CN101026430B - Method and system for removing interferences - Google Patents
Method and system for removing interferences Download PDFInfo
- Publication number
- CN101026430B CN101026430B CN 200610009189 CN200610009189A CN101026430B CN 101026430 B CN101026430 B CN 101026430B CN 200610009189 CN200610009189 CN 200610009189 CN 200610009189 A CN200610009189 A CN 200610009189A CN 101026430 B CN101026430 B CN 101026430B
- Authority
- CN
- China
- Prior art keywords
- symbol
- bit
- decoding
- decoder
- decode results
- 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.)
- Active
Links
Images
Landscapes
- Error Detection And Correction (AREA)
Abstract
The invention is in use for solving issue that current technique is unable to detect and remove paroxysmal interference. The method includes steps: (1) after receiving sequence to be decoded, the receiving end calculates measuration of each bit in each symbol of the sequence, and based on the measuration to add erasing marker for corresponding symbol; (2) erasing symbols of carrying the erasing marker, calculating the measuration of each bit in the erased symbols, and decoding the measuration; (3) based on the decoded result to determine whether it is needed to stop decoding; if yes, switching to step (4); otherwise, after adding erasing marker for corresponding symbol further based on decoded result, switching to step (2); (4) outputting decoded result.
Description
Technical field
The present invention relates to wireless communication field, particularly relate to a kind of method and system of removing interference.
Background technology
OFDM (OFDM) is proved to be a kind of technology very likely, and is the selection of many next generation communication systems.In OFDM, available bandwidth is divided into a series of orthogonal subcarriers on frequency domain.High speed signal can be on the subcarrier of a plurality of separation with than the low rate parallel transmission.By setting up a plurality of low speed subcarriers, the OFDM symbol period extends, and therefore reduces the intersymbol interference that is caused by multipath fading.
Because the communication resource is limited, a plurality of communication systems may be used identical frequency band and transmission medium.Many communication systems based on OFDM comprise that the communication system of wireless communication system, digital audio broadcasting and power line transmission all exists arrowband or part bandwidth interference.Ofdm system can be by avoiding reducing with the subcarrier that blocks the impact of arrowband or part bandwidth interference.This technology is effective under the channel condition known conditions.Yet sudden due to what disturb, transmitter may not know whether interference exists.Under this situation, interference may have a strong impact on the performance of system.
At present detect and remove to disturb two schemes is arranged:
Existing scheme one: in order to obtain acceptable bit error rate performance, interweaving and encoding of transmission information is essential.Given Noise and Interference distributes, and desirable Maximum Likelihood Detection can be finished the work.at document [Fazel94] (K.Fazel, " Narrow-band interference rejection in orthogonal multi-carrier spread-spectrumcommunications, " in Proceedings of Third Annual International Conference on Universal PersonalCommunications, September 1994, pp.46-50.) [PT_Bolinth04] (US 2004/0022175 A1, Bolinth et al., " Method and orthogonal frequency division multiplexing (OFDM) receiver for reducing the influence ofharmonic interference on OFDM transmission systems " .) in, the nil symbol that sends detects for receiver and disturbs the power that whether exists and estimate interference and noise.at document [Ghosh03] (M.Ghosh and V.Gaddam, " Bluetooth interference cancellation for 802.11g WLAN receivers, " in Proceedings of IEEE InternationalConference on Communications, vol.2, May 2003, pp.1169-1173.) [PT_Jones05] (US 2005/6973134 B1, Jones IV et al., " OFDM interference cancellation based on training symbol interference " .), Interference Estimation is based on the transmitted data symbols "ball-park" estimate.Document [PT_Oksanen03] (US 2003/6603743 B1, Oksanen etal., " Method for eliminating interference in an OFDM radio receiver " .) a kind of interference estimation method to constant enveloped modulation of proposition.No matter this thought hypothesis adjacent signals is in time-domain or frequency domain is not fast-changing, so the variation of amplitude has also just reflected the effect of disturbing.
Described optimal decoder based on Maximum Likelihood Detection, their defective is as follows.In document [Fazel94] [PT_Bolinth04], because pilots insertion is sparse, the same based on the Interference Detection scheme of pilot tone with all, interference may not can be fully detected.Yet for document [Ghosh03] [PT_Jones05], its accuracy may be transmitted error in data and estimate restriction, and the error in data that especially causes due to low signal-to-noise ratio (SNR) or deep fading or strong jamming is estimated.At last, in document [PT_Oksanen03], it has limited modulation scheme and can lose efficacy when the channel fading of considering the employing piece changes.
Existing scheme two: suboptimum wipe the estimation that decoder do not need to be considered to definite Noise and Interference power.This decoding scheme is wiped the signal that (ignoring) gets clogged in decode procedure, therefore avoided disturbing the adverse effect that causes.Do not need to know interference power although wipe decoder, it needs the information of interference position.document [Wong03] (K.K.Wong and T.O ' Farrell, " Coverage of 802.11g WLANs in thepresence of Bluetooth interference, " in Proceedings of IEEE International Symposium on Personal, Indoor andMobile Radio Communications, vol.3, September 2003, pp.2027-2031.) and [Aguado02] (L.E.Aguado, K.K.Wong and T.O ' Farrell, " Coexistence issues for 2.4GHz OFDM WLANS, " in Proceedings of ThirdInternational Conference on 3G Mobile Communication Technologies, May 2002, pp.400-404.) be evaluated at the performance of hypothesis interference position decoder when known.Document [PT_Laneman02] (US 2002/6430724 B1, Lanemanet al., " Soft selection combining based on successive erasures of frequency band components in acommunication system " .) use the special construction that the first neighboring trace FM disturbs also successfully to carry out wiping decoding in the digital audio broadcasting sideband.Under more general and practical condition, interference detector need to provide erasure information for decoder.
