WO2003094378A1

WO2003094378A1 - A method of and an arrangement for reduced real-time signal processing in a tdd-cdma telecommunications system

Info

Publication number: WO2003094378A1
Application number: PCT/EP2002/005058
Authority: WO
Inventors: Khalid Al-Murrani; Cezar Bruma
Original assignee: Telefonaktiebolaget Lm Ericsson (Publ)
Priority date: 2002-05-06
Filing date: 2002-05-06
Publication date: 2003-11-13
Also published as: CN100508413C; AU2002316897A1; CN1625846A

Abstract

There is disclosed a method of and arrangement for signal processing in a digital telecommunications system operating in a Time Division Duplex (TDD)-Code Division Multiple Access (CDMA) mode, wherein a plurality of users simultaneously transmit (11) signals (10) to a receiving unit (12) of a radio access unit. The signals (10) are comprised of a denumerable number of elements, and these signals (10) are simultaneously processed using a signal processing algorithm. Before processing of the signals (10), comprised in the original signal received (11) by the receiving unit (12) of the radio access unit, an assessment is made of elements that are likely to contain relevant information. Signal processing is performed by said elements, thereby reducing complexity and processing time of the signal processing algorithm used.

Description

Title

A method of and an arrangement for reduced real-time signal processing in a TDD-CDMA telecommunications system.

Field of the Invention

The present invention relates generally to signal processing in a telecommunications system, and more specifically, to real-time reduced signal processing in a telecommunications system based on Time Division Duplex (TDD)-Code Division Multiple Access (CDMA).

Background of the Invention

In mobile telecommunications, a number of mobile units simultaneously transmit signals to a single access unit, which access unit connects to a network. In order to distinguish the different signals sent by the multiple mobile units, a number of division-multiple-access- techniques have been introduced in practice. Given the relatively high rate of information exchange in third-generation (3G) mobile radio systems, these techniques are in particular important for 3G systems, although they might be applied in a different setting.

Division-multiple-access-techniques are in fact multiplexing techniques. A number of division-multiple-access-techniques can be distinguished, amongst which are Frequency Division Multiple

Access (FDMA), Code Division Multiple Access (CDMA), and Time Division Multiple Access (TDMA) .

In FDMA multiple signals are transmitted to a radio access unit in different frequency bands. The access unit distinguishes each signal by its frequency band.

In TDMA multiple signals are sent one after the other, and the access unit receives parts of the information of each signal subsequently.

In CDMA the signal of each mobile unit comprises a unique code, a spreading code, which codes, and hence the different signals, can be distinguished by the access unit.

In the above, the signal was assumed to be simplex. As communications are bi-directional, a duplex signal is required. Time division duplex (TDD) is, for example, a known duplexing technique based on TDMA. In TDD mode, inbound and outbound signal traffic to and from a radio access unit is sent subsequently over a channel. This could be symmetrically and asymmetrically, dependent on the system requirements. In the symmetric TDD mode, said access unit reserves equal transmission capacity in both directions, for example an alternating sequence of transmission capacity back and forth. Amongst others, some 3G mobile radio systems use CDMA in its TDD mode. This TDD-CDMA technique (also called time division-code division multiple access: TD-CDMA) is a combination of the above- mentioned FDMA, CDMA and TDMA/TDD techniques. The composite signal received by a radio access unit is divided in frequency bands, having an FDMA structure. Each frequency band or channel comprises multiple simultaneous signals of a number of mobile units. The simultaneous signals within a frequency band each comprise a unique code, a spreading sequence or spreading code, and are therefore distinguishable by their CDMA structure. Subsequently, within a frequency band, the access unit receives the combined signal of different groups of mobile units, which are present in the cell covered by said access unit. Duplexing has been established by an additional TDD feature (putting the CDMA in a TDD mode), whereby in combination with transmission to different user groups, transmission of a signal can take place back to the mobile units. TDD-CDMA adds significant complexity to the signal processing step in a radio access unit. For real-time data processing, such as speech, a signal processing algorithm has to be implemented, which can recognise and distinguish the spreading sequences or spreading codes of a large numbers of users as fast as possible. The algorithm used to decode the signal will have to perform this task in as few as possible computational steps.

The recognition and processing of the multiple signals in a TDD-CDMA based mobile radio system is called Joint Detection (JD). A major objective of this JD-technique, is the elimination of the interference. In CDMA, two important types of interference are recognised: inter-symbol interference (ISI) and multiple access interference (MAI).

