CN101026430B - Method and system for removing interferences - Google Patents

Abstract

The invention discloses a method and system for removing interference, which are used to solve the problem that the prior art cannot completely detect and remove sudden interference. The method of the present invention includes: A. After receiving the sequence to be decoded, the receiving end calculates the measure of each bit of each symbol in the sequence, and accordingly adds an erasure mark to the corresponding symbol; B. erasure carries an erasure Marked symbols, calculate the metric of each bit in the erased symbol, and decode the metric; C, judge whether the decoding needs to be terminated according to the decoding result, if so, then proceed to step D; otherwise, according to the decoding As a result, after further adding an erasure mark to the corresponding symbol, turn to step B; D, and output the decoding result.

Description

A method and system for removing interference

technical field

The present invention relates to the field of wireless communication, in particular to a method and system for removing interference.

Background technique

Orthogonal Frequency Division Multiplexing (OFDM) is proving to be a very promising technology and is the choice for many next generation communication systems. In OFDM, the available bandwidth is divided into a series of subcarriers that are orthogonal to each other in the frequency domain. High-speed signals can be transmitted in parallel at lower rates on multiple separated subcarriers. By establishing multiple low-speed subcarriers, the OFDM symbol period is extended, thereby reducing inter-symbol interference caused by multipath fading.

Because communication resources are limited, multiple communication systems may use the same frequency band and transmission medium. Many OFDM-based communication systems, including wireless communication systems, digital audio broadcasting and power line transmission communication systems, have narrowband or partial bandwidth interference. OFDM systems can reduce the impact of narrowband or partial bandwidth interference by avoiding the use of blocked subcarriers. This technique is effective when the channel condition is known. However, due to the bursty nature of the interference, the transmitter may not know if the interference is present. In this case, interference can seriously affect the performance of the system.

There are currently two solutions for detecting and removing interference:

Existing solution 1: In order to obtain an acceptable bit error rate performance, interleaving and coding of transmitted information are necessary. Given the noise and interference distributions, ideal maximum likelihood detection does the job. In the literature [Fazel94] (K.Fazel, "Narrow-band interference rejection in orthogonal multi-carrier spread-spectrum communications," in Proceedings of Third Annual International Conference on Universal Personal Communications, September 1994, pp.46-50.) [4]_Bolinth (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"), the transmitted zero symbols are used by the receiver to detect whether there is interference and Estimate the power of interference and noise. In the literature [Ghosh03] (M.Ghosh and V.Gaddam, "Bluetooth interference cancellation for 802.11g WLAN receivers," in Proceedings of IEEE International Conference 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".), the interference estimation is based on a rough estimate of the transmitted data symbols. The literature [PT_Oksanen03] (US 2003/6603743 B1, Oksanen et al., "Method for eliminating interference in an OFDM radio receiver".) proposes an interference estimation method for constant envelope modulation. This idea assumes that adjacent signals do not change rapidly in either the time domain or the frequency domain, so changes in amplitude reflect the effect of interference.

The above-mentioned optimal decoders based on maximum likelihood detection have the following defects. In [Fazel94][PT_Bolinth04], because the pilot insertion is sparse, as with all pilot-based interference detection schemes, the interference may not be adequately detected. However, for the literature [Ghosh03][PT_Jones05], its accuracy may be limited by misestimation of transmitted data, especially due to low signal-to-noise ratio (SNR) or data misestimation caused by deep fading or strong interference. Finally, in [PT_Oksanen03], it limits the modulation scheme and fails when considering channel fading variations with blocks.

Existing Solution 2: A sub-optimal erasure decoder is considered not to require exact noise and interference power estimation. This decoding scheme erases (ignores) blocked signals during the decoding process, thus avoiding adverse effects caused by interference. Although the erasure decoder does not need to know the interference power, it needs the information of the interference location. Literature [Wong03] (K.K.Wong and T.O'Farrell, "Coverage of 802.11g WLANs in the presence of Bluetooth interference," in Proceedings of IEEE International Symposium on Personal, Indoor and Mobile 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.) Evaluate the performance of the decoder assuming that the location of the interference is known. Document [PT_Laneman02] (US 2002/6430724 B1, Laneman et al., "Soft selection combining based on successful erasures of frequency band components in acommunication system".) Use the special structure of the first adjacent channel FM interference in the digital audio broadcasting sideband And erasure decoding is successfully performed. Under more general and practical conditions, the jammer detector needs to provide erasure information to the decoder.

