Research On The Encoding And Decoding Algorithms Of Systematic Polar Codes

Channel coding technology develops rapidly since the proposal of noisy-coding theorem,modern coding technologies,such as Turbo codes,LDPC codes,have the error performance which is close to the Shannon limit,but none of them achieves Shannon limit until the proposal of polar codes.Polar codes was proposed by Arikan,and its asymptotic performance was proven to be Shannon limit under binary discrete memoryless channel.Polar codes can be divided into non-systematic type and systematic type according to the construction methods,and the systematic type obtains better error performance while there is no explicit systematic decoding algorithm.Lot of researches are concentrated on non-systematic decoding algorithms,so non-systematic decoding algorithms and recoding becomes the universal concatenated decoding algorithm for systematic polar codes,which results in the great decoding latency of systematic polar codes.This thesis propose improved algorithms in both encoding and decoding to solve the decoding problems as well as enhance the practicability of systematic polar code.On this basis,a systematic decoding scheme is proposed with low latency and resources for the devices with limited resource.The main work is organized as follows:Firstly,we propose a modified systematic coding scheme according to the check properties of non-systematic decoding algorithm.Researches show that the proposed scheme can simplify the recoding procedure while keeps the same encoding complexity as the original systematic encoding scheme.As a result,this scheme reduces the decoding latency greatly and enhances the generality of systematic polar codes.Secondly,we propose a reverse sequence check aided successive cancellation list decoding scheme to break the lack of systematic decoding.Researches show that the proposed scheme eliminates the recoding procedure with the check of reverse sequence,which simplifies the decoding procedure and reduces decoding latency.Besides,the proposed scheme performs slightly better than the adaptive SCL.Finally,we propose an array-check successive cancellation list decoding scheme to meet the need of low latency and resources of micro-devices.Researches show that the proposed scheme achieves multiple check with the help of array check procedure,and solves the hysteretic check as well as reduces the decoding latency.Besides,a cut strategy excluding the best path is adopted in this scheme to reduce the resources occupation,and a direct mapping operation is adopted to simplify the data preprocessing before systematic encoding as well as reduce the latency of information collection.Simulations show that the proposed scheme reduces the latency of information collection and decoding,resources occupation greatly with little degradation in error performance.