CN100505725C - Channel equalization method of OFDM system - Google Patents

Abstract

This invention discloses a channel equalization method for OFDM system, which contains the receiver making FFT conversion to signal after receiving time domain signal to obtain the sub carrier signal Yk of each data in frequency domain, obtaining frequency domain Hk of each sub carrier by estimating channel estimation, equalizing signal to restore signal transmitted in sub carrier, calculating soft decision value of each bit Uki in each sub carrier, weighting and decoding to obtain final binary data, said invention makes the input of Viterbi decoder to be the log likelihood ratio of each sub carrier bit, the Viterbi decoding algorithm is the maximum likelihood decoding of convolution code, which greatly raises system performance.

Description

The channel equalization method of ofdm system

Technical field

The present invention relates to the channel equalization method of a kind of OFDM (Orthogonal Frequency Division Multiplexing) system.

Background technology

Along with the develop rapidly of the communication technology, realize that the broadband wireless communication technique of the high-speed transfer of data message is one of main direction of Communication Development.In recent years, OFDM (OFDM) technology has been considered in the broadband wireless communications a kind of basic fundamental of intersymbol interference ISI under the contrary frequency selective channel condition.Because OFDM will exist the broad-band channel of ISI to be decomposed into the narrow band channel of a plurality of no ISI, simultaneously data-signal at a high speed is decomposed into the low speed data signal of the outlet that can walk abreast, effectively resist ISI, therefore be regarded as resisting the key technology of multidiameter fading channel.The second class HiperLAN Hiperlan/2 and the mobile multimedia access communications (MMAC) of WLAN standard IEEE 802.11a, ETSI have all adopted the physical layer specification of OFDM technology as them.These systems all will encode and the OFDM technology combines the opposing multipath channel, effectively raise the reliability of transmission.

As shown in Figure 1, typical ofdm system transmitter architecture comprise convolutional encoding, interweave, mapping, IFFT and add Cyclic Prefix (CP).Its course of work is, at first binary source carried out convolutional encoding, interweaves then.Interweaving only changes the order of bit stream, and purpose is the burst error that the opposing multidiameter fading channel causes.The bit stream that will interweave is mapped to BPSK, QPSK or the QAM signal in the planisphere.For specific modulation system, every m bit u _{K, i}, i=1 ..., m is mapped to a signaling point X in the planisphere _k, k=0 ..., N-1.Here N represents the number of sub carrier wave of transmitting useful data, and k represents the subcarrier sequence number.If the system modulation exponent number is M, then the number of bits of each constellation point correspondence is m=log ₂M.Carry out the IFFT conversion then and be modulated on the different subcarriers, send into multipath channel after adding Cyclic Prefix at last.

As shown in Figure 2, typical ofdm system receiver structure comprises Cyclic Prefix (CP), FFT, channel estimating, channel equalization, soft-decision, deinterleaving and Viterbi decoding.The signal that receives is gone to do behind the Cyclic Prefix FFT conversion obtain signal Y on each data subcarrier of frequency domain _k, 0≤k≤N-1.According to the ofdm signal model Y is arranged _k=H _kX _k+ Z _k, k=0,1 ..., N-1, here, Z _kFor average is zero, variance is σ ²White Gaussian noise.For coherent demodulation, calculate the response H of each data subcarrier _k, 0≤k≤N-1.The subcarrier responses H that utilization estimates _kCarry out the data of channel equalization to recover to send on each subcarrier.Balanced data

Need soft-decision, deinterleaving and behind the Viterbi error-correcting decoding, obtain binary data.The performance that the soft-decision input is imported than hard decision for Viterbi decoding is well a lot, and this is because soft-decision-value not only provides the court verdict of each bit of received signal, and provides the confidence level of this court verdict.

According to the character of ofdm system, the received signal on each subcarrier with the pass that transmits is:

Y _k＝H _kX _k+Z _k，k＝0，1，…，N-1

Here, Z _kFor average is zero, variance is σ ²White Gaussian noise.