The main drawback of the sub-optimal erasure coder is the jammer detector. This is because pilot based detection schemes as mentioned in [PT_Saleh91] (US 1991/5048057, Saleh et al., "Wireless local area network".) cannot fully detect bursty interference as described. Undetected interference can seriously affect system performance.

To sum up, the existing technology needs to know relevant interference information to complete interference detection, and burst interference cannot be completely detected, so that undetected interference has a serious impact on system performance.

Contents of the invention

The invention provides a method and system for removing interference, which is used to solve the problem that the prior art cannot completely detect and remove sudden interference.

To further solve the problem that in the prior art, relevant interference information needs to be obtained to complete interference detection and interference removal.

The inventive method comprises the following steps:

A. After receiving the sequence to be decoded, the receiving end calculates the metric of each bit of each symbol in the sequence, and adds an erasure mark to the corresponding symbol accordingly;

B. Erase the symbol carrying the erasure mark, calculate the metric of each bit in the erased symbol, and decode the metric;

C. Judging whether the decoding needs to be terminated according to the decoding result, if so, proceed to step D; otherwise, after further adding an erasure mark to the corresponding symbol according to the decoding result, proceed to step B;

D. Output the decoding result.

In the step A, if the measure of one or more bits on a symbol is greater than a preset threshold value, an erasure mark is added to the symbol.

Before decoding in the step B, it is first judged whether the preset number of decoding iterations is exceeded, if not, the decoding is performed; otherwise, directly proceed to step D.

If the preset number of decoding iterations is not exceeded, deinterleaving is performed on each bit of the symbol carrying the erasure flag before decoding.

其中

表示所述符号序列中每个符号的各个比特的度量，Ek表示携带有擦除标记的符号的擦除状态，k表示符号的标号，m表示一个符号相应的比特数量，j表示在一个符号中的比特序号。The calculation of the metric for each bit in the erased symbol uses

in

Represents the measure of each bit of each symbol in the symbol sequence, Ek represents the erasure state of the symbol carrying the erasure mark, k represents the label of the symbol, m represents the number of bits corresponding to a symbol, and j represents the number of bits in a symbol bit sequence number.

The decoding described in step B comprises the following steps:

- Construct a bit trellis diagram of the state transition during convolutional encoding, which is used to generate the state transition and output bit sequence of the encoded symbol sequence during the encoding process;

- Constructing a bit erasure indication trellis diagram, which is used to represent the bit erasure distribution of a state transition in decoding;

- multiplying said bit trellis by the erasure indication trellis to obtain a product trellis representing the state transitions of a convolutional code with a certain number of bit erasures, and the distribution of erased bits;

-The decoder searches for the algorithmic shortest path of the bit carrying the erasure mark in the product trellis diagram under the given erasure number and the corresponding bit erasure position condition, and determines the corresponding maximum likelihood accordingly Codeword.

Described step C comprises the following steps:

- Obtain the path metric difference value of every two maximum likelihood codewords currently determined;

-Comparing the absolute value of the obtained path metric difference with the preset threshold value, if there is an absolute value of difference smaller than the threshold value, then stop decoding and go to step D; otherwise, according to the decoding result After further adding erasure marks to the corresponding symbols, go to step B.

Further adding an erasure mark to the corresponding symbol according to the decoding result described in step C includes the following steps:

- finding the path with the most bits that should be erased among the shortest paths;

- erase the bits that should be erased in this path;

- Among the symbols traversed by the path, if a symbol has one or more bits that should be erased, add an erasure flag to the symbol.

In the step D, the decoding result that does not exceed the preset number of decoding iterations is the bit that should be erased in the two maximum likelihood codewords whose absolute value of the path metric difference is less than the threshold value More maximum likelihood codewords;

The decoding result output beyond the preset number of decoding iterations is the maximum likelihood codeword with the most bits to be erased among the determined maximum likelihood codewords in any round of decoding.

The interference removal system of the present invention includes:

a demodulator, configured to calculate a bit metric according to the received symbol sequence, and add an erasure mark to a corresponding symbol according to the bit metric, and output the bit metric of the symbol carrying the erasure mark;

a decoder, configured to decode the received bit metrics and output a decoding result;

The second judging unit is used to receive the decoding result of the decoder, and judge accordingly whether the decoding needs to be terminated, and if it needs to continue decoding, then instruct the decoder to feed back the decoding result to the demodulator; otherwise, instruct The decoder outputs the decoding result to the outside.

The system also includes: a decoding iteration judging unit for judging whether the preset number of decoding iterations has been reached, and if it has reached, the decoding of the decoder is terminated, and the decoder is instructed to output the decoding result; otherwise , continue decoding.

