Noise gradient descent multi-bit flipping decoding early stop method for LDPC codes

Abstract

The invention provides a noise gradient descent multi-bit flipping decoding early stop method for LDPC codes, which comprises the following steps: (1) firstly, carrying out hard decision on a receivedsignal to obtain a vector, and initializing parameters; (2) calculating a syndrome s of the vector x, if each element in the s is 1, declaring that decoding succeeds, returning a result x and endingiteration, and otherwise, adding 1 to the number t of iterations; (3) calculating a flip equation E of x; (4) if the current iteration frequency t belongs to the early stop set S, starting a counter;(5) finding the positions of all bits to be flipped according to E and theta, flipping the bits, and updating x. Recording the number of flipped bits; (6) if the number t of iterations is smaller thanthe maximum number T of iterations and the value l of the counter is smaller than an early stop threshold lambda, turning to the step (2); otherwise, returning to x and ending iteration. According tothe method, the average iteration frequency of the decoder at a low signal-to-noise ratio can be remarkably reduced.

Description

technical field [0001] The invention relates to the technical field of channel coding in wireless communication, in particular to an LDPC code noise gradient descent multi-bit flip decoding early stopping method. Background technique [0002] Compared with the Sum-Product Algorithm (SPA), the traditional Bit Flipping (BF) decoding method of LDPC (Low-Density Parity-Check) code has very low complexity, and the corresponding bit error rate (BitError Rate, BER) performance is much worse than the SPA decoding algorithm. In order to improve the BER performance of the BF decoding algorithm, a BF decoding algorithm that adds soft information in the decoding process is proposed. As the BER performance of this type of BF algorithm is continuously improved, the complexity of the algorithm is also increasing. Among them, the Multi-Noisy Gradient Descent BitFlipping (M-NGDBF) decoding algorithm is a kind of decoding algorithm that uses soft information. Typical BF algorithm. For the ...

AI Technical Summary

Problems solved by technology

If the decoding fails, it will inevitably

Images