论文部分内容阅读
为在保障自由空间光(FSO)通信质量的同时,缩短传统的自动重传请求技术所需等待时间、减小缓存资源消耗,并随信道的变化适配数字喷泉码冗余符号的数量,提出一种基于Q学习算法的Raptor10码译码策略。该策略采用以减少冗余符号和降低误码率为联合回报目标的Q学习算法,可对接收端所接收的Raptor10码冗余符号进行动态调整,并且随着译码经验的不断累积可自动提升通信性能。仿真结果表明,在弱、中、强3种湍流条件下,与传统的采用固定冗余度Raptor10码译码策略和线性滤波调整算法相比,所提方法可有效减少3%的平均冗余符号数量。
In order to reduce the waiting time required by the traditional automatic repeat request technology and reduce the resource consumption of the buffer while ensuring the quality of free space optical (FSO) communication, the number of redundancies in the digital fountain code is adapted to the change of the channel A Raptor10 Decoding Strategy Based on Q Learning Algorithm. The strategy adopts the Q learning algorithm, which aims to reduce the redundant symbols and reduce the bit error rate. The algorithm can dynamically adjust the Raptor10 code redundant symbols received by the receiving end, and can automatically increase as the decoding experience accumulates Communication performance. The simulation results show that under the conditions of weak, moderate and strong turbulence, compared with the traditional Raptor10 code decoding strategy and linear filter adjustment algorithm, the proposed method can effectively reduce the average redundant symbols of 3% Quantity.