The system also includes: a deinterleaver, used for deinterleaving the bit metrics output by the demodulator, and then sending them to the decoder for decoding.

The system further includes: an interleaver, configured to interleave the decoding result fed back from the decoder to the demodulator.

The symbol sequence received by the demodulator comes from the channel output after front-end processing and the decoding result fed back by the decoder; or, the channel estimation output and the decoding result fed back by the decoder; or, the output of the interference detector and the decoding result fed back by the decoder.

The beneficial effects of the present invention are as follows:

The present invention gradually removes the symbols blocked by interference through circular decoding, and decides whether to end the decoding through adequacy judgment. That is, the path metric difference value of each two maximum likelihood codewords currently determined is obtained through the product trellis diagram; the absolute value of each path metric difference value obtained is compared with the preset threshold value, and if it is smaller than the threshold value If the absolute value of the difference is , the decoding is terminated; otherwise, the maximum likelihood codeword with the most bits to be erased is fed back to the decoder to enter the next round of processing.

Through the implementation of the present invention, sudden interference can be better detected and eliminated without knowing relevant interference information.

Description of drawings

Fig. 1 is a schematic structural diagram of the system of the present invention;

Fig. 2 is a flowchart of the method steps of the present invention;

Figure 3(a) shows the erasure indicator trellis with 0 and 1 erasures;

Fig. 3 (b) represents the bit trellis diagram during 1/2 code rate 2 state convolutional coding;

Figure 3(c) shows a product lattice diagram of 1/2 rate 2-state convolutional coding with 0 and 1 erasures;

Fig. 4 is the bit error rate (BER) performance based on the adequacy criterion of path metric difference under different interference quantities in illustrating traditional convolutional decoding, optimal maximum likelihood decoding and decoding of the present invention;

Fig. 5 is the bit error rate (BER) performance based on the adequacy criterion of the path metric difference under different SIRs illustrating traditional convolutional decoding, optimal maximum likelihood decoding and decoding of the present invention;

Fig. 6 is the impact of the path metric difference threshold ηDec based on the performance adequacy decision criterion under different amounts of interference on the bit error rate (BER);

Fig. 7 is the impact of the path metric difference threshold ηDec based on the performance adequacy decision criterion under different SIRs on the bit error rate (BER);

Fig. 8 is in the presence and absence of channel estimation errors, traditional convolutional decoding, optimal maximum likelihood decoding and the decoding of the present invention under CRC and the adequacy criterion based on path metric difference rate (BER) performance;

Fig. 9 is in the presence and absence of channel estimation errors, traditional convolutional decoding, optimal maximum likelihood decoding and the decoding of the present invention under the CRC and the error based on the sufficiency criterion of the path metric difference rate (WER) performance.

Detailed ways

The present invention provides a system for removing interference, which can be applied to but not limited to OFDM, and is used to better detect and remove sudden interference when there is noise and/or interference in the system. Referring to Fig. 1, it includes: a demodulator and a decoder connected to each other, and a second judging unit connected to the decoder; further comprising: a decoding iteration connected to the decoder A judging unit, a deinterleaver connected between the demodulator and the decoder, and an interleaver connected between the demodulator and the decoder.

The demodulator is configured to calculate a bit metric according to the received symbol sequence, add an erasure mark to a corresponding symbol according to the bit metric, and output the bit metric of the symbol carrying the erasure mark. The symbol sequence received by the demodulator comes from the channel output after front-end processing and the decoding result fed back by the decoder; or, the channel estimation output and the decoding result fed back by the decoder; or, the output of the interference detector and the decoding result The decoding result fed back by the encoder.

The decoding iteration judging unit is used to judge whether the preset number of decoding iterations has been reached, and if it has been reached, the decoding of the decoder is terminated, and the decoder is instructed to output the decoding result to the outside; otherwise, the decoding is continued .

The decoder is configured to decode the received bit metric and output a decoding result.

The second judging unit is used to receive the decoding result of the decoder, and judge accordingly whether the decoding needs to be terminated, and if it needs to continue decoding, instruct the decoder to feed back the decoding result to the demodulator; otherwise , instructing the decoder to output the decoding result externally.

The deinterleaver is used for deinterleaving the bit metrics output by the demodulator, and then sending them to the decoder for decoding.

The interleaver is configured to interleave the decoding result fed back from the decoder to the demodulator. The purpose is to scramble the code bits so that the jammed bits are separated from each other. Various types of interleavers can be used in the present invention, including block interleavers, convolutional interleavers, and random interleavers.

Applying the above system, the present invention provides a method for removing interference, as shown in Figure 2, comprising the following steps:

S1. Calculate the metric of the bit, and erase the corresponding symbol.

