CN102437980A - Improved Bussgang adaptive frequency domain balance method - Google Patents
Improved Bussgang adaptive frequency domain balance method Download PDFInfo
- Publication number
- CN102437980A CN102437980A CN2011104452875A CN201110445287A CN102437980A CN 102437980 A CN102437980 A CN 102437980A CN 2011104452875 A CN2011104452875 A CN 2011104452875A CN 201110445287 A CN201110445287 A CN 201110445287A CN 102437980 A CN102437980 A CN 102437980A
- Authority
- CN
- China
- Prior art keywords
- signal
- frequency domain
- improved
- weight vector
- ussgang
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention discloses an improved Bussgang adaptive frequency domain balance method. By the method, a balance process can be performed in a frequency domain by adding a fast Fourier transformation (FFT) module at the receiving end of a system; a balance algorithm can be simplified by rationally selecting an initial value of a weight value of a balancer and setting an error signal, so that the convergence speed of the algorithm is increased; and the problems of complicity and low convergence speed of a blind balance detection algorithm in the prior art can be solved. In the method, only a few pilot sequences are needed, and the implementation process is simple; the method is applicable to all types of systems, comprising orthogonal frequency division multiplexing (OFDM) systems and non OFDM systems.
Description
Technical field
The present invention relates to communication technical field, relate in particular to a kind of Improved B ussgang adaptive frequency domain equalization methods based on least mean-square error.
Background technology
Fast development along with (B3G) behind the three generations or the 4th third-generation mobile communication system (4G) and Internet; The design object of future mobile communication system not only requires the raising of traffic rate; Also require to realize bigger power system capacity and better communication quality, and want to realize better in the world seamless roam and the multimedia service that comprises voice, data and image etc. is provided for the user.How to realize that this target has become the focus of world communication and the research of information science academia.Yet; Abominable wireless channel environment is restricting the demand for development of traffic rate and quality; The new technology of many physical layers constantly is studied and uses, and mainly contains: multi-transceiver technology, new encoding and decoding technique, adaptive coding modulation, multi-antenna technology, channel estimating and equilibrium etc. receive detection technique.
In order to alleviate or to overcome the influence of wireless channel environment to communication system, balancing technique is a kind of vital technology.Because it can eliminate the influence of channel to the transmission signal, thereby improve the performance of receiving terminal, so receive much concern always.The equalization methods that adopts at present mainly is divided into two big types: one type is to carry out channel estimating through training sequence, then signal is carried out equilibrium, and this method need be transmitted training sequence, has sacrificed the bandwidth resources of system; Another kind of is the blind equalization that does not need training sequence, and blind equalization has improved the valid data transmission rate of system, but complex algorithm, convergence rate is slower.Become a research focus to the combine half-blindness equalization algorithm that produced of these two class methods gradually, become the another class methods in the balancing technique.
In existing blind equalization research; Mainly be based on the blind equalization algorithm of steady statistic of second-order cyclic or high-order statistic, based on the blind equalization algorithm of neural network theory etc.; These algorithms are relatively complicated; And existing equalization methods carries out from time domain mostly, and frequency domain equalization mainly is used in the ofdm system.
Summary of the invention
In view of above-mentioned existing in prior technology problem, the purpose of this invention is to provide a kind of algorithm and more simply be easy to the blind balance method realized, thereby improve the balanced performance that detects in the wireless communication system, improve the performance of wireless communication system.
The present invention is through adding a FFT (fast Fourier transform) module at system receiving terminal; Thereby make balancing procedure to carry out at frequency domain; Again through the equaliser weights initial value rationally choose the setting with error signal, simplified equalization algorithm, accelerated convergence rate simultaneously.
The objective of the invention is to realize through following technical scheme:
A kind of Improved B ussgang adaptive frequency domain equalization methods comprises the steps:
(1) receiving terminal extracts frequency domain and receives signal, defines its relevant target function one least mean-square error and corresponding equaliser weights vector;
(2) utilize Improved B ussgang adaptive frequency domain equalization algorithm to recover detection signal;
Described step (1) comprising:
Using FFT (fast Fourier transform) module to obtain frequency domain at receiving terminal and receive signal, define the weight vector of frequency-domain equalizer then, is the expression formula that benchmark draws target function one least mean-square error with the difference between restoring signal and the primary signal.
