Research On Polar Code Error Correction Technology For Flash Memory System

The theory of channel polarization shows that for any Binary-input Discrete Memoryless Channel,when the code rate is less than the channel capacity,a codeword that reaches the Shannon capacity limit can be constructed.Inspired by the phenomenon of channel polarization,Arikan gives a scheme for constructing polar codes,and proves that it has a lower complexity of coding and decoding.The good performance of polar code has attracted the attention of a large number of scholars,and has opened up a new direction for channel coding technology.The NAND multi-level cell?MLC?flash memory is a kind of non-volatile storage device,which has a higher storage density,smaller size,and smaller power consumption.And the MLC flash memory has become a mainstream storage device on the market.Compared with the Single-Level Cell?SLC?flash memory,the Multi-Level Cell?MLC?flash memory cell stores two bits of data,reduces the unit cost of the flash memory,and is more suitable for a large-capacity storage device.However,as the density of flash memory increases,the flash cells are more prone to errors.Therefore more effective error correction schemes are needed to improve the reliability of the flash memory system and extend the life of the flash memory chips.This article focuses on the application of polar codes in NAND MLC flash memories.The work content is divided into the following aspects:1.Determine the flash channel model.In this paper,two channel models are mainly men-tioned,Gaussian distributed channel model and 2-BBM channel model.By comparison,although the channel model based on Gaussian distribution has low complexity but its ac-curacy is not high,the 2BBM channel model based on actual bit error characteristics is not only highly accurate but also not very complicated.In contrast,the 2-BBM channel model is selected as the channel model of the flash memory.In addition,the characteristics and parameters of the 2-BBM channel model are given in detail.2.The flash memory has a high requirement for error correction code,which requires both a long code and a high code rate.Therefore,it is necessary to combine the characteristics of the flash memory when designing a flash memory error correction scheme.According to the asymmetry of the flash memory system error,the construction of the polarization code based on the Monte Carlo construction method is established,and a good performance SCL?Successive Cancellation List?decoding algorithm is used in the decoding algorithm.This paper studies the error frame rate characteristics of the SCL?Successive Cancelation List?decoding algorithm with different code lengths,different code rates,and different decoding widths.The simulation results show that for high bit rate flash memory systems,SCL??=1?decoding algorithm can be used to compensate for the performance loss due to high bit rate.3.Compare the performance of LDPC codes with power codes with a code length of 2 and LDPC code codes with the same code length on flash memory.The code length for flash memory is not necessarily a power of 2,and a scheme for shortening the polarization code is introduced,and the shortened polarization code is compared with an LDPC code having the same code length and the same code length.The performance of shortened polar code applications on flash memory systems.According to the simulation results,in the range of P/E cycle of 5000-10000,when the frame error rate reach 10^-1,compared with LDPC codes,polar codes can be used as error correction codes to perform nearly 500 P/E cycles,and the error correction performance of polarization codes is significantly better than that of LDPC codes.Therefore,we believe that the polar code can be well applied in a flash memory system,and as the code length becomes larger,the performance of the polar code will be better,which can effectively improve the performance of the flash memory and prolong the service life of the flash chip.