The input of the demodulator at the receiving end includes the channel output processed by the front end and the decoding result fed back by the decoder.

The channel output processed by the front end comes from the convolutionally encoded and bit-interleaved information sequence sent by the transmitter. Suppose the encoder output codeword is C=[c ₁ ,c ₂ ,...,c _N ](c _i ∈{0,1}) and the interleaved sequence D=[d ₁ ,d ₂ ,... , d _N ], that is, a permutation of sequence C. Each m-bit interleaved coding sequence is mapped to a symbol in the M-ary (M=2 ^m ) constellation diagram according to a mapping equation μ. That is x _k =μ([d _(k-1)m+1 , d _(k-1)m+2 ,...,d _km ])(k=1, 2,..., N/m ). The modulated symbol sequence is then sent through the interfering noisy channel. For OFDM systems, the modulated symbols are generally transmitted through subcarriers on flat fading channels. Some subcarriers may be blocked by interference, while others will be added with Gaussian distributed background noise. After the FFT at the receiving end, the k-th received symbol yk can be expressed as y _k =α _k x _k +n _k , where α _k is the fading factor and _nk is the additive noise. Each noise sample can be white Gaussian noise or high power interference. The invention applies to the general case and does not assume any particular distribution of interference. Interference may block consecutive received symbols or may block received symbols randomly.

After receiving the symbol sequence using the encoding structure, the demodulator at the receiving end performs the following operations:

1. Calculate the bit metric based on the channel output

The first step in demodulation is to calculate the bit metric using traditional methods, such as the literature [Caire98] (G.Caire, G.Taricco, and E.Biglieri, "Bit-interleaved coded modulation," IEEE Transactions on Information Theory, vol.44 , pp. 927-945, May 1998.). For each received modulation code sequence y, calculate the metric for each bit: d _(k-1)m+j = b(b=0,1; j=1,2,...m; k=1 , 2,..., N/m) is

$\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; \&Element;</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; b</span>}}^{\mathrm{<span\; class="notranslate">\; j</span>}}}{\mathrm{<span\; class="notranslate">\; min</span>}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span>{{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}}_{\mathrm{<span\; class="notranslate">\; k</span>}}\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

是信道衰落因子的估计值，

是一个第k个符号的第j位是b的信号子集。here

is the estimated value of the channel fading factor,

is a signal subset of the kth symbol whose jth bit is b.

2. Symbol erasure mark

The demodulator marks symbol erasures according to the metric computed in (1). If any m-bit metric corresponding to the k-th symbol is higher than a preset threshold η _Dem , the k-th symbol is marked as erased. The demodulator will also mark symbol erasure based on the symbol erasure indication fed back by the decoder after the previous round of decoding. The kth symbol is marked as erased if any m bits of the kth symbol are marked as erased. Let E _k ∈ {0, 1} (k = 1, 2, ..., N/m) be the erasure indication of the k-th symbol and e _i ∈ {0, 1} (i = 1, 2, . .., N) is the bit erasure indication fed back by the decoder after the previous round of decoding. For both erasure flags, zero indicates erasure. The decoder marks symbol erasures according to the following equation.

${\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; if</span>}<span\; class="notranslate">\; \&ForAll;</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; \&Element;</span><span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; 0,1</span>}<span\; class="notranslate">\; \}</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span><span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; \}</span>\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; .</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; .</span>\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ></span>{\mathrm{<span\; class="notranslate">\; \&eta;</span>}}_{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="Dem"\; filenames="H06109189620060306E000023.png,H06109189620060306E000031.png"\; state="\{\{state\}\}">Dem</figure-callout></span>}}\\ \mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; if</span>}<span\; class="notranslate">\; \&ForAll;</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span><span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; \}</span>\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; .</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; .</span>{\mathrm{<span\; class="notranslate">\; e</span>}}_{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; otherwise</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Before the first round of decoding, set e _i =1 for all i=1, 2, . . . , N.

Erasure marking can take advantage of the correlation of blocking symbols. For example, if a codeword is transmitted by multiple OFDM blocks in a packet, a blocked subcarrier may result in multiple blocked symbols. In this case, when any symbol on a subcarrier is marked as erasure, all symbols related to this subcarrier are erased when marked as erasure. In some cases, the bandwidth of the interferer occupies several adjacent subcarriers at the same time, and the demodulator may take advantage of such characteristics to mark erasures.