The major defect that described suboptimum is wiped decoder is interference detector.This is because can not detect bursty interference fully as described as the detection scheme based on pilot tone of mentioning in [PT_Saleh91] (US 1991/5048057, Saleh et al., " Wireless local area network " .).Undetected interference may have a strong impact on systematic function.
In sum, prior art need to know that relevant interfere information just can complete Interference Detection, and can't detect bursty interference fully, makes undetected interference cause systematic function and seriously influences.
Summary of the invention
The invention provides a kind of method and system of removing interference, can't detect and remove the problem of bursty interference in order to solve prior art fully.
Further solve, prior art need to know that relevant interfere information just can complete Interference Detection and remove the problem of disturbing.
The inventive method comprises the following steps:
After A, receiving terminal are received sequence to be decoded, calculate the tolerance of each bit of each symbol in this sequence, and wipe mark for corresponding symbol adds accordingly;
B, wipe and carry the symbol of wiping mark, calculate the tolerance of each bit in the symbol that is wiped free of, and this tolerance is carried out decoding;
C, judge whether that according to decode results needs stop decoding, if change step D over to; Otherwise, be further after mark is wiped in corresponding symbol interpolation, to change step B over to according to decode results;
D, output decode results.
In described steps A, if the tolerance of the one or more bits on symbol greater than default threshold value, is wiped mark for this symbol interpolation.
Before carrying out decoding in described step B, first judge whether to exceed default decoding iterations, if do not exceed, carry out decoding; Otherwise, directly change step D over to.
If do not exceed default decoding iterations, before decoding, first described each bit that carries the symbol of wiping mark is carried out deinterleaving.
In the symbol that described calculating is wiped free of, the tolerance of each bit adopts
Wherein
The tolerance that represents each bit of each symbol in described symbol sebolic addressing, Ek represents to carry the erase status of the symbol of wiping mark, and k represents the label of symbol, and m represents a corresponding amount of bits of symbol, and j is illustrated in a bit sequence in symbol.
Decoding described in step B comprises the following steps:
The bit lattice shape figure of the state transitions during-structure convolutional encoding is used for generating state transitions and the output bit sequence of symbol sebolic addressing at cataloged procedure that be encoded;
-structure bit is wiped indication lattice shape figure, and the bit that is used for a state transitions of expression decoding is wiped distribution;
-with described bit lattice shape figure with wipe indication lattice shape figure and multiply each other, obtain product grid shape figure, expression has the state transitions of the convolutional encoding that some bits wipe, and is wiped free of bit distribution;
-decoder is wiped under number and corresponding bits erasure location condition given, seeks the shortest path of algorithm that carries the bit of wiping mark in described product grid shape figure, and determines corresponding maximum likelihood code word with this.
Described step C comprises the following steps:
-obtain the path metric difference values of current definite every two maximum likelihood code words;
The absolute value of-each path metric difference values that will obtain and default threshold value relatively if absolute difference less than threshold value occurs, stop decoding, and change step D over to; Otherwise, be further after mark is wiped in corresponding symbol interpolation, to change step B over to according to decode results.
Further wipe mark for corresponding symbol adds according to decode results described in step C, comprise the following steps:
-search the maximum path of bit that should wipe in each shortest path;
-wipe the bit that should wipe in this path;
-in each symbol that this path is passed through, if on a symbol, the one or more bits that should wipe are arranged, for adding, this symbol wipes mark.
In described step D, the decode results that does not exceed described default decoding iterations output is to satisfy in two maximum likelihood code words of absolute value less than threshold value of path metric difference values, the maximum likelihood code word that the bit that should wipe is more;
The decode results that exceeds described default decoding iterations output is any one to take turns in decoding, the maximum maximum likelihood code word of bit that should wipe in each maximum likelihood code word of determining.
The system that removal of the present invention is disturbed comprises:
Demodulator is used for calculating bit metric according to the symbol sebolic addressing that receives, and wipes mark according to described bit metric for corresponding symbol adds, and output carries the bit metric of the symbol of wiping mark;
Decoder is used for the bit metric that receives is carried out decoding, and the output decode results;
The second judging unit is used for the decode results of receiver decoder, and judges whether accordingly to need to stop decoding, if need to continue decoding, indicates decoder that decode results is fed back to demodulator; Otherwise the indication decoder is externally exported decode results.
Described system also comprises: decoding iteration judging unit, and the decoding iterations for judging whether to reach default if reach, stops the decoding of decoder, and the indication decoder is externally exported decode results; Otherwise, continue decoding.
Described system also comprises: deinterleaver, be used for the bit metric of demodulator output is carried out deinterleaving, and re-send to decoder and carry out decoding.
