GB2362075A

GB2362075A - Spread-spectrum detection apparatus and method therefor

Info

Publication number: GB2362075A
Application number: GB0010548A
Authority: GB
Inventors: Leo Rademacher
Original assignee: Siemens AG
Current assignee: Siemens AG
Priority date: 2000-05-03
Filing date: 2000-05-03
Publication date: 2001-11-07
Anticipated expiration: 2020-05-03
Also published as: GB2362075B; GB0010548D0

Abstract

In Wideband-Code Division Multiple Access (W-CDMA) systems employing decimating filters for joint detection of time-variant signals it is desirable to directly calculate coefficients of the decimating filters. In order to calculate the coefficients, an input signal is de-spread (step 500) and a convolution matrix generated (step 502) based upon the de-spread input signal. A Moore-Penrose pseudo-inverse of the matrix is calculated (step 504) and used with a training sequence to determine initial coefficients. Advantageously, much of the complexity of known systems is avoided.

Description

DETECTION APPARATUS AND METHOD THEREFOR The present invention relates to

an apparatus for detection of timevariant spread spectrum signals of the type used in a spread spectrum system 5 having a time-variant spreading code, for example, Wideband-Code Division Multiple Access (W-CDMA) systems. The present invention also relates to a method of detection of time-variant spread spectrum signals.

In a spread spectrum radio communications system, baseband information from each user is combined with a spreading code specific to the user in order to yield corresponding wideband signals resembling noise. At a spread spectrum receiver, a signal is received which corresponds to a superposition of all the signals of all users in the system. In order to extract the baseband information of a given user from the received signal, a technique known as "Joint Detection" has been proposed by A. Klein and P.

W. Baier in "Linear Unbiased Data Estimation in Mobile Radio Systems Applying CDMA" (IEEE Transactions on Selected Areas in Conununications, Vol. SAC- 11, No. 3, September 1993, pages 105 8 to 1066).

In Joint Detection, a decimating filter is used at the spread spectrum receiver to extract symbols corresponding to the given user, coefficients of the decimating filter being specific to the extraction of symbols corresponding to the given user.

GB-A-2 337 673 discloses a method of directly calculating coefficients of the decimating filter for time-invariant spreading codes, due to the same matrix-vector mathematical description shared by the joint detection systems and adaptive antenna systems. By "time invariant spreading" it is meant that successive symbols are spread by convolution with the same segment of a spreading code.

However, the method described in the above document relates to timeinvariant spreading codes and so is inapplicable to W-CDMA, which uses time variant spreading which, in contrast with time invariant spreading, involves the convolution of symbols with different segments of the spreading 5 code.

It is therefore an object of the present invention to obviate or at least mitigate the above disadvantage associated with direct calculation of filter coefficients for spread spectrum systems employing time invariant spreading.

According to the present invention there is provided an apparatus for detection of a time-variant spread-spectrum signal comprising a processor arranged to provide a plurality of de-spreading filters capable of generating a respective plurality of output signals representable as a vector, the plurality of de-spreading filters being respectively coupled to a plurality of decimating filters, and means for generating a convolution matrix of the vector and a pseudo-inverse of the convolution matrix, wherein coefficients of the plurality of decimating filters are calculated based upon the pseudo- inverse of the convolution matrix and a training sequence.

Preferably, the apparatus further comprises a channel model filter unit coupled to an adaptive algorithm unit, wherein the adaptive algorithm unit is coupled to the plurality of decimating filters.

Preferably, the adaptive algorithm unit is arranged to adapt the coefficients of the channel model filter unit and the coefficients of the plurality of decimating filters in response to an error signal.

Preferably, the error signal is a difference between an output signal generated by the channel model filter unit and a sum of respective output signals generated by the plurality of decimating filters.

According to the present invention, there is also provided a method of detection of a time-variant spread-spectrum signal comprising the step of. . de- spreading an input signal to form a plurality of output signals representable as a vector; generating a convolution matrix of the vector; generating a pseudo-inverse of the convolution matrix; providing a plurality of decimating filters; calculating the coefficients of the decimating filters using the pseudo- inverse and a training sequence, and passing the de-spread signal through the plurality of decimating filters.

Preferably, the method further comprises the step of. adapting the coefficients of the decimating filters based upon a sum of output signals of the decimating filters and an output signal of a channel model filter unit.