The inter-symbol interference results from the Channel Impulse Response (CIR) of the channel used. The CIR is the impulse response of a channel to a delta signal transmitted by a mobile unit. This signal tends to blur due to reflections of the signal along its propagation path to the access unit, for instance reflection against buildings, trees, mountains and other objects along its path. As a mobile unit often moves within the environment, the CIR will be variable in time. Inter-Symbol Interference (ISI) is the interference between two subsequent symbols within a signal transmitted by a mobile unit to a radio access unit, which interference is caused by blurring of the signal .

Another source of interference in CDMA, related to the orthogonality of spreading codes, is the major cause of multiple access interference. Multiple Access Interference (MAI) is the interference between multiple signals received simultaneously by the access unit. If the spreading codes, used to delimit the signals from multiple units from each other, are not orthogonal (due to multipath propagation or other disturbing factors), the access unit might fail to a certain extend to distinguish differently coded signals. This will cause interference between the multiple signals transmitted simultaneously. In order to effectively eliminate the interference from the signals received, JD often makes use of the correlation of interference with the signals received. A mathematical description of the various JD techniques used within TDD-CDMA can be found in "Zero Forcing and Minimum Mean-Square-Error Equalisation for Multi User Detection in Code-Division Multiple-Access Channels", by Anja Klein, Ghassan Kawan Kaleh, and Paul Walter Baier, in IEEE transactions on vehicular technology, volume 45, number 2, May 1996. Another study of JD techniques can be found in "Comparative Study of Joint-Detection Techniques for TD-CDMA based Mobile Radio Systems", by Marius Vollmer, Martin Haardt, and Jϋrgen Gδtze, in IEEE journal on selected areas in communications, volume 19, number 8, August 2001.

Amongst others, these articles describe the construction of a signal matrix, having a typical block-Toeplitz structure, the use of Cholesky factorisation and the construction of a covariance matrix. A number of other statistical algorithms and techniques, such as minimum mean-square-error block linear equalizing (MMSE-BLE), are also described herein.

The JD algorithm, in addition to the signal received, requires knowledge about the orthogonal spreading codes and the CIR in order to process the signal. Knowledge about the CIR is provided by a separate channel estimator. This channel estimator usually analyses the midamble of a signal, which contains the midamble code, to determine the CIR. In the JD algorithm, the information about spreading codes and CIR is used to compose a convolution matrix. This convolution matrix forms the basis of a number of processes within the algorithm, amongst which is the construction of a covariance matrix, and Cholesky decomposition thereof. It is in particular these last two steps that form the bottleneck of JD algorithms used in prior art.

If M is the number of data symbols in a data field for each spreading code, and N is the number of spreading codes used in the algorithm, the covariance matrix comprises (N*M)² elements. The calculation of these (N*M)² elements in the covariance matrix, and the Cholesky decomposition step of this covariance matrix put a significant claim on the total number of instruction cycles used per decomposed time slot in the TDD-CDMA system. These two steps can be regarded as a performance bottleneck within the signal processing process.

In prior art the JD receiver processes the signals corresponding to all possible spreading codes in each slot of the TDD- CDMA system. If a system can handle N_max spreading codes simultaneously, and N codes are actually used in the system at a specific moment, the JD algorithm decomposes the signals corresponding to all codes, assuming N=N_max.

In prior art, it is further assumed that the ISI span of each signal is the maximum span allowed by the midamble structure. The CIR thus used by the JD algorithm is the maximum CIR produced by the channel estimator and is based on the preset midamble structure. The number of elements M for a certain code in the TDD-CDMA signal is determined by the CIR used, resulting in M=M_max. The above two assumptions made in prior art result in a standard (maximum) size of (N_max*M_max)² elements in the covariance matrix.

Although JD has evolved over the years, the complexity of JD in a TDD-CDMA system still remains a challenge to the designer. The maximum rate at which information can be sent and received to and from a mobile unit will be limited to the performance of the JD algorithm.

Summary of the Invention

It is an object of the present invention to provide a signal processing method of and an arrangement for a mobile telecommunications system operating with TDD-CDMA, which is optimised for its performance.

It is in particular an object of the present invention to provide a real-time signal processing method and arrangement with reduced complexity. These and further objects of the present invention are solved in accordance with the present invention by a method for signal processing in a telecommunications system based on Time Division Duplex (TDD) -Code Division Multiple Access (CDMA), wherein a plurality of users simultaneously transmit signals to a radio access unit over at least one channel, each of which signals comprises at least a unique spreading sequence and a denumerable number of elements, and which signals are simultaneously processed by said access unit using a signal processing algorithm, which signal processing algorithm comprises a step of construction of a convolution matrix of the combined effects of the spreading sequences and the channel, characterized in that complexity of said signal processing algorithm is reduced by reducing the number of elements in said convolution matrix, by making an assessment of elements that are likely to contain relevant information and constructing said convolution matrix at least of said elements that are likely to contain relevant information.