At present the method for channel equalization have direct phase division (direct equalization) or least mean-square error (minimummean square error, MMSE) 2 kinds:

1, direct phase division.Specifically can be with reference to " Chien-Fang Hsu; Yuan-Hao Huang and Tzi-Dar Chiueh; " Design of an OFDM receiver for high-speed wireless LAN "; IEEE 200 1InternationalSymposium on Circuits and Systems; vol.4; pp.558-561, May2001 ", the signal Y on each subcarrier that it will receive _kDivided by the response H on this subcarrier _k, that is:

$X_k^{\′}=\frac{Y_k}{H_k},$ k＝0，1，…，N-1

2, least mean-square error method.Specifically can be with reference to " Boumard; S.and Mammela; A.; " Channel estimationversus equalization in an OFDM WLAN system "; IEEE2001 Vehicular Technology Conference; pp.653-657, May 2001 ", the data that least mean-square error method (MMSE criterion) goes out equilibrium

With its true value X _kMean square error

Minimum can obtain:

$X_k^{\′}=\frac{Y_k}{H_k(1+\frac{1}{{\left|H_k\right|}^2\mathrm{SNR}})}$

By above content as can be known, the channel equalization of traditional receiver and decoding are separately carried out: after receiver estimates the response of each subcarrier, the frequency domain data that sends on employing method 1 or method 2 each subcarrier of recovery, carry out soft decision demodulation then, decipher sending into the Viterbi decoder after the soft-decision-value deinterleaving again.In multipath channel, a part of subcarrier is serious decline, and it is 0 that its response amplitude is close to, and another part subcarrier decline degree is then lighter.The judgement that two kinds of different data that subcarrier transmitted of decline degree were done the Viterbi decoder when this channel equalization caused Viterbi decoding with decoding structure separately is identical, to such an extent as to the performance of receiver is bad.

Summary of the invention

At the existing problem and shortage of the channel equalization method of above-mentioned existing ofdm system, the purpose of this invention is to provide and a kind ofly improve the ofdm system receiver performance, channel equalization and Viterbi decoding is joined together to make the channel equalization method of ofdm system of entire system performance the best.

The present invention is achieved in that a kind of channel equalization method of ofdm system, may further comprise the steps:

1), after receiver receives time-domain signal, this signal is carried out the FFT conversion, obtains the subcarrier signals Y of each data in the frequency domain _k

2), obtain each sub-carrier signal Y by channel estimating _kFrequency domain response H _k

3), channel is carried out balanced signal to recover to send on each subcarrier, the signal after the equilibrium $X_k^{\′}=\frac{Y_k}{H_k};$

4), calculate the soft-decision-value U of each bit on each subcarrier of received signal _{K, i}, i.e. the log-likelihood ratio of this bit correspondence: $U_{k,i}=R\·\ln \frac{\mathrm{Pr}\{u_{k,i}=1/X_k^{\′}\}}{\mathrm{Pr}\{u_{k,i}=0/X_k^{\′}\}},$ Wherein R is a constant;

5), to each soft-decision-value U _{K, i}Be weighted: W _{K, i}=U _{K, i}| H _k| ²Again it is deciphered and promptly obtain final binary data.

Further, described channel estimating can be to carry out before the FFT conversion.

Further, described interpretation method is the Viterbi decoding algorithm.

Further, described soft-decision-value U _{K, i}Also can carry out U _{K, i}| H _k| weighting scheme.

Further, if described modulation system BPSK and QPSK then can carry out equilibrium to channel with following manner: with the signal Y that receives on each subcarrier _kMultiply by this subcarrier responses H _kConjugation, promptly $W_k=Y_k{H}_k^{*}.$

By above-mentioned steps as can be known this moment the Viterbi decoder input be the log-likelihood ratio of each subcarrier transmitted bit, and the Viterbi decoding algorithm is the maximum-likelihood decoding of convolution code, the inventive method is optimum in this sense.In multipath channel, when subcarrier declines, its amplitude is very little, and the bit that adopts weighting algorithm of the present invention to make to transmit on the bad subcarrier is very little to the influence of the judgement that the Viterbi decoder is done, thereby can improve the performance of system greatly.The present invention joins together the channel equalization and the decoding of receiver to consider, makes overall performance reach optimum, effectively raises the performance of system.

Description of drawings

Below in conjunction with accompanying drawing, the present invention is made detailed description.

