系统工程与电子技术
繫統工程與電子技術
계통공정여전자기술
SYSTEMS ENGINEERING AND ELECTRONICS
2014年
11期
2288-2294
,共7页
低密度奇偶校验码%加权比特翻转%附加偏移项%有限域几何
低密度奇偶校驗碼%加權比特翻轉%附加偏移項%有限域幾何
저밀도기우교험마%가권비특번전%부가편이항%유한역궤하
low density parity-check (LDPC)code%weighted bit flipping (WBF)%additive offset term%fi-nite geometry (FG)
大量仿真表明，基于幅度和的改进型加权比特翻转（modified sum of the magnitude based weighted bit flipping，MSMWBF）译码算法对于行重／列重较小的低密度奇偶校验（low density parity-check，LDPC）码而言，展现出巨大的性能优势，但对于行重／列重较大的基于有限域几何（finite-geometry，FG）的 LDPC 码，性能损失严重。首先对此现象进行理论分析。其次，引入附加的偏移项对 MSMWBF 算法的校验方程可靠度信息进行修正，提高了算法对行重／列重较大的 LDPC 码的译码性能。仿真结果表明，在加性高斯白噪声信道下，误比特率为10－5时，相比于 MSMWBF 算法，在适度增加实现复杂度的条件下，所提算法可获得约0．63 dB 的增益。
大量倣真錶明，基于幅度和的改進型加權比特翻轉（modified sum of the magnitude based weighted bit flipping，MSMWBF）譯碼算法對于行重／列重較小的低密度奇偶校驗（low density parity-check，LDPC）碼而言，展現齣巨大的性能優勢，但對于行重／列重較大的基于有限域幾何（finite-geometry，FG）的 LDPC 碼，性能損失嚴重。首先對此現象進行理論分析。其次，引入附加的偏移項對 MSMWBF 算法的校驗方程可靠度信息進行脩正，提高瞭算法對行重／列重較大的 LDPC 碼的譯碼性能。倣真結果錶明，在加性高斯白譟聲信道下，誤比特率為10－5時，相比于 MSMWBF 算法，在適度增加實現複雜度的條件下，所提算法可穫得約0．63 dB 的增益。
대량방진표명，기우폭도화적개진형가권비특번전（modified sum of the magnitude based weighted bit flipping，MSMWBF）역마산법대우행중／렬중교소적저밀도기우교험（low density parity-check，LDPC）마이언，전현출거대적성능우세，단대우행중／렬중교대적기우유한역궤하（finite-geometry，FG）적 LDPC 마，성능손실엄중。수선대차현상진행이론분석。기차，인입부가적편이항대 MSMWBF 산법적교험방정가고도신식진행수정，제고료산법대행중／렬중교대적 LDPC 마적역마성능。방진결과표명，재가성고사백조성신도하，오비특솔위10－5시，상비우 MSMWBF 산법，재괄도증가실현복잡도적조건하，소제산법가획득약0．63 dB 적증익。
Extensive simulations show that for some low density parity-check (LDPC)codes of low row/colum weight,the modified sum of the magnitude based weighted bit flipping (MSMWBF)decoding algorithm performs extraordinarily well.However,this dose not hold when it comes to large row/colum weight finite-ge-ometry LDPC codes,which is first theoretically analyzed.Secondly,improvement in performance is observed for large row/colum weight LDPC codes by considering an additive offset term for the parity-check-equation reliabi-lity of the above algorithm.Simulation results show that the performance of the improved scheme is better than that of the MSMWBF algorithm about 0.63 dB at bit-error rate of 10 -5 over the additive white Gaussian noise channel with only a modest increase in computational complexity.