According to a preferred embodiment of the present invention, a receiver (or access unit) makes an assessment of elements that are likely to contain relevant information by determining whether a signal corresponding to a spreading code is idle (which at the time of receiving are not used in the cell), and ignoring spreading codes corresponding to idle signals in the JD algorithm. The complexity of the calculation during JD is reduced extensively by reducing the amount of spreading sequences to be processed.

In accordance with a further embodiment of the invention, the idleness of a signal can be decided upon if the signal power corresponding to each spreading code is below a certain threshold. This task could, for example, be performed each time a channel estimate is made by the joint channel estimator of the receiver.

The threshold, upon which the idleness of a signal is determined, can be made dependent upon particular situation in which the receiver will be used. One can think of accuracy requirements and available processing power as an example of these dependencies.

In another embodiment of the invention a receiving method is provided, wherein assessment of elements that are likely to contain relevant information comprises the step of reduction of the number of elements to be processes per received signal. In order to establish this, one can think of a receiver that determines and makes use of the actual CIR of a channel .

The length of the CIR can be determined from the channel estimate by using a number of taps that contain a sufficiently high percentage of CIR energy. When the percentage of CIR energy drops below a threshold, the measurement is stopped and the CIR is determined.

Again, the threshold value of the percentage of the CIR energy can be made dependent on parameters that determine the particular case in which the receiver will be used, such as accuracy requirements and available processing power.

The invention further relates to an arrangement for signal processing comprising means arranged for performing the method disclosed, and a radio access unit comprising such an arrangement.

In addition to the above the invention relates to a radio access unit comprised of means arranged for making an assessment of elements that are likely to contain relevant information.

The above mentioned and other features and advantages of the invention are illustrated in the following description of a preferred embodiment of the present invention, with reference to the enclosed drawings. Brief Description of the Drawings

Figure 1 shows a typical structure of a data burst of a signal of a mobile unit operating in a TDD-CDMA telecommunications systems.

Figure 2 shows the structure of the combined signal of a plurality of mobile units transmitting signals to a radio access unit in a TDD-CDMA telecommunications system.

Figure 3 is an illustration of a receiver operation, which may be comprised in a method according the present invention.

Figure 4 shows a method and arrangement according to the present invention.

Detailed description of the embodiments

In figure 1 a typical structure of a data burst 1 is shown as transmitted by a mobile unit to a radio access unit in a Time Division Duplex (TDD) -Code Division Multiple Access (CDMA) telecommunications system. The data burst 1 ends with a guard period 2, to delimit the data burst 1 from other data bursts. Between the guard period from a prior data burst (not shown) and guard period 2, data blocks 3 and 4 are transmitted to the radio access unit. The data blocks 3 and 4, comprising data sequences, are separated by a midamble 5, which comprises a midamble code. Information regarding the spreading code used by the signal is conveyed by the midamble code.

In TDD-CDMA the total or sum signal received by a radio access unit from a plurality of groups of radio mobile units (not shown) operative in the coverage area of the radio access unit can be represented by a matrix structure, shown schematically in figure 2. Here the data bursts 6 (as shown in figure 1) are represented, and each block in the matrix represents a different data burst (such as block 6). The different sets of blocks 7 represent the signal transmitted in different frequency bands, indicated by the abbreviation 'freq.¹ on the top left side of the figure.

Within a single frequency band, simultaneously a plurality of signals from a group of users, like 8 or 9, are transmitted to the radio access unit. All signals 8 belong to a single user group, in this case user group I, which consists of N signals each coded by a different spreading code, as indicated by the abbreviation 'code.' on the middle left side of the figure. Directly following in time upon the signals transmitted in user group I, as shown in 8, are the signals from user group II, as shown by 9, again these are N signals. Subsequently signals from user group III and user group IV, and so on will be received by the radio access unit. As the signals are duplex (which is not explicitly shown in figure 2), the radio access unit will transmit signals back to the mobile units, which can also be slotted in blocks as shown in figure 2.