Described step (2) comprising:
A, the balanced weight vector of initialization;
B, confirm current restoring signal according to the input signal of current equaliser weights and filter;
C, error signal, this error signal utilize the judgement sensing method of Bussgang algorithm to calculate, and promptly error signal is represented by restoring signal and to the difference that restoring signal carries out between the signal that symbol judgement obtains;
D, carry out the renewal of balanced weight vector according to corresponding relation, obtain the used weight vector of balancing procedure next time according to the current input vector of error signal and equalizer;
E, repeat above-mentioned steps B, C and D, realize the equilibrium of signal, recover primary signal.
Described Improved B ussgang adaptive frequency domain equalization methods, and described equalizer is frequency-domain equalizer, weight vector is the tap coefficient of frequency-domain equalizer.
Among the present invention, the initialization procedure of the described balanced weight vector of steps A comprises:
Utilize some known pilot signals;, gradient obtains optimal value when being zero according to target function; Calculate the expression formula of optimum weight vector, the statistical average in the optimum weights computational process (being mathematic expectaion) replaces with the arithmetic average of limited number of time, and the weight vector that calculates is as initial value.
Described step B comprises:
Balanced weight vector is received signal multiplication with corresponding frequency domain, then it is transformed to time domain, the signal that is restored, promptly
Restoring signal d (i)=W
H(i) Y (i) F;
Wherein, W (i) is balanced weight vector, and Y (i) receives signal for frequency domain, and F is the transformation matrix of IFFT (fast fourier inverse transformation), subscript
HThe expression conjugate transpose.
Described step C comprises:
Error signal e (i)=d (i)-sgn (d (i));
Wherein, described e (i) is current error signal, and d (i) is the restoring signal among the step B, and sgn () representes sign function.
Described step D comprises:
Balanced next time used filter weights vector W (i+1)=W (i)-μ Y (i) Fe
H(i);
Wherein, described W (i) is current weight vector, and μ is a step factor, can be a constant or by preferred variable someway, Y (i) is frequency domain reception signal, and e (i) is the error signal among the step C.
Described Improved B ussgang adaptive frequency domain equalization methods also comprises:
When system starts working; Transmitting terminal need send the known symbol of setting quantity as training sequence or pilot tone; Receiving terminal need utilize these frequency pilot signs to carry out the calculating of near-optimization weight vector, and the near-optimization weights that obtain are as the initial value of weight vector.
Among the present invention, the modulation system that adopts during this system works is BPSK (binary phase shift keying).Half-blindness equalization methods of the present invention simply is easy to realize, can transform the wireless communication system that is used for adopting the BPSK modulation, improves the balanced performance that detects in the wireless communication system, improves the performance of wireless communication system.
Improved B ussgang adaptive frequency domain equalization methods of the present invention is applicable to sorts of systems, comprises ofdm system and non-ofdm system.
Technical scheme by the invention described above provides can find out that the present invention makes balancing procedure to carry out at frequency domain through add a FFT module at system receiving terminal; Adopt improved LMS (least mean-square error) algorithm to carry out equilibrium to received signal based on Bussgang judgement sensing method; The setting with error signal of choosing through equaliser weights vector initial value; Make algorithm the convergence speed accelerate; Simultaneously algorithm is realized simply, and it is complicated and restrain slow problem to have solved in the prior art blind equalization algorithm; The present invention only needs a spot of pilot frequency sequence, and implementation procedure is simple.
Description of drawings
Fig. 1 is the structural representation of applicable system of the present invention;
Fig. 2 is the theory diagram of Improved B ussgang adaptive frequency domain equalization algorithm among the present invention.
Embodiment
Method of the present invention is a kind of Improved B ussgang adaptive frequency domain equalization methods; For realizing the present invention; Begin to carry out the balanced known symbol that needs before to send some earlier as training sequence in system; Just can get into the balancing procedure that iterates then, each iteration has comprised equalization filtering, Error Calculation and three steps of adaptive updates equaliser weights vector altogether.Carry out in the process of transfer of data in system, except initial a small amount of pilot frequency sequence, no longer need insert pilot tone or training symbol, have higher spectrum efficiency.