3. Bit metrics for the decoder

The demodulator then outputs bit metrics to the decoder according to the metric formula (1) and the symbol erasure indication (2). If the k-th symbol is marked as erased, all m bits associated with the k-th symbol will also be marked as erased. That is, each bit of d _(k-1)m+j = b (b=0,1; j=1,2,...m) of the k-th symbol, the bit metric output to the decoder for

$\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; E.</span>}}_{{\mathrm{<span\; class="notranslate">\; k</span>}}^{<span\; class="notranslate">\; \&Center\; Dot;</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

表示所述符号序列中每个符号的各个比特的度量，E_k表示符号的擦除状态，k表示符号的标号，m表示一个符号相应的比特数量，j表示在一个符号中的比特序号。in

Represents the metric of each bit of each symbol in the symbol sequence, E _k represents the erasure state of the symbol, k represents the label of the symbol, m represents the number of bits corresponding to a symbol, and j represents the bit sequence number in a symbol.

Before the bit metrics output by the demodulator are output to the decoder, the deinterleaver is used for deinterleaving.

S2. Judging whether the preset number of decoding iterations is exceeded, if not, proceed to step S3; otherwise, directly proceed to step S5.

When a bit metric is output to the decoder, the decoding iteration judging unit first judges whether the preset number of decoding iterations is exceeded.

S3, decoding.

Given a bit metric, the decoder can find K (K ≥ 1) maximum likelihood codewords, each with 0, 1, ..., K-1 additional bit erasure markers. For example, the theoretical minimum decoding metrics for maximum likelihood codewords with 0 and 1 additional bit erasure markers can be respectively

${\mathrm{<span\; class="notranslate">\; \&psi;</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

and

${\mathrm{<span\; class="notranslate">\; \&psi;</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; \&Element;</span><span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; 1,2</span>}<span\; class="notranslate">\; ,</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; \}</span>}{\mathrm{<span\; class="notranslate">\; min</span>}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; \&NotEqual;</span>\mathrm{<span\; class="notranslate">\; l</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; c</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

In a practical implementation, for convolutional codes, the minimization problem can be solved by finding the shortest path in the trellis graph. There are many ways to achieve this. Proposed in [Li03] T.Li, W.H.Mow, and M.Siu, "Serial JEVA forefficient decoding in impulsive noise channels," in Proceedings of IEEE Semiannual Vehicular Technology Conference, vol.1, October 2003, pp.308-312. a possible implementation. Alternatively, the shortest path can also be obtained through a product lattice graph constructed in the following steps.

1. By inserting L-1 (L is the number of coded bits of each trellis branch) intermediate states between two connected states of adjacent trellis layers, the convolutional coded trellis graph is expanded into a bit trellis graph. Each branch of the bit trellis is thus marked with a corresponding coded bit. The bit trellis is equivalent to the original coding trellis in the sense that they represent the same set of codewords.

2. The erasure indication of the i-th bit is represented by a binary symbol e _i , 0 means that the i-th bit is erased, and 1 means other meanings. Moreover, the erasure counter ε _i can also be used to calculate the total number of bits marked as erasure from the beginning to the i-th bit. The erasure counter ε _i depends only on its previous value ε _i-1 and the current pointer e _i , which is non-decrementing. The counter sequence, like the instruction sequence, can be represented by an instruction lattice diagram, the erasure counter ε _i is represented as a state, and the erasure indicator _ei is represented by a branch instruction. Each path in the trellis corresponds to a possible instruction sequence e. Assuming that the total number of marked symbols is limited to K-1, then the indicator lattice diagram contains K states, that is, ε _i ∈ {0, 1, ..., K-1}. As an example, Figure 3(a) depicts the lattice pattern when K=2. The two corresponding states are ε _i =0, 1 respectively.

指示格型图的第i层包含了

个状态ε_i ¹，ε_i ²，...，这两个格型的乘积在第i层包含了

个状态，每个状态是

(j＝1，...，N_i，，

)对。当且仅当比特格型图中的q_i-1 ^l和q_i ^l可以用分支标识c_i，来连接，且指示格型图中的

和

两个状态可以用分支标识e_i来连接时，相邻层的两个状态

和

可以用一个(c_i，e_i)分支标识来连接。相应的分支度量可以定义为：3. The product trellis can be obtained by multiplying the bit trellis in step 1 and the indicator trellis in step 2. Assume that the i-th layer of the bitmap contains N _i states q _i ¹ , q _i ² ,...,

Indicates that layer i of the trellis contains

states ε _i ¹ , ε _i ² ,..., The product of these two lattice types at level i contains

states, each of which is

(j=1, . . . , N _i ,

)right. If and only if q _i-1 ^l and q _i ^l in the bit trellis diagram can be connected by branch label c _i , and indicate that

and

When two states can be connected by branch identifier e _i , two states of adjacent layers

and

Can be connected with a ( _ci , e _i ) branch identifier. The corresponding branch metric can be defined as:

φ(c _i , e _i )=λ(c _i )·e _i (6)

The path metric is the sum of the branch metrics. Each path in the product lattice corresponds to a specific coding sequence and instruction sequence. A path with an end state of erasure counter k (0≦k<K) has k additional erased codewords. Therefore, the path with the shortest path metric corresponds to a maximum likelihood codeword with an erasure flag.