Described system also comprises: interleaver is used for decoder is carried out interleaving treatment to the decode results of demodulator feedback.
The symbol sebolic addressing that described demodulator receives comes from the decode results that channel is exported and decoder feeds back after front-end processing; Perhaps, the decode results of channel estimating output and decoder feedback; Perhaps, the decode results of the output of interference detector and decoder feedback.
Beneficial effect of the present invention is as follows:
The present invention progressively removes the symbol of disturbed obstruction by circulation decoding, and judges by adequacy, determines whether to finish decoding.Namely obtain the path metric difference values of current definite every two maximum likelihood code words by product grid shape figure; With the absolute value of each path metric difference values of obtaining and default threshold value relatively, if absolute difference less than threshold value occurs, stop decoding; Otherwise, will have the maximum likelihood code word that should wipe at most bit and feed back to decoder, enter next round and process.
By enforcement of the present invention, can better detect and remove bursty interference, and need not to know relevant interfere information.
Description of drawings
Fig. 1 is system configuration schematic diagram of the present invention;
Fig. 2 is the inventive method flow chart of steps;
Fig. 3 (a) expression has 0 and 1 the indication lattice type of wiping figure that wipes;
Bit lattice type figure during Fig. 3 (b) expression 1/2 code check 2 state convolutional encoding;
The product grid type figure that Fig. 3 (c) expression has 0 and 1 1/2 code check 2 state convolutional encoding of wiping;
Fig. 4 is being for to illustrate traditional folding coding, optimum maximum-likelihood decoding and the adequacy criterion error rate (BER) performance based on path metric difference values of decoding of the present invention under disturbance quantity;
Fig. 5 is being for to illustrate traditional folding coding, optimum maximum-likelihood decoding and the adequacy criterion error rate (BER) performance based on path metric difference values of decoding of the present invention under different SIR;
Fig. 6 be under varying number disturbs based on the path metric difference threshold η Dec of the performance adequacy decision rule impact on bit error rate (BER);
Fig. 7 be under different SIR based on the path metric difference threshold η Dec of the performance adequacy decision rule impact on bit error rate (BER);
Fig. 8 exists and does not exist under the channel estimation errors situation, traditional folding coding, and optimum maximum-likelihood decoding and decoding of the present invention is the error rate (BER) performance at CRC with under based on the adequacy criterion of path metric difference values;
Fig. 9 exists and does not exist under the channel estimation errors situation, traditional folding coding, and optimum maximum-likelihood decoding and decoding of the present invention is word error probability (WER) performance at CRC with under based on the adequacy criterion of path metric difference values.
Embodiment
The invention provides a kind of system of removing interference, can be applicable to but be not limited to OFDM, in the situation that has noise and/or interference in system, be used for better detecting and removing bursty interference.Shown in Figure 1, it comprises: interconnective demodulator and decoder, and the second judging unit that is connected with described decoder; Further also can comprise: the decoding iteration judging unit that is connected with described decoder is connected in the deinterleaver between described demodulator and decoder, and is connected in the interleaver between described demodulator and decoder.
Described demodulator is used for calculating bit metric according to the symbol sebolic addressing that receives, and wipes mark according to described bit metric for corresponding symbol adds, and output carries the bit metric of the symbol of wiping mark.The symbol sebolic addressing that demodulator receives comes from the decode results that channel is exported and decoder feeds back after front-end processing; Perhaps, the decode results of channel estimating output and decoder feedback; Perhaps, the decode results of the output of interference detector and decoder feedback.
Described decoding iteration judging unit, the decoding iterations for judging whether to reach default if reach, stops the decoding of decoder, and the indication decoder is externally exported decode results; Otherwise, continue decoding.
Described decoder is used for the bit metric that receives is carried out decoding, and the output decode results.
Described the second judging unit is used for the decode results of receiver decoder, and judges whether accordingly to need to stop decoding, if need to continue decoding, indicates decoder that decode results is fed back to demodulator; Otherwise the indication decoder is externally exported decode results.
Described deinterleaver is used for the bit metric of demodulator output is carried out deinterleaving, re-sends to decoder and carries out decoding.
Described interleaver is used for decoder is carried out interleaving treatment to the decode results of demodulator feedback.Thereby purpose is the disturbing code bit makes the bit that gets clogged be separated from each other.Various types of interleavers can with in the present invention, comprise block interleaver, convolutional deinterleaver and random interleaver.
Use said system, the invention provides a kind of method of removing interference, shown in Figure 2, comprise the following steps:
The tolerance of S1, calculating bit, and wipe corresponding symbol.
The input of the demodulator of receiving terminal comprises the decode results that channel output after front-end processing and decoder feed back.