Existing embodiment to method of the present invention is in conjunction with the accompanying drawings:
System construction drawing is referring to Fig. 1; The data-signal of incoming bit stream representative input comprises the pilot frequency sequence of beginning and active traffic afterwards, then incoming bit stream is carried out the BPSK mapping; Be that each input bit is mapped to accordingly+1 or-1; Signal arrives receiving terminal through channel afterwards, and the FFT of receiving terminal (fast Fourier transform) module receives signal transformation to frequency domain with time domain, carries out equilibrium then, adjudicates, separates process such as mapping and recover original bit stream.Wherein, signal processing module and signal processing inverse process module are meant that this system can add other signal processing, handle as long as carry out corresponding inverse process at receiving terminal.
The concrete implementation procedure of equalization methods of the present invention comprises the steps:
Step 1: when initially getting into operating state in system; In order to use improved LMS (least mean-square error) algorithm to carry out equilibrium based on Bussgang judgement sensing method; At first need transmitting terminal to send length and be that the training sequence of L symbol, L are that (K carries out counting of FFT to K*N, and N is an integer; Usually less than 10) so that initial value W (the 0)=E of receiving terminal calculating filter weights [Y (i) FF
HY
H(i)]
-1E [Y (i) Fz
*(i)]=KE [Y (i) Y
H(i)]
-1E [Y (i) Fz
*(i)], wherein, Y (i) receives the diagonal matrix of signal as its diagonal element, [Y (i) Y for frequency domain
H(i)]
-1Be simple and easy to ask, F is the transformation matrix of normalized IFFT (fast fourier inverse transformation), and its dimension is K * K (K=2
n, n is positive integer such as value 0,1,2 as required), z (i) is a pilot signal, the mathematic expectaion E here carries out arithmetic average with the corresponding signal on all frequency pilot signs and is similar to and obtains.
Step 2: receiving terminal receives signal multiplication with balanced weight vector with corresponding frequency domain, then it is transformed to time domain, and signal is restored;
W (i) is current equaliser weights vector; Y (i) is the input of equalization filter, and as its diagonal element, dimension is K * K by K frequency domain reception signal value of K symbol before the current time; The weight vector dimension of equalization filter is K * 1, and i is the currency of iterations counter.Signal after the filtering should be from the frequency domain transform to the time domain in addition, and F is the transformation matrix of IFFT (fast fourier inverse transformation), and restoring signal is d (i)=W
H(i) Y (i) F.
Step 3: receiving terminal calculates error signal by restoring signal and difference that restoring signal is carried out between the signal that symbol judgement obtains;
Use judgement direction calculation error signal e (i)=d (the i)-sgn (d (i)) of Bussgang algorithm, d (i) is the restoring signal in the above-mentioned steps 2.
Step 4: referring to Fig. 2, through sef-adapting filter weights controlling mechanism filter weights is adjusted, the promptly adaptive adjustment of carrying out filter weights according to e (i) value is upgraded, to obtain to be used for the weight vector of equalization filtering next time;
Be specially the filter weights W that next time uses in the balancing procedure (i+1)=W (i)-μ Y (i) Fe
H(i), wherein, described W (i) is current equaliser weights, and μ is a step factor, can be a constant or by preferred variable someway.Y (i) is that corresponding frequency domain receives signal matrix.
Among the present invention; Only need to send L known symbol as training sequence in the training stage (being that system starts working the stage); And after data transfer phase no longer need insert pilot tone or training symbol, this moment, the input signal Y (i) of filter was that frequency domain receives the diagonal matrix of signal as its diagonal element.
Carry out in the normal data transmission procedure in system, method of the present invention mainly comprises three processing procedures: (1) is confirmed current signal equalization d (i) as a result with reference to above-mentioned steps 2 according to the input signal Y (i) of current filter weights W (i) and filter; (2) calculate the difference e (i) the current demand signal equilibrium result is got sign function with it after with reference to above-mentioned steps 3; (3) carry out the adjustment of filter weights with reference to above-mentioned steps 4.Circulation execution (1), (2), (3) three processes just can realize the equilibrium of signal in system.
Among the present invention, add various signal processing modes through the transmitting terminal in system, at receiving terminal various signal processing modes are carried out contrary the processing, Improved B ussgang adaptive frequency domain equalization methods provided by the invention also is suitable for for other system.
The above; Be merely the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, any technical staff who is familiar with the present technique field is in the technical scope that the present invention discloses; The variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection range of claims.