As an example, the example in Fig. 3 illustrates the process of forming a product trellis graph from a rate 1/2 2-state convolutional code with 0 and 1 erasure marks. Figure 3(a) shows the erasure indicator lattice diagram. In Figure 3(b), an intermediate state is inserted between each pair of connected states to form a bit lattice diagram. Figure 3(c) illustrates that the product trellis can be obtained from the bit trellis and the indicator trellis.

4. Obtain the maximum likelihood codeword in the product lattice graph. Given a product lattice, use the Viterbi algorithm to find the most reliable path. The shortest paths of erasure marks with 0, 1, ..., K-1 (K≥1) can be obtained simultaneously or sequentially according to the implementation, and then correspondingly obtain a series of maximum likelihood codewords.

S4. Determine whether the decoding needs to be terminated according to the decoding result, and if so, proceed to step S5; otherwise, further add an erasure mark to the corresponding symbol according to the decoding result, and then proceed to step S1.

The second judging unit uses the maximum likelihood codeword (ie, the decoding result) output by the decoder and uses an error correction code, such as CRC or other block codes, which can be used as adequacy judging criteria. Once a candidate codeword satisfies the error check, the decoding process is terminated. The corresponding codeword will be selected as the output codeword.

Adequacy decision criteria are independent of additional error checking. After each round of decoding is completed, the path metric difference values of the new candidate codewords are sequentially calculated starting from zero erasure codewords. If the absolute value of the difference is smaller than the preset threshold η _Dec , the decoding process is terminated. (Adequacy decision criteria include error-checking codes and methods based on path metric differences.) According to the product lattice diagram in Figure 3(c), the Viterbi algorithm can simultaneously find ] terminates the two shortest paths. The two paths correspond to two maximum likelihood codewords, and one of the two codewords has an additional erasure indication relative to the other codeword.

If the absolute value of the difference is larger than the preset threshold value η _Dec , continue decoding, then the erasure instruction sequence of the maximum likelihood codeword with the maximum number of erasures will be fed back to the demodulator through the interleaver . The demodulator will update the bit metrics according to the new bit erasure indication according to formulas (2) and (3). That is, return to step S1.

S5. Outputting the decoding result.

The decoding result that does not exceed the preset number of decoding iterations is the maximum likelihood codeword with more bits that should be erased among the two maximum likelihood codewords whose absolute value of the path metric difference is less than the threshold value. Codeword. That is, step S4 is continued.

The decoding result output beyond the preset number of decoding iterations is the maximum likelihood codeword with the most bits to be erased among the determined maximum likelihood codewords in any round of decoding. That is, step S2 is continued.

In the following, computational simulations demonstrate the robustness and effectiveness of the invention. All simulations examine rate 1/2 64-state convolutional codes using 16QAM. The basic decoding schemes used for comparison are the traditional decoding scheme and the optimal maximum likelihood decoding scheme. A conventional decoding scheme refers to a design for white Gaussian noise without interference.

In the simulation results in Fig. 4-7, each codeword is mapped to an OFDM block containing 864 subcarriers. There are six paths in a multipath fading channel, the corresponding average power is [0 -7 -15 -22 -24-19]dB, and the corresponding delay is [0 3 8 11 13 21] ^μs . The fading factors are assumed to be fully known by the receiver. The threshold of the demodulator is η _Dem =∞. Adequacy decision criteria based on path metric differences are used.

Fig. 4 shows the traditional decoding scheme, the optimal maximum likelihood decoding scheme and the decoding scheme proposed in the present invention at SIR=0dB, the bit error rate (BER) performance under SNR=20dB and different number of interferences . Figure 5 shows the BER performance of the three decoding schemes at SNR=20dB, 5 interfered sub-carriers and different SIRs. From Fig. 4 and Fig. 5, it can be found that the scheme of the present invention is obviously superior to the traditional decoding scheme, and can achieve the actual best decoding performance for different numbers of disturbed subcarriers and a wide range of interference power.