The output of channel after described front-end processing, come from that transmitting terminal sends by the information sequence of convolutional encoding and Bit Interleave.If the encoder output codons is C=[c
1, c
2..., c
N] (c
i∈ 0,1}) and the sequence D after interweaving=[d
1, d
2..., d
N], i.e. sequence C displacement.The interweaving encoding sequence of every m bit can be mapped to M-ary (M=2 according to a mapping equation μ
m) symbol in planisphere.X namely
k=μ ([d
(k-1) m+1, d
(k-1) m+2..., d
km]) (k=1,2 ..., N/m).Then the symbol sebolic addressing after the modulation is by sending in noisy noisy communication channel.To ofdm system, the symbol after modulation generally transmits by subcarrier on falt fading channel.The disturbed obstruction of some subcarriers possibilities, and other meeting is added into the background noise of Gaussian Profile.Through after the FFT of receiving terminal, k symbol yk that receives can be expressed as y
k=α
kx
k+ n
k, α here
kFading factor and n
kIt is additive noise.Each noise samples can be white gaussian noise or high-power interference.The present invention is applied to general situation, not any special distribution of hypothesis interference.Interference may be blocked continuous receiving symbol also may random barrage reception symbol.
After the demodulator of receiving terminal is received the symbol sebolic addressing that has used coding structure, carry out following operation:
1, calculate the bit metric of exporting based on channel
The first step of demodulation is to adopt conventional method to calculate bit metric, as document [Caire98] (G.Caire, G.Taricco, and E.Biglieri, " Bit-interleaved coded modulation, " IEEE Transactions on Information Theory, vol.44, pp.927-945, May 1998.).To each modulation code sequence y that receives, calculate the tolerance of each bit: d
(k-1) m+j=b (b=0,1; J=1,2 ... m; K=1,2 ..., be N/m)
Here
The estimated value of the channel fading factor,
The j position that is a k symbol is the signal subset of b.
2, symbol is wiped mark
Demodulator is wiped according to the alignment marks symbol that calculates in (1).If corresponding to the tolerance of any m bit of k symbol higher than default threshold value η
Dem, k symbol just is marked as to be wiped.Demodulator also can be wiped the cue mark symbol based on decoder feedback symbol after previous round decoding and wipe.If any m bit of k symbol is marked as wipe, k symbol will be marked as to be wiped.Make E
k∈ 0,1} (k=1,2 ..., be N/m) that wiping of k symbol indicated and e
i∈ 0,1} (i=1,2 ..., N) feed back after previous round decoding for decoder bit wipe indication.Wipe mark for these two kinds, null representation is wiped.Decoder comes label symbol to wipe according to following formula.
Before first round decoding, to all i=1,2 ..., N establishes e
i=1.
Wipe mark and can utilize the correlation of blocking symbol.Give an example, if code word at a bag by a plurality of OFDM block movements, may cause a plurality of obstruction symbols on occluder carrier wave.Any one symbol under this situation on the subcarrier is marked as to be wiped, and all symbols relevant with this subcarrier all are wiped free of being marked as when wiping.In some cases, the bandwidth of interference has occupied several adjacent subcarriers simultaneously, and demodulator may utilize this class feature to come erasure of marks.
3, give the bit metric of decoder
Then demodulator wipes indication (2) to decoder output bit metric according to tolerance formula (1) and symbol.If k symbol is marked as and wipes, also will be marked as with k all relevant m bits of symbol and wipe.Namely, each d of k symbol
(k-1) m+j=b (b=0,1; J=1,2 ... position m), the bit metric of exporting to decoder is
Wherein
The tolerance that represents each bit of each symbol in described symbol sebolic addressing, E
kThe erase status of expression symbol, k represents the label of symbol, and m represents a corresponding amount of bits of symbol, and j is illustrated in a bit sequence in symbol.
Before the bit metric with demodulator output outputs to decoder, first utilize deinterleaver to carry out deinterleaving.
S2, judge whether to exceed default decoding iterations, if do not exceed, change step S3 over to; Otherwise, directly change step S5 over to.
When having bit metric to output to decoder, first whether exceed default decoding iterations by decoding iteration judgment unit judges.
S3, decoding.
When given bit metric standard, decoder can be found the individual maximum likelihood code word of K (K 〉=1), and each code word has 0,1 ..., K-1 added bit wiped mark.The maximum likelihood code word of for example, wiping mark with 0 and 1 added bit minimum decoding tolerance in theory can be respectively
With
In actual embodiment, for convolution code, minimization problem can in lattice type figure, shortest path solves by finding.There are a lot of methods to realize.At [Li03] T.Li, W.H.Mow, and M.Siu, " Serial JEVA forefficient decoding in impulsive noise channels; " in Proceedings of IEEE Semiannual Vehicular TechnologyConference, vol.1, October 2003, proposed a kind of possible implementation in pp.308-312..As selection, the product grid type figure that shortest path can also be constructed as follows obtains.
1, by insert the individual intermediateness of L-1 (L is the number of coded bits of each lattice branch) between two of adjacent compartment connected states, expansion convolutional encoding lattice type figure is bit lattice type figure.Be labeled as like this coded-bit of a correspondence in each branch of bit lattice type figure.Represent from them on the meaning of same code word set, bit lattice type figure is equivalent to original coding lattice type figure.