Claims (9)
1. an Improved B ussgang adaptive frequency domain equalization methods is characterized in that comprising the steps:
(1) receiving terminal extracts frequency domain and receives signal, defines its relevant target function-least mean-square error and corresponding equaliser weights vector;
(2) utilize Improved B ussgang adaptive frequency domain equalization algorithm to recover detection signal;
Described step (2) comprising:
A, the balanced weight vector of initialization;
B, confirm current restoring signal according to the input signal of current equaliser weights and filter;
C, error signal, this error signal utilize the judgement sensing method of Bussgang algorithm to calculate, and promptly error signal is represented by restoring signal and to the difference that restoring signal carries out between the signal that symbol judgement obtains;
D, carry out the renewal of balanced weight vector according to corresponding relation, obtain the used weight vector of balancing procedure next time according to the current input vector of error signal and equalizer;
E, repeat above-mentioned steps B, C and D, realize the equilibrium of signal, recover primary signal.
2. Improved B ussgang adaptive frequency domain equalization methods according to claim 1 is characterized in that described step (1) comprising:
Using the FFT module to obtain frequency domain at receiving terminal and receive signal, define the weight vector of frequency-domain equalizer then, is the expression formula that benchmark draws target function-least mean-square error with the difference between restoring signal and the primary signal.
3. Improved B ussgang adaptive frequency domain equalization methods according to claim 1 is characterized in that described equalizer is a frequency-domain equalizer, and weight vector is the tap coefficient of frequency-domain equalizer.
4. Improved B ussgang adaptive frequency domain equalization methods according to claim 1 is characterized in that the initialization procedure of the described balanced weight vector of steps A comprises:
Utilize some known pilot signals;, gradient obtains optimal value when being zero according to target function; Calculate the expression formula of optimum weight vector, the statistical average in the optimum weights computational process replaces with the arithmetic average of limited number of time, and the weight vector that calculates is as initial value.
5. Improved B ussgang adaptive frequency domain equalization methods according to claim 1 is characterized in that step B comprises:
Balanced weight vector is received signal multiplication with corresponding frequency domain, then it is transformed to time domain, the signal that is restored, promptly
Restoring signal d (i)=W
H(i) Y (i) F;
Wherein, W (i) is balanced weight vector, and Y (i) receives signal for frequency domain, and F is the transformation matrix of IFFT, subscript
HThe expression conjugate transpose.
6. Improved B ussgang adaptive frequency domain equalization methods according to claim 5 is characterized in that among the step C:
Error signal e (i)=d (i)-sgn (d (i));
Wherein, described e (i) is current error signal, and d (i) is the restoring signal among the step B, and sgn () representes sign function.
7. Improved B ussgang adaptive frequency domain equalization methods according to claim 1 is characterized in that step D comprises:
Balanced next time used filter weights vector W (i+1)=W (i)-μ F (i) Fe
H(i);
Wherein, described W (i) is current weight vector, and μ is a step factor, is a constant or preferred variable, and Y (i) receives signal for frequency domain, and e (i) is the error signal among the step C.
8. according to each described Improved B ussgang adaptive frequency domain equalization methods of claim 1 to 7, it is characterized in that this method also comprises:
When system started working, transmitting terminal need send the known symbol of setting quantity as training sequence or pilot tone, and receiving terminal need utilize these frequency pilot signs to carry out the calculating of balanced weight vector initial value.