Figure 3 illustrates an operation of a receiver unit in a TDD-CDMA system, which may be comprised in an arrangement according to the present invention. A sum signal 33 corresponding to a certain frequency, e.g. the first column 8 as shown in figure 2, is received by a receiver (as shown in figure 4, receiver 12). The receiver separates 34 the block of mida bles 35 from the data blocks 36. The part of the sum signal 36 representing the data blocks will be used later in the process, and is forwarded 38 to an element of the system performing signal processing. The part of the sum signal representing the midambles 35 will be forwarded 37 to a channels estimator (as shown in figure 4). Both parts of the signal 35 and 36 can be seen as a stack of overlapping individual signal parts representing data blocks and midambles respectively. Figure 4 shows a schematic representation of the method and arrangement according to the invention. Here, a plurality of N signals 10 are transmitted 11 to a radio access unit 12. The radio access unit 12 receives the sum signal 10 and separates a stack of midambles 14 and a stack of data blocks 13 from the sum signal 10 (as is illustrated in figure 3), note that at this point there is no knowledge about the individual midambles present in the sum signal 10. The stack of data blocks 13 will be used in the signal processing, as will be discussed below, and the stack of midambles 14 is forwarded to the channel estimator 16.

Based on built-in system knowledge, the channel estimator 16 determines a maximum possible (long) CIR 17 from the structure of the midambles in the stack received, and forwards this to the CIR analyser 18. In the CIR analyser 18, signal power is compared to a threshold 15, and from this comparison an improved (short) CIR estimate 19 is determined as well as the spreading codes 20 that correspond to idle signals.

During the abovementioned comparison, the signal power of the midambles is measured and compared to threshold 15 for all spreading codes. In case the signal power, corresponding to a certain spreading code, is below threshold 15, the signal is regarded an idle signal and the spreading code is classified as such. If the signal power is above threshold 15, a sample of the signal power is taken each time cycle or tap for a number of consecutive taps. When the signal power drops below the threshold at a certain tap, the number of taps from the first until this last sample is taken as the actual CIR corresponding to the spreading code. This comparison process is not explicitly shown in figure 3.

Preferably, the threshold value 15 is continuously updated in accordance with dynamically varying system parameters, such as available processing power and performance quality requirements. However, instead of or in addition to measuring the received signal power, other signal parameters indicative for the validity or relevance of a received signal may be applied by the method and arrangement according to the present invention. Such as a relative signal energy dependent parameter, being a percentage of the total signal energy received by a radio access unit. This relative signal energy parameter can be determined from counting a number of taps or time cycles of a received signal having an energy level above a minimum threshold value, for example.

The spreading codes 20 corresponding to idle signals are eliminated from a list of all available spreading codes 21 in selector 22, leaving only the active spreading codes 23. Accordingly, the selector also removes the information about the CIR's that correspond to idle signals (as these do not contain valid or relevant values), leaving only the active CIR's 24.

The active spreading codes 23 and CIR's 24 are forwarded to an element 25 that generates the convolution matrix of both active spreading codes and active CIR's. Since only the active spreading codes and the actual CIR's are used, a reduced number of signal elements need to be processed. The performance of this calculation is therefore improved due to the reduced computational complexity as compared to the case wherein all known spreading codes would be used in combination with an upper limit of the CIR.

The convolution matrix 26 is forwarded to an element 27 that constructs the system matrix. The system matrix 28 is fed to the element 29 which generates the covariance matrix. The covariance matrix 30 is fed to the signal processing unit 31 which performs Cholesky factorization and combines the incoming data blocks 13 with the result of the factorisation. The N estimated data signals 32 are sent on (for example to a telecommunications network).

In a further embodiment of the invention, the input signals 13 are processed with a suitable signal processing algorithm, such as the so-called Joint Detection (JD) algorithm. The JD algorithm comprises the steps of construction of a covariance matrix based on the system matrix, and processing of the covariance matrix using Cholesky factorisation.

It will be appreciated that numerous modifications and variations of the present inventions are possible in the light of the above teachings. Therefore, it will be understood that within the scope of the attached claims, the invention may be practised otherwise than as specifically described herein.

Claims

1. A method for signal processing in a telecommunications system based on Time Division Duplex (TDD)-Code Division Multiple Access (CDMA), wherein a plurality of users simultaneously transmit signals to a radio access unit over at least one channel, each of which signals comprises at least a unique spreading sequence and a denumerable number of elements, and which signals are simultaneously processed by said access unit using a signal processing algorithm, which signal processing algorithm comprises a step of construction of a convolution matrix of the combined effects of the spreading sequences and the channel, characterized in that complexity of said signal processing algorithm is reduced by reducing the number of elements in said convolution matrix, by making an assessment of elements that are likely to contain relevant information and constructing said convolution matrix at least of said elements that are likely to contain relevant information.