2, wiping of i bit indicated with a binary character e
iRepresent, 0 means that i bit is wiped free of, 1 other meaning of expression.And use erase counters ε
iCan also calculate from beginning to be marked as the sum of wiping bit to i bit process.Erase counters ε
iOnly rely on its previous value ε
i-1With current indicating device e
i, it is non-decreasing.The counter sequence is the same with indicator sequence can be represented by indication lattice type figure, erase counters ε
iBe expressed as state, wipe indicating device e
iIndicated to represent by branch.Each paths in lattice type figure is corresponding possible indicator sequence e all.Suppose that the total number of symbols that is labeled is restricted to K-1, indicate so lattice type figure just to comprise K state, namely ε
i∈ 0,1 ..., K-1}.As an example, the lattice type when Fig. 3 (a) has described K=2.The state of two correspondences is respectively ε
i=0,1.
3, product grid type figure can obtain by the bit lattice type figure in step 1 and the indication lattice type product in step 2.The i layer of supposing bit lattice type figure has comprised N
iIndividual state q
i 1, q
i 2...,
The i layer of indication lattice type figure has comprised
Individual state ε
i 1, ε
i 2...,
The product of these two lattice types has comprised at the i layer
Individual state, each state is
(j=1 ..., N
i,,
) right.Q in and if only if bit lattice type figure
i-1 lAnd q
i lCan identify c with branch
i, connect, and in indication lattice type figure
With
Two states can identify e with branch
iWhen connecting, two states of adjacent layer
With
Can be with (a c
i, e
i) branch identifies to connect.Corresponding branch metric can be defined as:
φ(c
i，e
i)＝λ(c
i)·e
i (6)
Path metric is the summation of branch metric.Each paths in product grid type figure is a corresponding specific coded sequence and indicator sequence all.Done state is that (path of 0≤k＜K) has k code word of additionally wiping to erase counters k.Therefore, the path correspondence of shortest path tolerance the maximum likelihood code word of wiping mark corresponding to having arranged.
As an example, the example of Fig. 3 has illustrated that the 2 state convolution codes that possess 0 and 1 1/2 code check of wiping mark form the process of product grid type figure.Fig. 3 (a) has showed and has wiped indication lattice type figure.In Fig. 3 (b), insert an intermediateness and form bit lattice type figure between every pair of connection status.Fig. 3 (c) has illustrated that product grid type figure can obtain by bit lattice type figure and indication lattice type figure.
4, obtain maximum likelihood code word in product grid type figure.Given product grid is sought the most reliable path with Viterbi Viterbi algorithm.Have 0,1 ..., the shortest path of wiping mark of K-1 (K 〉=1) can obtain simultaneously or successively according to realizing, and then correspondence obtains a series of maximum likelihood code words.
S4, judge whether that according to decode results needs stop decoding, if change step S5 over to; Otherwise, be further after mark is wiped in corresponding symbol interpolation, to change step S1 over to according to decode results.
The maximum likelihood code word (being decode results) that the second judging unit is exported according to decoder, and utilize a kind of error correcting code, as CRC or other block code, can be used as adequacy decision rule.In case a candidate codewords has satisfied error checking, decode procedure is terminated so.Corresponding code word just will be chosen as output codons.
Adequacy decision rule is independent of extra error checking and correction.After each was taken turns decoding and completes, the path metric difference values of new candidate codewords can be wiped code word from zero and be begun order computation.If the absolute value of difference is than default threshold value η
DecLittle, decode procedure just stops so.(adequacy decision rule comprises error check code and based on the method for path metric difference values.) according to the product grid type figure of Fig. 3 (c), viterbi algorithm can find to locate to terminate in state [a, 0] and [a, 1] two shortest paths simultaneously.Article two, the road is through corresponding two maximum likelihood code words, and one of them code word of two code words another many one of code word meeting is relatively wiped indication.
If the absolute value of difference is than default threshold value η
DecGreatly, continue decoding, have so wiping indicator sequence and will being fed to demodulator by interleaver of maximum erasure number purpose maximum likelihood code word.Demodulator will be wiped indication update bit tolerance according to new bit according to formula (2) and (3).Namely return to step S1.
S5, output decode results.
The decode results that does not exceed described default decoding iterations output is to satisfy in two maximum likelihood code words of absolute value less than threshold value of path metric difference values, the maximum likelihood code word that the bit that should wipe is more.Namely be connected step S4.
The decode results that exceeds described default decoding iterations output is any one to take turns in decoding, the maximum maximum likelihood code word of bit that should wipe in each maximum likelihood code word of determining.Namely be connected step S2.
Below, Calculation Simulation proves robustness and the validity of this invention.The convolution code that 1/2 code check 64 states use 16QAM is investigated in all emulation.Being used for basic decoding scheme relatively is traditional decoding scheme and optimum maximum likelihood decoding scheme.Traditional decoding scheme refers to one for not having noisy white Gauss noise design.
In the simulation result of Fig. 4-7, each code word is mapped on an OFDM piece that comprises 864 subcarriers.Multidiameter fading channel has six paths, and corresponding average power is [0-7-15-22-24-19] dB, and corresponding time delay is [0 38 11 13 21]
μ sFading factor is assumed to be received machine and predicts fully.The thresholding of demodulator is η
Dem=∞.Adequacy decision rule based on path metric difference values is used.
Figure 4 shows that traditional decoding scheme, the decoding scheme that proposes in optimum maximum-likelihood decoding scheme and the present invention is at SIR=0dB, bit error rate (BER) performance under SNR=20dB and different number disturb.Figure 5 shows that three kinds of decoding schemes at SNR=20dB, 5 BER performances that are disturbed under subcarrier and different SIR.Observe from Fig. 4 and Fig. 5, can find that the present invention program significantly is better than traditional decoding scheme, and for different numbers be disturbed subcarrier and on a large scale interference power can obtain actual optimal decoding performance.