It is thus possible to provide apparatus for and method of generating filter coefficients for detection in a spread spectrum communication system employing time invariant spreading codes.

At least one embodiment of the invention will now be described, by way of example only, with reference to the accompanying drawings, in which:

Figure 1 is a schematic diagram of an apparatus constituting a communications link; Figure 2 is a schematic diagram of a mobile terminal shown in Figure 1; Figure 3 is a schematic diagram of a base station shown in Figure 1; Figure 4 is a schematic diagram of a detector apparatus employed by the base station of Figure 3 constituting an embodiment of the invention, and Figure 5 is a flow diagram of the operation of the detector apparatus of Figure 4.

Throughout the following description like parts will be identified by identical reference numerals.

In a cellular teleconununications network supported by, for example, a W-MMA system 100 (Figure 1), a base station 102 supports a geographical area, or cell 104, the base station 102 being in communication with a mobile subscriber unit 106 via a radio frequency (RF) interface 108.

As an example only, communications between the base station 102 and a Public Switched Telecommunications Network (PSTN) 110 can be supported by any telecommunications architecture 112 known in the art. A fixed-line telephone 114 is also coupled to the PSTN I 10.

It should be appreciated that although reference has been made above to particular types of terminals, other terminals can be used instead of the base station 102 or the mobile subscriber unit 106, including, for example, fixed cellular terininals, or laptop computers/PDAs suitably adapted to function within the W-CDMA system 100. Similarly, although a fixed-line telephone 114 has been described above, other communications devices are envisaged, for example, a personal computer (PC) and a modem, or another mobile subscriber unit operating in the W-CDMA system.

Referring to Figure 2, the mobile subscriber unit 106 comprises a terminal antenna 200 coupled to a terminal duplexer 202. A first terminal of the terminal duplexer 202 is coupled to a terminal Digital Signal Processor (DSP) 204 via a terminal transmitter chain 206. Similarly, a second terminal of the terminal duplexer 202 is coupled to the terminal DSP 204 via a terminal receiver chain 208. The terminal DSP 204 is coupled to a terminal Random Access Memory (RAM) 210, a display 212, for example, a liquid crystal display, a speaker unit 214, a keypad 216 and a microphone 218.

The base station 102 (Figure 3) comprises a base station antenna 300 coupled to a base station duplexer 302. A first terminal of the base station duplexer 302 is coupled to a base station microprocessor 304 via a base station transmitter chain 306. Similarly, a second terminal of the base station duplexer 302 is coupled to the base station microprocessor 304 via a base station receiver chain 308.

The base station microprocessor 304 is coupled to a base station RAM 3 10 and a DSP unit 311. Information is communicated to and from other parts of the cellular telecommunications network (not shown) by means of an 1/0 interface 312 coupled to the base station microprocessor 304.

For the purposes of simplicity of description and hence clarity, a detector 400 (Figure 4) according to an embodiment of the invention will now be described in relation to the base station 102 only. However, it should be appreciated that the detector can equally be employed in the mobile terminal 106.

Referring to Figure 4, the DSP unit 311 of the base station 102, is arranged to comprise the detector 400, the detector 400 comprising a terminal input port 402 for receiving an input signal corresponding to a superposition of signals received from all transmitting users. The input port 400 is coupled to a first input port 406 of a bank of de-spreading filters 404 and a second input port 408 of the bank of de-spreading filters 404 via a first delay 412.

The input port 402 is also coupled to a third input port 410 of the bank of de spreading filters 404 via the first delay 412 and a second delay 414.

The bank of de-spreading filters 404 comprises a first sub-bank of de spreading filters 416, a second sub-bank of de-spreading filters 418, and a third sub-bank of de-spreading filters 420. The first sub-bank of de spreading filters 416, the second sub-bank of de-spreading filters 418, and the third sub-bank of de-spreading filters 420 are coupled to the first input port 406, the second input port 408 and the third input port 410 of the bank of de spreading filters 404, respectively.