2. A method according to claim 1, wherein an assessment of elements that are likely to contain relevant information comprises a step of calculation of signal power corresponding to said spreading sequence, and comparison of said signal power to a threshold value.

3. A method according to claim 2, wherein said threshold value is continuously updated in accordance with dynamically varying system parameters, such as available processing power and performance quality requirements.

4. A method according to any of the previous claims, wherein an assessment of elements that are likely to contain relevant information comprises a step of identification of spreading sequences which correspond to idle signals.

5. A method according to claim 4 and any of the claims 2 or 3, wherein spreading sequences are identified as corresponding to idle signals if said signal power corresponding to said spreading sequence is smaller than said threshold value.

6. A method according to any of the claims 4 or 5, wherein said spreading sequences which correspond to idle signals are used to determine which elements are classified as elements that are not likely to contain relevant information.

7. A method according to any of the previous claims, wherein an assessment of elements that are likely to contain relevant information comprises determination of a propagation path Channel Impulse Ratio (CIR) for each signal, which CIR is determined by the time interval in which a majority of the received signal power is comprised.

8. A method according to claim 7, wherein said CIR is determined for each signal by counting a number of samples wherein a signal energy dependent parameter is measured, which parameter exceeds a threshold value.

9. A method according to claim 8, wherein said signal energy dependent parameter is the percentage of CIR energy present in said signal .

10. A method according to any of the claims 7-9, wherein said CIR is being used to determine which elements are classified as elements that are likely to contain relevant information.

11. A method according to any of the previous claims, wherein said convolution matrix is used to construct a system matrix, which system matrix is used to construct a covariance matrix, which covariance matrix is decomposed by using Cholesky factorisation, and wherein the result of said Cholesky factorisation is used in said signal processing algorithm.

12. An arrangement for signal processing in a telecommunications system based on Time Division Duplex (TDD) -Code Division Multiple Access (CDMA), wherein a plurality of users simultaneously transmit signals to a radio access unit over at least one channel, each of which signals comprises at least a unique spreading sequence and a denumerable number of elements, said arrangement comprising means for simultaneously processing said signals using a signal processing algorithm, and means for constructing a convolution matrix of the combined effects of said spreading sequences and said channel, characterized in that said arrangement comprises means for making an assessment of elements that are likely to contain relevant information and means for constructing said convolution matrix at least of said elements that are likely to contain relevant information.

13. An arrangement according to claim 12, comprising means for calculation of signal power corresponding to said spreading sequence, and means for comparison of said signal power to a threshold value.

14. An arrangement according to claim 13, comprising means for continuously updating said threshold value in accordance with dynamically varying system parameters, such as available processing power and performance quality requirements.

15. An arrangement according to any of the previous claims 12-14, wherein said means for making an assessment of elements that are likely to contain relevant information comprises means for identifying spreading sequences which correspond to idle signals.

16. An arrangement according to claim 15, comprising means for classifying spreading sequences which correspond to idle signals as elements that are not likely to contain relevant information.

17. An arrangement according to any of the previous claims, wherein said means for making an assessment of elements that are likely to contain relevant information comprises means for determination of a propagation path Channel Impulse Ratio (CIR) for each signal,

18. An arrangement according to claim 17, wherein said means for determination of a propagation path Channel Impulse Ratio (CIR) for each signal comprises means for determining the time interval in which a majority of the received signal power is comprised.

19. An arrangement according to claim 18, wherein said means for determining the time interval in which a majority of the received signal power is comprised, comprises means for counting a number of samples wherein a measured signal energy dependent parameter exceeds a threshold value.

20. An arrangement according to any of the claims 17-19, comprising means for using said CIR to classify elements as elements that are likely to contain relevant information.

21. An arrangement according to any of the previous claims, comprising, means for constructing a system matrix with said convolution matrix, means for construction of a covariance matrix from said system matrix, means for decomposing said covariance matrix using Cholesky factorisation, and signal processing means arranged for using the result of said Cholesky factorisation.

22. A radio access unit for use in a telecommunications system based on Time Division Duplex (TDD)-Code Division Multiple Access (CDMA), comprising signal processing means, which signal processing means comprise means for decomposing a plurality of signals on the basis of a unique spreading sequence comprised in each signal, and said signal processing means further comprise means for making an assessment of elements that are likely to contain relevant information and means for constructing a convolution matrix at least of said elements that are likely to contain relevant information.