Research And Application Of Unequal Error Protection In Storage And Transmission

Information has become the most important resource with the popularization of Internet and the progress of VLSI technology.Therefore,it is of great importance to ensure the reliability of information.Information error in the medium can be modeled as channel error in the communication system,and the error recovery is usually performed by error control.Forward Error Correction(FEC)does not require feedback and only needs to be encoded once in the input channel,so it becomes the most commonly used error correction method.However,FEC itself can not adapt to the channel situation and automatically adjust the protection ability,so we put forward the UEP(Unequal Error Protecion)method to solve this problem.In the storage process,the most widely used electronic storage NAND flash memory is selected as research background.The increase of the storage density of the flash memory leads to the increase of the error code,and the error correction capability of the traditional ECC is insufficient.Moreover,the ECC itself may fail and the lost error caused by the flash bad track can not be recovered.Therefore,Raptor code word is adapted for protection.The existing method for flash memory protection does not take into account the actual situation of different block error situation,so we put forward Raptor's adaptive code rate mechanism further,we get the approximate formula of the adaptive code rate according to the situation,and encodes each memory block with the obtained code rate to achieve the effect of UEP.It can be seen through simulation that no matter how the error condition of a block changes,the mechanism can adaptively adjust the error correction capability so that the bit error rate after decoding stabilizes in a lower range and achieves better protection effect.We choose the increasing video data in network transmission as the object of protection.Different frames in video have dependencies and have different impact on video quality.Channels are not perceptible to data.Therefore,UEP is required when encoding to provide targeted protection.1)We found that expanding window mechanism needs to modify the original codeword,resulting in the undermining of original codeword performance and poor compatibility.Therefore,we explore another Layer-aware mechanism.We propose a simplified decoding algorithm that implements the LA mechanism based on the RaptorQ codeword.Through simulation,it is found that the error performance of all importance data of LA mechanism is improved by 1 to-3 times,which is more balanced than expanding window mechanism,and the overall bit error rate is lower,protection effect is better.2)On this basis,we also propose a rate-adaptive FEC mechanism based on broadcast and4 / 5G heterogeneous networks,establish an optimization model for user experience,and propose a fast search algorithm to attempt to find the optimal code rate adaptively under limited circumstances,and finally verifying the feasibility of the algorithm.Under different conditions,the algorithm can accurately search for the solution corresponding to the best coding rate.In this paper,we propose a corresponding rate-adaptive algorithm in the common scenarios of storing and transmitting,it solves the problem that the FEC itself can not adapt to the channel feature and adjust protection capability,and the data is better protected.