Fig. 1 is existing ofdm system transmitter architecture schematic diagram;

Fig. 2 is existing ofdm system receiver structure schematic diagram;

Fig. 3 is a receiver structure schematic diagram of the present invention;

Fig. 4 is a receiver channel equalization methods flow chart of the present invention;

Fig. 5 is the receiver structure schematic diagram of BPSK of the present invention and QPSK modulation;

Fig. 6 is the receiver channel equalization methods flow chart of BPSK of the present invention and QPSK modulation;

Fig. 7 is the system BER performance schematic diagram of 6Mb/s for speed among the IEEE802.11a;

Fig. 8 is the system PER performance schematic diagram of 6Mb/s for speed among the IEEE802.11a;

Fig. 9 is the system BER performance schematic diagram of 12Mb/s for speed among the IEEE802.11a;

Figure 10 is the system PER performance schematic diagram of 12Mb/s for speed among the IEEE802.11a;

Figure 11 is the system BER performance schematic diagram of 24Mb/s for speed among the IEEE802.11a;

Figure 12 is the system PER performance schematic diagram of 24Mb/s for speed among the IEEE802.11a;

Figure 13 is the balanced hardware embodiment of an ofdm system receiver channel of the present invention schematic diagram;

Figure 14 is the balanced hardware embodiment of the ofdm system receiver channel schematic diagram of BPSK of the present invention and QPSK modulation.

Embodiment

As shown in Figure 3, receiver of the present invention is similar to existing receiver structure, and difference is that channel estimating is connected with weighting block, and carries | H _k| ²Information to weighting block.Here, the total number of sub-carriers of supposing to be used for transmitting useful data is N, k=0, and 1 ..., N-1.

As shown in Figure 4, specific implementation step of the present invention is, after signal receiver receives signal, this signal carried out the FFT conversion, obtains the subcarrier signals Y of each data in the frequency domain _kObtain each sub-carrier signal Y by channel estimating _kFrequency domain response H _kAnd signal carried out the data of direct phase division equilibrium to recover to send on each subcarrier, balanced back data $X_k^{\′}=\frac{Y_k}{H_k};$ k＝0，1，…，N-1。Then calculate the soft-decision-value U of each bit on each subcarrier of received signal _{K, i}, i.e. the log-likelihood function of this bit correspondence: $U_{k,i}=R\·\ln \frac{\mathrm{Pr}\{u_{k,i}=1/X_k^{\′}\}}{\mathrm{Pr}\{u_{k,i}=0/X_k^{\′}\}},$ Wherein R is a constant, and Pr{} represents to ask probabilistic operation.Then to each soft-decision-value U _{K, i}With | H _k| ²Be weighted: U _{K, i}| H _k| ²At last it is deciphered and promptly obtain final binary data.

Ofdm system channel equalization method for BPSK and QPSK modulation can also be reduced to top step: $W_k=Y_k{H}_k^{*},$ k＝0，1，…，N-1。Its receiver structure comprises Cyclic Prefix (CP), FFT, channel estimating, channel equalization, soft-decision, deinterleaving and Viterbi decoding as shown in Figure 5, and channel equalization partly adopts $W_k=Y_k{H}_k^{*},$ K=0,1 ..., N-1 gets final product.Its flow chart as shown in Figure 6, with the signal Y that receives on each subcarrier _kMultiply by this subcarrier responses H _kConjugation, promptly $W_k=Y_k{H}_k^{*},$ K=0,1 ..., N-1 carries out soft-decision then.

The also available another kind of weight coefficient of channel equalization method comes balanced, is specially W _k=Y _k| H _k|, k=0,1 ..., N-1.Because this weighting does not change the noise size of each subcarrier, and only having changed the phase place of noise, promptly is phase equalization (phase equalization).

The present invention has carried out emulation to the OFDM baseband system based on WLAN (wireless local area network) IEEE 802.11a agreement.Transmission rate adopts 6Mb/s, 12Mb/s and 24Mb/s, corresponds respectively to BPSK, QPSK and 16QAM modulation system.The code rate of convolution code is 1/2.Wireless channel adopts the frequency selectivity slow fading channel, and root mean square (rms) time delay is 75ns, and maximum delay is 10 times of rms time delay.The average power in each footpath is exponential damping, and every footpath whose amplitude obeys rayleigh distributed, and phase place is obeyed the even distribution of [0,2 π].The length of each frame is 600 bytes (Byte) in the emulation, and the hypothesis receiver can obtain the ideal channel response estimated result.