9. according to each described Improved B ussgang adaptive frequency domain equalization methods of claim 1 to 7, it is characterized in that the modulation system that adopts during this system works is BPSK.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011104452875A CN102437980A (en) | 2011-12-28 | 2011-12-28 | Improved Bussgang adaptive frequency domain balance method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011104452875A CN102437980A (en) | 2011-12-28 | 2011-12-28 | Improved Bussgang adaptive frequency domain balance method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102437980A true CN102437980A (en) | 2012-05-02 |
Family
ID=45985861
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011104452875A Pending CN102437980A (en) | 2011-12-28 | 2011-12-28 | Improved Bussgang adaptive frequency domain balance method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102437980A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105138789A (en) * | 2015-09-08 | 2015-12-09 | 浪潮集团有限公司 | Non-linear system equivalent analysis method based on Bussgang theory |
| CN106338743A (en) * | 2016-11-11 | 2017-01-18 | 西安航天恒星科技实业(集团)公司 | Space-frequency combined self-adaptive zeroing algorithm based on CCD resolving |
| CN111342905A (en) * | 2018-12-18 | 2020-06-26 | 深圳市中兴微电子技术有限公司 | Signal processing method and device and computer storage medium |
| CN111935747A (en) * | 2020-08-17 | 2020-11-13 | 南昌航空大学 | Method for predicting link quality of wireless sensor network by adopting GRU (generalized regression Unit) |
-
2011
- 2011-12-28 CN CN2011104452875A patent/CN102437980A/en active Pending
Non-Patent Citations (1)
| Title |
|---|
| 居美艳,李岳衡，李黎: "预编码单载波系统中的简单半盲均衡方法", 《北京邮电大学学报》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105138789A (en) * | 2015-09-08 | 2015-12-09 | 浪潮集团有限公司 | Non-linear system equivalent analysis method based on Bussgang theory |
| CN106338743A (en) * | 2016-11-11 | 2017-01-18 | 西安航天恒星科技实业(集团)公司 | Space-frequency combined self-adaptive zeroing algorithm based on CCD resolving |
| CN106338743B (en) * | 2016-11-11 | 2019-05-21 | 西安航天恒星科技实业(集团)公司 | Combine adaptive nulling algorithm based on the null tone that CCD is resolved |
| CN111342905A (en) * | 2018-12-18 | 2020-06-26 | 深圳市中兴微电子技术有限公司 | Signal processing method and device and computer storage medium |
| CN111342905B (en) * | 2018-12-18 | 2022-11-04 | 深圳市中兴微电子技术有限公司 | Signal processing method and device and computer storage medium |
| CN111935747A (en) * | 2020-08-17 | 2020-11-13 | 南昌航空大学 | Method for predicting link quality of wireless sensor network by adopting GRU (generalized regression Unit) |
| CN111935747B (en) * | 2020-08-17 | 2021-04-27 | 南昌航空大学 | Method for predicting link quality of wireless sensor network by adopting GRU (generalized regression Unit) |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP1836817A2 (en) | System and method for utilizing different known guard intervals in single/multiple carrier communication systems | |
| CN102123115B (en) | Particle swarm optimization based orthogonal wavelet blind equalization method | |
| CN102624652B (en) | LDPC decoding method and apparatus, and receiving terminal | |
| CN101478510B (en) | Adaptive equalizer and receiver system using the equalizer | |
| EP1716681A1 (en) | Channel equalization | |
| CN105245478B (en) | A kind of adaptive equalization algorithm based on qam mode | |
| CN103873406B (en) | Underwater sound orthogonal FDM communication system inter-frame-interference removing method | |
| WO2013121958A1 (en) | Receiving device, receiving method and receiving program | |
| CN101409694A (en) | Method and system for implementing single carrier frequency domain equilibrium | |
| CN105306396A (en) | Optimization method for iterative equalization of wireless wideband communication channel | |
| CN102437980A (en) | Improved Bussgang adaptive frequency domain balance method | |
| WO2017183631A1 (en) | Los-mimo demodulation device, communication device, los-mimo transmission system, los-mimo demodulation method and program | |
| CN110311876A (en) | The implementation method of underwater sound OFDM receiver based on deep neural network | |
| CN112769726A (en) | Passive beam forming optimization method based on intelligent reflection surface auxiliary communication system | |
| Kari et al. | Adaptive and efficient nonlinear channel equalization for underwater acoustic communication | |
| CN101656697B (en) | Frequency-domain blind equalization method (T/2FF-CMA) based on T/2 fraction space | |
| Iqbal et al. | Adaptive Frequency-Domain RLS DFE for Uplink MIMO SC-FDMA | |
| WO2007111198A1 (en) | Transmission method and transmission device | |
| CN102685060B (en) | Multi-user multiple input multiple output (MIMO) receiving method and device for orthogonal frequency division multiplexing system | |
| CN103348622A (en) | Signal processing method and apparatus in microwave communication system | |
| CN105049095A (en) | SC-MIMO system double-flow tri-antenna or tetra-antenna diversity transmit-receive method and device | |
| CN102420790A (en) | Equalization processing method in single carrier ultra wide band communication system and system thereof | |
| JP4299302B2 (en) | Apparatus and method for precoding multi-carrier signals | |
| CN107959648A (en) | The design method of double prototype FBMC-OQAM systems Central Plains mode filter | |
| CN104135455A (en) | Iterative receiving method for communication system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20120502 |