CN101232295B

CN101232295B - Training method of adaptive equalizer in wireless communicating system receiver

Info

Publication number: CN101232295B
Application number: CN200710154157XA
Authority: CN
Inventors: 陈晓春; 周亮
Original assignee: Mawier International Trade Co Ltd
Current assignee: Marvell World Trade Ltd; Mawier International Trade Co Ltd
Priority date: 2006-10-04
Filing date: 2007-09-19
Publication date: 2013-06-26
Anticipated expiration: 2027-09-19
Also published as: CN101232295A

Abstract

The invention discloses a training method of equalizer in a wireless communication system modulation process, comprising a training process and a judging process executed one after another. In the training process, processing an iterative computation by each symbol in a known sequence in received data so as to adjust a filter coefficient of the self-adapted equalizer. In the continuous judging process, modulating the received data by adjusted filter coefficient; in the training process, at least one symbol in the known sequence is used more than once for the iterative computation.

Description

The training method of the adaptive equalizer in wireless communication system receiver

Technical field

The present invention relates to wireless communication system, more particularly, relate to the equalizer that is used for the receiving end signal demodulation in wireless communication system.

Background technology

PHS (personal handyphone system) is a kind of simple and low wireless communication system of cost, it adopts simple sending/receiving machine technology, and for example modulation and demodulation method adopts respectively DQPSK (difference quadrature phase shift keying) modulation of pi/4 shift and the Non-coherent Demodulation in receiver.The characteristic that PHS is simple due to it and cost is low becomes the communication system of extensive use.

Fig. 1 shows the frame format that the PHS system is transmitted.As shown in Figure 1, the frame format of PHS transmission mainly contains two kinds: control channel (CCH) frame and Traffic Channel (TCH) frame.The CCH frame be mainly used between base station (CS) and terminal (PS) synchronize and other is controlled etc.The TCH frame is mainly used in the transport communication data.Each time slot is for frame of transmission, and therefore the time slot in whole PHS channel also can be divided into for the CCH time slot of transmission CCH frame with for the TCH time slot that transmits the TCH frame.Each time slot is 120 symbols, totally 625 μ s.These two kinds of frames all comprise rise/fall time (R), initial symbol (SS), lead code (PR), unique word (UW), data (DATA), cyclic redundancy check (CRC) code (CRC) and guard time (GT).In these two kinds of frames, the symbolic number of corresponding each field is different.For example, as shown in Figure 1, the UW field of CCH frame reaches 16 symbols, and the UW field of TCH frame only has 8 symbols.

Receiver in terminal (PS) is used for receiving CCH frame and TCH frame.Fig. 2 shows the part block diagram of the PHS receiver that designs based on the coherent demodulation that utilizes adaptive equilibrium method.As shown in Figure 2,201 pairs of baseband receiving signals of AFE (analog front end) (AFE) part carry out analog-to-digital conversion.Digital signal after conversion is filtered and extraction by decimation filter 202.This signal is through carrier recovery block 203 and circulator module 204.The function of carrier recovery block 203 is the carrier frequency shifts that detect between burst and estimating received signal and transmitted signal.Therefore, ensuing circulator module 204 is used for the compensation carrier shift.Signal after circulator 204 is processed is imported into correlator 206 and equalizer 207 through A/B buffer 205.

The mode that correlator 206 is used for by correlation computations positions the UW field of frame.In correlation module, will input data and UW and carry out correlation computations, obtain relevance values, in the peak value that this relevance values the detected position of corresponding UW in bursty data just.Then equalizer 207 compensates it for the intersymbol interference of channel is assessed, thereby realizes the equilibrium of channel.Signal after equilibrium is output to the frame demodulator.

Equalizer 207 is generally with filter realization, for example sef-adapting filter.Adaptive equalizer operates mainly in two kinds of patterns: training mode and tracing mode.Fig. 3 shows the schematic diagram of mode of operation of the adaptive equalizer of prior art.For each frame that receives, at first equalizer works in training mode, assess the characteristic of channel by known algorithm and channel is compensated with unique word (UW) symbol (known array) that correlator 206 detects, using simultaneously the coefficient of the mode correction wave filter of iterative computation; After the UW sign-off, adaptive equalizer works in tracing mode, utilize the filter factor that under training mode, iterative computation obtains that the data symbol is adjudicated, and the symbol or the known array that obtain according to judgement, adaptive algorithm by equilibrium is constantly finely tuned filter factor, thereby follows the tracks of the channel that constantly changes.Obviously, training process is vital for demodulation result, therefore should guarantee filter factor convergence or approaching convergence as much as possible before training process finishes, otherwise ensuing data symbol will be increased by the probability of demodulation mistakenly.