Arrive shown in Figure 12 as Fig. 7, provided the simulation result of the error rate (BER) and the frame error rate (PER) of 6Mb/s, 12Mb/s and these three kinds of speed of 24Mb/s respectively, and be that direct phase division, MMSE criterion, phase equalization method and method provided by the present invention compare four kinds of methods.As can be seen from the figure, the BER of algorithm proposed by the invention and PER are than MMSE algorithm, phase equalization and directly division is all good mutually, and along with the increase performance gap of order of modulation shows more obviously.

As shown in figure 13, present embodiment can adopt the IEEE802.11a WLAN standard to implement based on the CPLD chip EP20K400EBC652-2X of Alter company.The subcarrier responses of input remains unchanged in a frame, calculate simultaneously each subcarrier responses amplitude square, the two is saved in RAM the inside, when next frame arrives, refresh.Sub-carrier signal that receives and corresponding response signal are controlled by counter.The remaining equalization step according to ofdm system is finished and is got final product.

Ofdm system channel equalization method overall plan about BPSK and QPSK modulation of the present invention as shown in figure 14, present embodiment adopts IEEE 802.11a WLAN standard as developing example based on the CPLD chip EP20K400EBC652-2X of Alter company.The realization master clock is 60MHz, and the sub-carrier signal speed of input is 12MHz.The subcarrier responses of input remains unchanged in a frame, and it is saved in a RAM the inside, refreshes when next frame arrives.The sub-carrier signal and the corresponding response that receive are controlled by counter, finish equilibrium then in multiplier module.Because what adopt later is the Viterbi decoder of 4 bits, gets final product so export the data flow of 4 bits.For the BPSK modulation, only need the output solid part signal to get final product, corresponding conjugation multiplier module also only needs to calculate real part and gets final product.

Claims (5)

1, a kind of channel equalization method of ofdm system may further comprise the steps:

1), after receiver receives time-domain signal, this signal is carried out the FFT conversion, obtains the subcarrier signals Y of each data in the frequency domain _k, k=0,1 ..., N-1, N represent the number of sub carrier wave of transmitting useful data;

2), obtain each sub-carrier signal Y by channel estimating _kFrequency domain response H _k

3), channel is carried out balanced signal to recover to send on each subcarrier, the signal after the equilibrium $X_k^{\′}=\frac{Y_k}{H_k};$

4), calculate the soft-decision-value U of each bit on each subcarrier of received signal _{K, i '}The i.e. log-likelihood ratio of this bit correspondence: $U_{k,i}=R\·\ln \frac{\mathrm{Pr}\{u_{k,i}=1/X_k^{\′}\}}{\mathrm{Pr}\{u_{k,i}=0/X_k^{\′}\}},$ Wherein R is a constant, and Pr represents the value in { } is asked probabilistic operation, and i represents bit number;

5), to each soft-decision-value U _{K, i}Be weighted: W _{K, i}=U _{K, i}| H _k| ²Again it is deciphered and promptly obtain final binary data.

2, the channel equalization method of ofdm system as claimed in claim 1 is characterized in that, described channel estimating is carried out before the FFT conversion.

3, the channel equalization method of ofdm system as claimed in claim 1 is characterized in that, the described Viterbi that is decoded as deciphers.

4, the channel equalization method of ofdm system as claimed in claim 1 or 2 is characterized in that, described soft-decision-value U _{K, j}Carry out U _{K, i}| H _k| weighting scheme.

5, the channel equalization method of ofdm system as claimed in claim 1 or 2 is characterized in that, if modulation system BPSK and QPSK then carry out equilibrium with following weighting scheme to channel: with the signal Y that receives on each subcarrier _kMultiply by this subcarrier responses H _kConjugation, promptly $W_k=Y_k{H}_k^{*},$

Expression H _kConjugation.

Images