In Fig. 4 and Fig. 5 use is for thresholding η best based on the adequacy decision rule of path metric difference values
Dec, Fig. 6 and Fig. 7 use function η
Dec/ σ
2Estimate the BER performance, here σ
2It is background Gaussian noise variance.Can see by setting thresholding at a suitable value range and just can obtain best error rate.And optimum threshold value can't occur significantly to change along with the number that is disturbed subcarrier or different interference powers.In fact, best thresholding is only to rely on background Gaussian noise power, can determine by off-line like this.Therefore, the adequacy decision rule based on path metric difference values is healthy and strong and feasible in reality to different interference environments.
In Fig. 8 and Fig. 9, each code word is mapped on 10 OFDM pieces that 200 subcarriers are arranged.Thresholding is η
Dem=η
Dec=26 σ
2Channel fading for each subcarrier is random the generation.Decoding scheme is assessed in two kinds of situations of channel estimation errors channel estimation errors being arranged and do not have.The channel estimator that adopts is frequency domain LS estimator, two pilot tones on each subcarrier.Interference for each subcarrier produces at random with probability 0.04, and power evenly distributes between [20,10] dB.Adequacy decision rule based on path metric difference values and CRC is used together.Specify, if any one of two criterions is satisfied, decode procedure all will stop.The Generator polynomial of CRC is 435 (octal system).Zero quilt is filled up in the back of CRC code word makes convolution code be terminated.Because information sequence is very long, the frequency spectrum loss that causes due to CRC and zero padding can be left in the basket.
Fig. 8 and Fig. 9 have represented traditional decoding scheme, the BER of optimum maximum-likelihood decoding scheme and decoding scheme of the present invention and error code word rate (WER).In order to show the performance gain that utilizes code structure to obtain, assess two kinds of possible realizations of the present invention.Sign " Proposed 1 " represents that only utilizing demodulator to export based on channel carries out the scheme that symbol is wiped.The scheme that sign " Proposed 2 " representative utilizes demodulator and decoder to wipe.In scheme " Proposed 1 ", decoder is the same with traditional Viterbi decoder not to be wiped.In this scheme, wipe and do not utilize code structure, can be counted as individual wiping and decoding scheme of separating.The gap of " Proposed 1 " and " Proposed 2 " has been showed to unite and has been wiped decoding with respect to the performance gain of separately wiping decoding.Solid line has represented the performance that the accurate channel fading factor is arranged.Can see that traditional decoding scheme almost lost efficacy.The decoding scheme that we propose will obviously be better than traditional scheme, especially when high SNR.In addition, utilized and unite " Proposed 2 " of wiping decoding and obviously be better than " Proposed 1 " and obtain almost optimum solution code performance in the whole SNR scope of examining or check.These results show that demodulator can only utilize channel output to wipe the symbol of the disturbed obstruction of part.Utilize code structure and use to unite and wipe decoding, almost the symbol of whole disturbed obstructions all can be removed.
In these figure, the decoding scheme that we propose and traditional decoding scheme are having the evaluation form under channel estimation errors to be shown dotted line.Can find, the performance loss of the scheme " Proposed 2 " that we propose for BER in 3dB, for WER in 2dB.Scheme " Proposed 1 " is because channel estimation errors shows larger performance loss.For WER=0.1, performance loss surpasses 4dB.Proposed by the invention the uniting of this explanation wiped decoding scheme not only for the interference of Unknown Distribution but also be all healthy and strong when the channel fading evaluated error is arranged.The decoding scheme that the computational complexity of scheme " Proposed 1 " is basic and traditional is similar.The computational complexity of scheme " Proposed 2 " is for interested error rate scope, such as WER less than 0.1, within several times of traditional decoding scheme.For instance, when WER=0.01,1.5 times of the decoding scheme that proposes the chances are under desirable channel estimating traditional scheme.Due to channel estimation errors, 3 times of left and right of the complexity of calculating the chances are traditional scheme.In addition, complexity descends along with the rising of SNR.
In a word, need not interfere information, the decoding scheme that we propose obviously is better than traditional decoding scheme, and can obtain to utilize the just maximum-likelihood decoding performance of available optimum of complete interference statistics.For a system that autonomous retransmission requirement (ARQ) arranged, the throughput of the decoding scheme that proposes can be very large raising system.Cost is to have increased the complexity of calculating.For low channel estimation errors and high SNR, computational complexity and traditional decoding scheme are very less than being.And the decoding scheme that proposes can be used for and other interference detector cooperates.Under the help of interference detector, the complexity of the decoding scheme of proposition can further reduce.
Obviously, those skilled in the art can carry out various changes and modification and not break away from the spirit and scope of the present invention the present invention.Like this, if within of the present invention these are revised and modification belongs to the scope of claim of the present invention and equivalent technologies thereof, the present invention also is intended to comprise these changes and modification interior.