In Fig. 4 and Fig. 5, the optimal threshold η _Dec is used for the adequacy decision criterion based on the path metric difference, and Fig. 6 and Fig. 7 use the function η _Dec /σ ² to estimate the BER performance, where σ ² is Background Gaussian noise variance. It can be seen that the best error rate can be obtained by setting the threshold in an appropriate range. Moreover, the optimum threshold value does not change significantly with the number of interfered subcarriers or different interference power. In fact, the best threshold depends only on the background Gaussian noise power, which can be determined off-line. Therefore, the adequacy decision criterion based on path metric difference is robust and feasible to different interference environments in practice.

In Figure 8 and Figure 9, each codeword is mapped to 10 OFDM blocks with 200 subcarriers. The threshold is η _Dem =η _Dec =26σ ² . The channel fading for each subcarrier is randomly generated. The decoding scheme is evaluated both with and without channel estimation errors. The channel estimator used is a frequency domain LS estimator with two pilots on each subcarrier. The interference for each subcarrier is randomly generated with a probability of 0.04, and the power is evenly distributed between [-20, 10] dB. The adequacy decision criterion based on path metric difference and CRC is used together. In particular, if either of the two criteria is satisfied, the decoding process will terminate. The generator polynomial of the CRC is 435 (octal). Zeros are padded at the end of the CRC codeword so that the convolutional code is terminated. Since the information sequence is very long, the spectral loss due to CRC and zero padding can be ignored.

8 and 9 show the BER and word error rate (WER) of the conventional decoding scheme, the optimal maximum likelihood decoding scheme and the decoding scheme of the present invention. In order to demonstrate the performance gain obtained with the code structure, two possible implementations of the invention are evaluated. The designation "Proposed 1" represents a scheme that uses only the demodulator for symbol erasure based on the channel output. The designation "Proposed 2" represents a scheme that utilizes both demodulator and decoder for erasure. In the scheme "Proposed 1", the decoder does not perform erasure like a traditional Viterbi decoder. In this scheme, the erasure of unused code structures can be viewed as a separate erasure and decoding scheme. The gap between "Proposed 1" and "Proposed 2" shows the performance gain of joint erasure decoding over separate erasure decoding. The solid line represents the performance with the exact channel fading factor. It can be seen that the traditional decoding scheme is almost ineffective. Our proposed decoding scheme significantly outperforms conventional schemes, especially at high SNR. In addition, "Proposed 2" utilizing joint erasure decoding is significantly better than "Proposed 1" and achieves almost the best decoding performance across the entire SNR range examined. These results show that the demodulator can use only the channel output to erase some of the symbols blocked by the interference. With the code structure and using joint erasure decoding, almost all symbols blocked by interference can be removed.

In these figures, the evaluations of our proposed decoding scheme and conventional decoding schemes with channel estimation error are shown as dashed lines. It can be found that the performance loss of our proposed scheme "Proposed 2" is within 3dB for BER and within 2dB for WER. The scheme "Proposed 1" exhibits larger performance loss due to channel estimation errors. For WER = 0.1, the performance loss exceeds 4dB. This shows that the joint erasure decoding scheme proposed by the present invention is robust not only to interference with unknown distribution but also to channel fading estimation errors. The computational complexity of the scheme "Proposed 1" is basically similar to that of the traditional decoding scheme. The computational complexity of the scheme "Proposed 2" is within several times of the traditional decoding scheme for the range of bit error rates of interest, such as WER less than 0.1. For example, when WER=0.01, the proposed decoding scheme is about 1.5 times better than the traditional scheme under ideal channel estimation. Due to the channel estimation error, the computational complexity is about three times that of the traditional scheme. Also, complexity falls as SNR rises.

In conclusion, without interference information, our proposed decoding scheme significantly outperforms traditional decoding schemes and achieves the optimal maximum likelihood decoding performance that can only be obtained with full interference statistics. For a system with automatic repeat request (ARQ), the proposed decoding scheme can greatly improve the throughput of the system. The price is increased computational complexity. For low channel estimation error and high SNR, the computational complexity is small compared to conventional decoding schemes. Moreover, the proposed decoding scheme can be used to cooperate with other jammer detectors. With the help of an interference detector, the complexity of the proposed decoding scheme can be further reduced.

Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these modifications and variations.