A first output port 422 of the first sub-bank of de-spreading filters 416 is coupled to a first input port 424 of a bank of adaptive decimating filters 446, a second output port 426 of the first sub-bank of despreading filters 416 being coupled to a second input port 428 of the bank of adaptive decimating filters 446. A first output port 430 of the second sub-bank of de spreading filters 418 is coupled to a third input port 432 of the bank of adaptive decimating filters 446, a second output port 434 of the second sub bank of de-spreading filters 418 being coupled to a fourth input port 436 of the bank of adaptive decimating filters 446. A first output port 438 of the third sub-bank of de-spreading filters 420 is coupled to a fifth input port 440 of the bank of adaptive decimating filters 446, a second output port 442 of the third sub-bank of de-spreading filters 420 being coupled to a sixth input port 444 of the bank of adaptive decimating filters 446.

The bank of adaptive decimating filters 446 comprises a first sub bank of adaptive decimating filters 448, a second sub-bank of adaptive decimating filters 450, a third sub-bank of adaptive decimating filters 452.

The first sub-bank of decimating filters 448 comprises a first decimating filter 454 having an input port coupled to the first input port 424 of the bank of adaptive decimating filters 446 and a second decimating filter 456 having an input port coupled to the second input port 428 of the bank of adaptive decimating filters 446. The second sub-bank of decimating filters 450 comprises a third decimating filter 458 having an input port coupled to the third input port 432 of the bank of adaptive decimating filters 446 and a fourth decimating filter 460 having an input port coupled to the fourth input port 436 of the bank of adaptive decimating filters 446. The third sub- bank of decimating filters 452 comprises a fifth decimating filter 462 having an input port coupled to the fifth input port 440 of the bank of adaptive decimating filters 446 and a sixth decimating filter 464 having an input port coupled to the sixth input port 444 of the bank of adaptive decimating filters 446. Output terminals of the first, second, third, fourth, fifth and sixth adaptive decimating filters 454, 456, 458, 460, 462, 464 are coupled to a first summation unit 466, an output port of the first summation unit 466 being coupled to a first input port of a second summation unit 468. A second input port of the second summation unit 468 is coupled to an output terminal of a model filter unit 470, for example, a transversal model filter. An output port of the second summation unit 468 is coupled to a first input port of an adaptive algorithm unit 472 arranged to execute any adaptive algorithm known in the art, for example, a Least Mean Square (LMS) algorithm. A second input port of the adaptive algorithm unit 472 is coupled to a symbol input port 474, the symbol input port 474 also being coupled to a first input port of the model filter unit 470.

A first output port of the adaptive algorithm unit 472 is coupled to a second input port of the model filter unit 470 in order to adapt coefficients of the model filter unit 470. Sin-filarly, a second output port of the adaptive algorithm unit 472 is coupled to each of the first, second, third, fourth, fifth and sixth adaptive decimating filters 454, 456, 458, 460, 462, 464 in order to adapt coefficients of the first, second, third, fourth, fifth and sixth adaptive decimating filters 454, 45 6, 45 8, 460, 462, 464.

In operation, the input signal present at the terminal input port 402 is fed to the bank of de-spreading filters 404 (step 500). The coefficients of each of the first, second and third sub-banks of de-spreading filters are selected so as to de-spread the input signal.

For a given symbol to be detected, the input signal also includes a precursor component and a post-cursor component corresponding to symbols preceding and following the given symbol to be detected, respectively. The precursor and post-cursor components are present in the input signal because, due to the nature of time variant spreading, the preceding, following and given symbol to be detected are spread with different segments of a spreading code. Consequently, the bank of de-spreading filters 404 comprises the first, second and third sub-banks of de-spreading filters 416, 418, 420. The bank of de-spreading filters 404 can be expressed as a Moore-Penrose pseudo inverse matrix of the segments used of the spreading code, E,+, corresponding to the coefficients of the bank of de-spreading filters 404.

After processing by the bank of de-spreading filters 404, signals present at the output ports 422, 426, 430, 434, 438, 442 of the bank of de-spreading filters 404 correspond to a de-spread version of the input signal which can be expressed mathematically as a vector of the de-spread input signal, iF.

In order to detect the symbols of the given user, the symbols of the given user have to be extracted from the de-spread signal E comprising the precursor and post-cursor symbols for all user signals which can be received.