Claims (15)
1. a method of removing interference, is characterized in that, comprises the following steps:
After A, receiving terminal are received sequence to be decoded, calculate the tolerance of each bit of each symbol in this sequence, and wipe mark for corresponding symbol adds accordingly;
B, wipe and carry the symbol of wiping mark, calculate the tolerance of each bit in the symbol that is wiped free of, and this tolerance is carried out decoding;
C, judge whether that according to decode results needs stop decoding, if change step D over to; Otherwise, be further after mark is wiped in corresponding symbol interpolation, to change step B over to according to decode results;
D, output decode results.
2. the method for claim 1, is characterized in that, in described steps A, if the tolerance of the one or more bits on symbol greater than default threshold value, is wiped mark for this symbol interpolation.
3. the method for claim 1, is characterized in that, before carrying out decoding in described step B, first judges whether to exceed default decoding iterations, if do not exceed, carries out decoding; Otherwise, directly change step D over to.
4. method as claimed in claim 3, is characterized in that, if do not exceed default decoding iterations, before decoding, first described each bit that carries the symbol of wiping mark carried out deinterleaving.
5. the method for claim 1, is characterized in that, in the symbol that described calculating is wiped free of, the tolerance of each bit adopts
Wherein
The tolerance of each bit of each symbol in the expression symbol sebolic addressing, E
kExpression carries the erase status of the symbol of wiping mark, and k represents the label of symbol, and m represents a corresponding amount of bits of symbol, and j is illustrated in a bit sequence in symbol.
6. method as claimed in claim 3, is characterized in that, the decoding described in step B comprises the following steps:
The bit lattice shape figure of the state transitions during-structure convolutional encoding is used for generating state transitions and the output bit sequence of symbol sebolic addressing at cataloged procedure that be encoded;
-structure bit is wiped indication lattice shape figure, and the bit that is used for a state transitions of expression decoding is wiped distribution;
-with described bit lattice shape figure with wipe indication lattice shape figure and multiply each other, obtain product grid shape figure, expression has the state transitions of the convolutional encoding that some bits wipe, and is wiped free of bit distribution;
-decoder is wiped under number and corresponding bits erasure location condition given, seeks the shortest path of algorithm that carries the bit of wiping mark in described product grid shape figure, and determines corresponding maximum likelihood code word with this.
7. method as claimed in claim 6, is characterized in that, described step C comprises the following steps:
-obtain the path metric difference values of current definite every two maximum likelihood code words;
The absolute value of-each path metric difference values that will obtain and default threshold value relatively if absolute difference less than threshold value occurs, stop decoding, and change step D over to; Otherwise, be further after mark is wiped in corresponding symbol interpolation, to change step B over to according to decode results.
8. method as claimed in claim 7, is characterized in that, further wipes mark for corresponding symbol adds according to decode results described in step C, comprises the following steps:
-search the maximum path of bit that should wipe in each shortest path;
-wipe the bit that should wipe in this path;
-in each symbol that this path is passed through, if on a symbol, the one or more bits that should wipe are arranged, for adding, this symbol wipes mark.
9. method as claimed in claim 8, is characterized in that, utilizes the correlation of blocking symbol, further wipes mark for corresponding symbol adds.
10. method as claimed in claim 7, it is characterized in that, in described step D, the decode results that does not exceed described default decoding iterations output is to satisfy in two maximum likelihood code words of absolute value less than threshold value of path metric difference values, the maximum likelihood code word that the bit that should wipe is more;
The decode results that exceeds described default decoding iterations output is arbitrary one to take turns in decoding, the maximum maximum likelihood code word of bit that should wipe in each maximum likelihood code word of determining.
11. a system of removing interference is characterized in that, comprising:
Demodulator is used for calculating bit metric according to the symbol sebolic addressing that receives, and wipes mark according to described bit metric for corresponding symbol adds, and output carries the bit metric of the symbol of wiping mark;
Decoder is used for the bit metric that receives is carried out decoding, and the output decode results;
The second judging unit is used for the decode results of receiver decoder, and judges whether accordingly to need to stop decoding, if need to continue decoding, indicates decoder that decode results is fed back to demodulator; Otherwise the indication decoder is externally exported decode results.
12. system as claimed in claim 11 is characterized in that, described system also comprises:
Decoding iteration judging unit, the decoding iterations for judging whether to reach default if reach, stops the decoding of decoder, and the indication decoder is externally exported decode results; Otherwise, continue decoding.
13. system as described in claim 11 or 12 is characterized in that, described system also comprises:
Deinterleaver is used for the bit metric of demodulator output is carried out deinterleaving, re-sends to decoder and carries out decoding.
14. system as claimed in claim 13 is characterized in that, described system also comprises:
Interleaver is used for decoder is carried out interleaving treatment to the decode results of demodulator feedback.