But as shown in Figure 1, the UW field of the TCH frame of PHS agreement regulation only has 8 symbols.Therefore, when equalizer 207 utilizes UW to train according to mode shown in Figure 3, when finishing, the UW field is difficult to also guarantee that filter factor restrains well, especially when SNR (signal to noise ratio) is not high, the more difficult convergence of filter factor, thereby cause the demodulation performance of data symbol relatively poor, BER (error rate) is higher.

Summary of the invention

in view of above-mentioned and other problem of the prior art, a technical scheme of the present invention provides a kind of training method of the adaptive equalizer for wireless communication system receiver, the method comprises training process and the judging process of in succession being carried out, in described training process, each symbol in known array in the data that utilization receives carries out iterative computation, to adjust the filter factor of described adaptive equalizer, in follow-up described judging process, utilize the filter factor after adjusting to carry out demodulation to the data that receive, wherein in described training process, at least one symbol in described known array is by more than once being used for carrying out described iterative computation.

another technical scheme of the present invention provides a kind of training method of the adaptive equalizer for wireless communication system receiver, the method comprises training process and the judging process of in succession being carried out, in described training process, each symbol in known array in the data that utilization receives carries out iterative computation, to adjust the filter factor of described adaptive equalizer, in follow-up described judging process, utilize the filter factor after adjusting to carry out demodulation to the data that receive, wherein said training process comprises the following steps: from first sign-on of described known array, carry out successively iterative computation along the data receiver order, judgement has been carried out iterative computation to last symbol of described known array, if and last symbol of described known array has been carried out iterative computation, return to first symbol of described known array.

another technical scheme of the present invention provides a kind of training method of the adaptive equalizer for wireless communication system receiver, the method comprises training process and the judging process of in succession being carried out, in described training process, each symbol in known array in the data that utilization receives carries out iterative computation, to adjust the filter factor of described adaptive equalizer, in follow-up described judging process, utilize the filter factor after adjusting to carry out demodulation to the data that receive, wherein said training process comprises the following steps: from last sign-on of described known array, along carrying out successively iterative computation with the order of data receiver reversed in order, judge whether first symbol of described known array has been carried out iterative computation, if first symbol of described known array carried out iterative computation, begin to carry out successively iterative computation along the data receiver order.

Description of drawings

Read by reference to the accompanying drawings following detailed description the in detail of the present invention, the present invention may be better understood and advantage and further feature, wherein:

Fig. 1 shows the time slot format that the PHS system is transmitted;

Fig. 2 shows the part block diagram of PHS receiver;

Fig. 3 shows the schematic diagram of mode of operation of the adaptive equalizer of prior art;

Fig. 4 shows the schematic diagram of mode of operation of the adaptive equalizer of one embodiment of the invention;

Fig. 5 shows the flow chart of training method of the adaptive equalizer of one embodiment of the invention;

Fig. 6 shows the schematic diagram of mode of operation of the adaptive equalizer of another embodiment of the present invention;

Fig. 7 shows the flow chart of training method of the adaptive equalizer of another embodiment of the present invention.

Embodiment

In order to understand better the present invention, the below will describe specific embodiments of the invention in detail.But recognize, these embodiment are only illustrative rather than restrictive, those skilled in the art under the instruction of these embodiment, be easy to make variously substitute, distortion and improving.In addition, in this manual, some known structure, parts, feature, method are not elaborated, in order to avoid main contents of the present invention are thickened.In the accompanying drawings, unless specialize, same or similar label represents same or similar parts.

In one embodiment of the invention, design LMS (lowest mean square) adaptive equalizer is realized coherent demodulation.In the situation that it is shorter to be used for the known array of training, the situation of 8 symbols is for example only arranged for UW in the TCH frame of PHS, in order to allow filter factor restrain better, can carry out two or more times training to UW.Fig. 4 shows the schematic diagram of mode of operation of the adaptive equalizer of one embodiment of the invention.As shown in Figure 4, after known array UW finished, training mode did not finish, but begins once to train from the primary sign of UW again, then just adjudicates, and namely altogether UW is carried out twice training.Frequency of training in figure is only exemplary, also can train more frequently UW according to the specific requirement of filter factor convergence, so that filter factor can be restrained well.UW is repeatedly trained the length that has been equivalent to increase known array, can guarantee better the convergence of filter factor.

