining a preliminary set of data symbols based on at least the demodulated signal; and estimating the multipath component based on the preliminary set of data symbols.

6. The multipath interference cancellation method according to claim 5, wherein the estimation of each multipath component further comprises combining the demodulated signal with a sum of previously demodulated signals, if any, and wherein the preliminary set of data symbols is determined based on the combined signal.

7. The multipath interference cancellation method according to claims 5 or 6, wherein channel estimation is performed to determine the at least one characteristic feature of each of the plurality of paths in the multipath communication channel .

8. The multipath interference cancellation method according to any one of claims 1 to 7, wherein the sum of previously estimated multipath components is weighted by a factor λ_n before being subtracted from the composite signal.

9. The multipath interference cancellation method according to claim 8, wherein the factor is defined according to the following: < λ₂ ≤ λ_j ≤ „ ≤ λ_n ≤ .. ≤ λ_L__x ≤ λ_L ≤ \

wherein λ_n is the factor for weighting the sum of previously estimated multipath components in (n)^th path in the plurality (L) of paths.

10. The multipath interference cancellation method according to any one of claims 1 to 9, wherein estimating the set of data symbols of the multicode signal comprises: for each path: subtracting from the composite signal a sum of ^• estimated multipath components of the other paths; and demodulating the subtracted composite signal based on at least one characteristic feature of the path; combining the demodulated signals; and estimating the set of data sysmbols of the multicode signal based on the combined demodulated signals.

11. The multipath interference cancellation method according to claim 10, wherein the sum of estimated multipath components of the other paths is weighted by a factor β before being subtracted from the composite signal.

12. The multipath interference cancellation method according to claim 11, wherein the factor β is defined according to the following:

λ_L ≤ β ≤ l

wherein λ_L is the factor for weighting the sum of previously estimated multipath components, if present, when estimating the multipath component in the (L^th) path in the plurality (L) of paths, and β > is the factor for weighting the sum of estimated multipath components of the other paths.

13. A receiver for a multicode code division multiple access (cdma) system for receiving a composite signal from a transmitter of the cdma system, wherein the transmitter transmits a multicode signal that propagates along a plurality (L) of paths in a multipath communication channel as respective multipath components that are received by the receiver as the composite signal, the receiver comprising: a first estimation module being adapted to estimate each multipath component, path by path, based on a modified composite signal, wherein the modified composite signal is the result of subtracting a sum of previously estimated multipath components, if any, from the composite signal; and a second estimation module being adapted to estimate at least a set of data symbols of the multicode signal based on the estimated multipath components.

14. A receiver according to claim 13, wherein the first estimated module estimates each multipath component, path by path, according to a decreasing order of path gains associated with the plurality of paths.

15. A receiver according to claims 13 or 14, wherein the first estimation module comprises a plurality of estimation sub-modules, each or which corresponds to one of the plurality (L) of paths, wherein each estimation sub-module comprises : a demodulation unit being adapted to demodulate the modified composite signal based on at least one characteristic feature of the respective path to generate a demodulated signal; a decision unit being adapted to determine a preliminary set of data symbols based on at least the demodulated signal; and an estimation unit being adapted to estimate a multipath component that propagates along the respective path based on the preliminary set of data symbols.

16. A receiver according to claim 15, wherein the estimation sub-module further comprises: a combining unit being adapted to combine the demodulated signal with a sum of previously demodulated signals, if any, to generate a combined signal, and wherein the decision unit is adapted to determine the preliminary set of data symbols based on the combined signal.

17. A receiver according to any one of claims 13 to 16, wherein the second estimation module comprises: a plurality of demodulation sub-modules, each of which corresponds to one of the plurality of paths, wherein each demodulation sub-module comprises: a subtraction unit being adapted to subtract from the composite signal a sum of estimated multipath components of the other paths; and a demodulation unit being adapted to demodulate the subtracted composite signal based on at least one characteristic feature of the path; a combining unit being adapted to combine the demodulated signals; and an decision unit being adapted to estimate the set of data symbols of the multicode signal based on the combined demodulated signals.