15. system as claimed in claim 13 is characterized in that, the symbol sebolic addressing that described demodulator receives comes from the decode results that channel is exported and decoder feeds back after front-end processing;
Perhaps, the decode results of channel estimating output and decoder feedback;
Perhaps, the decode results of the output of interference detector and decoder feedback.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610009189 CN101026430B (en) | 2006-02-20 | 2006-02-20 | Method and system for removing interferences |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610009189 CN101026430B (en) | 2006-02-20 | 2006-02-20 | Method and system for removing interferences |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101026430A CN101026430A (en) | 2007-08-29 |
| CN101026430B true CN101026430B (en) | 2013-11-06 |
Family
ID=38744386
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200610009189 Active CN101026430B (en) | 2006-02-20 | 2006-02-20 | Method and system for removing interferences |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101026430B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102780540B (en) * | 2012-06-26 | 2015-01-07 | 华为技术有限公司 | Erasure decoding method, device and receiver |
| CN102882651B (en) * | 2012-09-27 | 2015-04-29 | 华为技术有限公司 | Cyclic redundancy check method and cyclic redundancy check device |
| WO2014107899A1 (en) * | 2013-01-14 | 2014-07-17 | 华为技术有限公司 | Decoding method and device |
| CN108347298B (en) * | 2017-01-24 | 2021-01-15 | 华为技术有限公司 | Coding method and communication device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1181663A (en) * | 1996-10-18 | 1998-05-13 | 三菱电机株式会社 | Connecting code error correcting coding device |
| US6081919A (en) * | 1996-12-27 | 2000-06-27 | Nec Corporation | Coding and decoding system using CRC check bit |
| US6430724B1 (en) * | 1999-05-28 | 2002-08-06 | Agere Systems Guardian Corp. | Soft selection combining based on successive erasures of frequency band components in a communication system |
| WO2005099100A1 (en) * | 2004-03-22 | 2005-10-20 | Matsushita Electric Industrial Co., Ltd. | Local erasure map decoder |
-
2006
- 2006-02-20 CN CN 200610009189 patent/CN101026430B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1181663A (en) * | 1996-10-18 | 1998-05-13 | 三菱电机株式会社 | Connecting code error correcting coding device |
| US6081919A (en) * | 1996-12-27 | 2000-06-27 | Nec Corporation | Coding and decoding system using CRC check bit |
| US6430724B1 (en) * | 1999-05-28 | 2002-08-06 | Agere Systems Guardian Corp. | Soft selection combining based on successive erasures of frequency band components in a communication system |
| WO2005099100A1 (en) * | 2004-03-22 | 2005-10-20 | Matsushita Electric Industrial Co., Ltd. | Local erasure map decoder |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101026430A (en) | 2007-08-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101020461B1 (en) | Method and apparatus for channel quality measurements | |
| CN102195672B (en) | OFDM (Orthogonal Frequency Division Multiplexing) system inter-cell interference elimination method based on IDMA (Interleave-Division Mutiple-Access) | |
| EP1798920B1 (en) | Iterative detection and decoding in MIMO communication systems | |
| US20060172716A1 (en) | Ofdm reception device and ofdm reception method | |
| US7702050B2 (en) | Method and system for an adaptive VBLAST receiver for wireless multiple input multiple output (MIMO) detection | |
| CN102404257B (en) | Narrow-band interference detecting method and device in MIMO-OFDM (Multiple Input Multiple Output-Orthogonal Frequency Division Multiplexing) system | |
| EP3457605B1 (en) | Selective sub-carrier processing | |
| US7724853B2 (en) | Enabling mobile switched antennas | |
| CN101026430B (en) | Method and system for removing interferences | |
| JP5024197B2 (en) | Soft decision value correction method and receiving apparatus | |
| US7876864B2 (en) | Method and device for enhancing signal detection in a frequency selective fading channel | |
| US20070165757A1 (en) | Method and system for an improved cellular interference cancelling diversity receiver | |
| US7653117B2 (en) | Method for decoding digital data in a frequency hopping communication system | |
| US7809090B2 (en) | Blind data rate identification for enhanced receivers | |
| CN110445554B (en) | Non-coherent underwater acoustic communication method and system based on actual channel fading statistics | |
| CN103188176B (en) | The method of single-frequency or arrowband interference effect is reduced in ofdm system | |
| CN101668330B (en) | Signal receiving device, signal receiving method and global mobile communication system telephone | |
| CN103188003B (en) | Based on method for low complexity parallel interference cancellation and the system of iterative detection | |
| US20080049821A1 (en) | Modulation scheme deciding apparatus, receiving apparatus, modulation scheme deciding method and modulation scheme deciding program | |
| Jung et al. | Receiving colliding lora packets with hard information iterative decoding | |
| KR100747595B1 (en) | Apparatus and method for softly detecting symbol in multiple-input/multiple-output communication system | |
| Li et al. | Joint erasure marking and list Viterbi algorithm for decoding in unknown non-Gaussian noise | |
| Zhu et al. | A new HF radio prototype based on dynamic spectrum anti-jamming concept | |
| EP1885079B1 (en) | Method for measuring radio channel quality and related equipment | |
| CN117978206A (en) | Combined decoding method and system for diversity copy transmission in PLC 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 | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20211223 Address after: 450046 Floor 9, building 1, Zhengshang Boya Plaza, Longzihu wisdom Island, Zhengdong New Area, Zhengzhou City, Henan Province Patentee after: Super fusion Digital Technology Co.,Ltd. Address before: 518129 Bantian HUAWEI headquarters office building, Longgang District, Guangdong, Shenzhen Patentee before: HUAWEI TECHNOLOGIES Co.,Ltd. |
|
| TR01 | Transfer of patent right |