Fig. 5 shows the flow chart of training method of the adaptive equalizer of one embodiment of the invention.As shown in Figure 5, training starts from step 501.In step 502, the 1st symbol of location UW.As previously mentioned, can obtain by correlation module the exact position of UW, therefore can accurately locate the 1st symbol of UW.Then UW is carried out forward (from left to right, also namely pressing the reception order of data Fig. 4) training (step 503) from the 1st sign-on.In step 504, judged whether to arrive the 8th symbol of UW, i.e. last symbol of known array.If do not arrive last symbol of known array, continue UW is carried out the forward training.If judgement has arrived last symbol of known array in step 504, return to the 1st symbol (step 505) of UW, and then carry out forward training (step 506).In step 507, judged whether to arrive the 8th symbol of UW, i.e. last symbol of known array.If do not arrive the 8th symbol, continue UW is carried out the forward training.If judgement has arrived the 8th symbol in step 507, training finishes (step 508).Training is adjudicated and filter factor is followed the tracks of data after finishing.

Fig. 6 shows the schematic diagram of mode of operation of the adaptive equalizer of another embodiment of the present invention.As shown in Figure 6, the training of equalizer is not the first sign-on from UW, but first carries out reverse train one time from its last sign-on, and then UW is carried out a forward training.This method not only can obtain the effect of the method shown in Fig. 4 b, and has guaranteed the amplitude of channel pulse in whole training process and the continuity of phase place, has especially guaranteed the continuity of the phase place of equalizer coefficients, thereby can further improve performance.

Fig. 7 shows the flow chart of training method according to another embodiment of the present invention.As shown in Figure 7, training starts from step 701.In step 702, the 8th symbol of location UW, i.e. last symbol of known array.As previously mentioned, can obtain by correlation module the exact position of UW, therefore can accurately locate the 8th symbol of UW.Then UW is carried out oppositely (Fig. 4 from right to left, also namely with the reception reversed in order of data) training (step 703) from the 8th sign-on.In step 704, judged whether to arrive the 1st symbol of UW, if do not arrive the 1st symbol, continue UW is carried out reverse train.If judgement has arrived the 1st symbol of UW in step 704, UW is carried out forward (from left to right, also namely pressing the reception order of data Fig. 4) training (step 705) from the 1st sign-on.In step 706, judged whether to arrive the 8th symbol of UW, i.e. last symbol of known array.If do not arrive the 8th symbol, continue UW is carried out the forward training.If judgement has arrived the 8th symbol in step 706, training finishes (step 707).Training is adjudicated and filter factor is followed the tracks of data after finishing.

It should be noted, the specific embodiment that more than provides not is restrictive.For example, although the applied environment that provides in embodiment is the reception of the TCH channel of PHS receiver, the present invention also can be applied to other communication system.And known array also is not limited to UW, and any known array that can be used for carrying out the adaptive equalizer training all is fine.

And the frequency of training in embodiment can be also more times, and for example 4 times, the training process of this moment can be the training of 4 forwards, or a reverse train, a forward training, reverse train and forward training again again.In addition, training process can be also other any combination of frequency of training and training direction.

In addition, the inventor has also expected other embodiment.For example, training process of the present invention not necessarily will carry out identical number of times to each symbol in known array.In a specific embodiment, training is from the 8th (being last) sign-on of UW, first carry out reverse train to the 1st symbol of UW, carry out again forward and train the 3rd symbol of UW, can obtain to increase so too the effect of filter factor convergent probability, but also can carry out with the 4th to the 8th symbol of UW other some calculate, such as the 4th to the 8th symbol to UW adjudicated, with estimation of carrying out the error rate (BER) etc.

Above the present invention is described in detail, wherein stated a lot of specific features in conjunction with specific embodiment, but the present invention is not limited to above-mentioned specific implementations.In each specific embodiment, above-mentioned various specific features also are not limited to described combination.In any limited range that falls into appended claims all should be considered as falling into protection scope of the present invention to the substituting of the specific embodiment of the invention, distortion and improvement and equivalent thereof within.

Claims

1. training method that is used for the adaptive equalizer of wireless communication system receiver, the method comprises training process and the judging process of in succession being carried out, in described training process, each symbol in known array in the data that utilization receives carries out iterative computation, to adjust the filter factor of described adaptive equalizer, in follow-up described judging process, utilize the filter factor after adjusting to carry out demodulation to the data that receive, wherein

In described training process, last sign-on from described known array, along carrying out successively iterative computation with the order of data receiver reversed in order, and after first symbol to described known array has carried out iterative computation, carry out successively iterative computation from first sign-on of described known array along the data receiver order.

2. the method for claim 1, wherein after each symbol in described known array is carried out the iterative computation of same number, carry out described judging process.

3. method as claimed in claim 2, wherein said same number is 2 times.

4. the method for claim 1, wherein in described training process, the number of times that each symbol in described known array is carried out iterative computation is incomplete same.

5. method as described in any one in claim 1-4, wherein said wireless communication system is the PHS system, and described known array is the unique word in traffic channel frame.