Consequently, the de-spread version of the input signal is processed by the bank of adaptive decimating filters 446. The coefficients of the bank of decimating filters 446 can be expressed as a vector of decimating filter coefficients, k-u 5 and are initially determined using a training sequence, du in accordance with the following equation:

du = kiu By re-arranging equation (1), it is possible to obtain an equation for the vector of decimating filter coefficients, u:

ku = E+du (2) where: is a convolution matrix of the vector of the de-spread input signal e- (step 502), E+ being the pseudo-inverse of the convolution matrix of the vector of the de-spread input signal j (step 504).

As should be appreciated by a person skilled in the art, the structure of the convolution matrix E- is dependent upon the number of coefficients ku selected for the bank of decimating filters 446.

Using the pseudo-inverse of the convolution matrix and the training sequence d, the coefficients, of the bank of decimating filters 446 are calculated. Next, the de-spread version of the input signal is processed (step 506) by the decimating filters 454, 4565 458,460,462,464, the outputs of the decimating filters 454, 45 6, 45 8,460, 462,464 being summed by the first summation unit 466.

Once the bank of decimating filters 446 have been provided with initial coefficients, adaptation (step 508) of the initial coefficients is performed by the adaptive algorithm unit 472. In order to adapt the coefficients of the bank of decimating filters 446, the adaptive algorithm unit 472 receives an error signal, e, generated by the second summation unit 468 based upon a difference between an output signal generated by the bank of decimating filters 446 and the channel model unit 470. The channel model unit 470 generates a model of the communications link between the given user and the receiver. i.e. the transmitter chain of a terminal of the given user, the channet the receive chain of the base station 102, the bank of de spreading filters 404 and the bank of decimating filters 446. The error signal, e, generated using the output signal of the model filter 470 is used as a reference signal for the adaptation performed by the adaptive algorithm unit 472 and for detection using a detection device, for example, a slicer or a Viterbi detection algorithm. Coefficients of the channel model filter 470 are also adapted by the adaptive algorithm unit 472 in accordance with changes in the channel and the bank of decimating filters 446.

Although the above example has been described in the context of a detector for two users, U=1 and U=2, it should be appreciated that the above described apparatus can be expanded for a larger number of users U=n by the provision of additional de-spreading filters in each of the first, second and third sub-banks of de-spreading filters 416, 418, 420, each additional despreading filter corresponding to an additional user. Also, for additional users, additional decimating filters corresponding to the additional despreading filters must be provided.

In the event that inter-symbol interference is caused by symbols more than one symbol apart from a symbol to be detected, additional sub-banks of de-spreading filters and corresponding decimating filters can be provided to extract the symbol to be detected from additional interfering precursor and post-cursor symbols.

Claims

CLAIMS:

1. An apparatus for detection of a time-variant spread-spectrum signal comprising a processor arranged to provide a plurality of de-spreading filters capable of generating a respective plurality of output signals representable as a vector, the plurality of de-spreading filters being respectively coupled to a plurality of decimating filters, and means for generating a convolution matrix of the vector and a pseudo-inverse of the convolution matrix, wherein coefficients of the plurality of decimating filters are calculated based upon the pseudo-inverse of the convolution matrix and a training sequence.

2. An apparatus as claimed in Claim 1, further comprising a channel model filter unit coupled to an adaptive algorithm unit, wherein the adaptive algorithm unit is coupled to the plurality of decimating filters.

3. An apparatus as claimed in Claim 2, wherein the adaptive algorithm unit is arranged to adapt the coefficients of the channel model filter unit and the coefficients of the plurality of decimating filters in response to an error signal.

4. An apparatus as claimed in Claim 3, wherein the error signal is a difference between an output signal generated by the channel model filter unit and a sum of respective output signals generated by the plurality of decimating filters.

5. A method of detection of a time-variant spread-spectrum signal comprising the step of.

de-spreading an input signal to form a plurality of output signals representable as a vector; generating a convolution matrix of the vector; generating a pseudo-inverse of the convolution matrix; providing a plurality of decimating filters; calculating the coefficients of the decimating filters using the pseudo inverse and a training sequence, and passing the de-spread signal through the plurality of decimating filters.

6. A method as claimed in Claim 5, further comprising the step of.

adapting the coefficients of the decimating filters based upon a sum of output signals of the decimating filters and an output signal of a channel model filter unit.

7. An apparatus for detection of a time-variant spread-spectrum signal substantially as hereinbefore described with reference to Figure 4.

8. A method of detection of a time-variant spread-spectrum signal substantially as hereinbefore described with reference to Figure 5.