Claims (15)

1. A method for removing interference, characterized in that, comprising the following steps: A. After receiving the sequence to be decoded, the receiving end calculates the metric of each bit of each symbol in the sequence, and adds an erasure mark to the corresponding symbol accordingly; B. Erase the symbol carrying the erasure mark, calculate the metric of each bit in the erased symbol, and decode the metric; C. Judging whether the decoding needs to be terminated according to the decoding result, if so, proceed to step D; otherwise, after further adding an erasure mark to the corresponding symbol according to the decoding result, proceed to step B; D. Output the decoding result. 2. The method according to claim 1, wherein in the step A, if the measure of one or more bits on a symbol is greater than a preset threshold value, an erasure mark is added to the symbol. 3. The method according to claim 1, wherein, before decoding in the step B, it is first judged whether the number of decoding iterations exceeds the preset number of iterations, if not, decoding is performed; otherwise, directly transfer to Go to step D. 4. The method according to claim 3, wherein, if the preset number of decoding iterations is not exceeded, deinterleaving is performed on each bit of the symbol carrying the erasure flag before decoding. 5. The method according to claim 1, wherein the calculation of the metric of each bit in the erased symbol adopts

in

Represents the measure of each bit of each symbol in the symbol sequence, E _k represents the erasure state of the symbol carrying the erasure mark, k represents the label of the symbol, m represents the number of bits corresponding to a symbol, and j represents the number of bits in a symbol bit sequence number. 6. The method according to claim 3, characterized in that, the decoding described in step B comprises the following steps: -Construct the bit lattice diagram of the state transition during convolutional coding, which is used to generate the state transition and output bit sequence of the coded symbol sequence during the coding process; -construct bit erasure indicating trellis diagram, which is used to represent the bit erasure distribution of a state transition in decoding; - multiplying the bit trellis with the erasure indication trellis to obtain a product trellis, representing the state transition of a convolutional code with a certain number of bit erasures, and the distribution of erased bits; -The decoder searches for the algorithmic shortest path of the bit carrying the erasure mark in the product trellis diagram under the given erasure number and the corresponding bit erasure position condition, and determines the corresponding maximum likelihood accordingly Codeword. the 7. method as claimed in claim 6, is characterized in that, described step C comprises the following steps: - Obtain the path metric difference value of every two maximum likelihood codewords currently determined; -Comparing the absolute value of the obtained path metric difference with the preset threshold value, if there is an absolute value of difference smaller than the threshold value, then stop decoding and go to step D; otherwise, according to the decoding result After further adding erasure marks to the corresponding symbols, go to step B. the 8. The method according to claim 7, wherein, according to the decoding result described in the step C, further adding an erasure mark for the corresponding symbol, comprising the following steps: -Find the path with the most bits that should be erased in each shortest path; - erase the bits that should be erased in this path; - Among the symbols traversed by the path, if a symbol has one or more bits that should be erased, add an erasure flag to the symbol. the 9. The method according to claim 8, characterized in that, using the correlation of the blocked symbols, further adding erasure marks to the corresponding symbols. the 10. The method according to claim 7, characterized in that, in the step D, the output decoding result that does not exceed the preset number of decoding iterations satisfies that the absolute value of the path metric difference is less than the threshold value Among the two maximum likelihood codewords of , the maximum likelihood codeword with more bits to be erased; The output decoding result exceeding the preset number of decoding iterations is the maximum likelihood codeword with the most bits to be erased among the determined maximum likelihood codewords in any round of decoding. the 11. A system for removing interference, comprising: The demodulator is used to calculate the bit metric according to the received symbol sequence, and add an erasure mark to the corresponding symbol according to the bit metric, and output the bit metric of the symbol carrying the erasure mark; Decoder, used to decode the received bit metric, and output the decoding result; The second judging unit is used to receive the decoding result of the decoder, and judge accordingly whether the decoding needs to be terminated, and if it needs to continue decoding, then instruct the decoder to feed back the decoding result to the demodulator; otherwise, instruct The decoder outputs the decoding result to the outside. the 12. The system of claim 11, further comprising: The decoding iteration judging unit is used for judging whether the preset number of decoding iterations has been reached, and if it has been reached, the decoding of the decoder is terminated, and the decoder is instructed to output the decoding result; otherwise, the decoding is continued. the 13. The system according to claim 11 or 12, wherein the system further comprises: The deinterleaver is used for deinterleaving the bit metrics output by the demodulator, and then sending them to the decoder for decoding. the 14. The system of claim 13, further comprising: The interleaver is configured to interleave the decoding result fed back from the decoder to the demodulator. the 15. The system according to claim 13, wherein the symbol sequence received by the demodulator comes from the channel output after front-end processing and the decoding result fed back by the decoder; Or, the decoding result of channel estimation output and decoder feedback; Or, the output of the interference detector and the decoding result fed back by the decoder. the

Images