Research On Adaptive Algorithm Based On Rateless Codes For Satellite Communication

With the feature of time-varying channel state, the satellite communication requires adaptive transmission technology. But the feature of long round trip time makes it very difficult to design the adaptive technology. In order to maximize the throughput of the systems within the constraints of minimizing the transmission delay and the decoding complexity, this paper proposes an adaptive transmission strategy, which is based on rateless codes and IR-HARQ(Increase Redundancy-Hybrid Automatic Repeat Request) technology, and designs the corresponding transmission frame format and two adaptive algorithms.Firstly, this paper introduces the two features of the satellite communication systems, and constructs the corresponding slow-fading channel model and fast-fading channel model. Secondly, this paper elaborates the fundamental of rateless codes and HARQ technology, and chooses an adaptive transmission strategy based on RCM(Rate Compatible Modulation) codes and IR-HARQ technology. We also deduce the mathematical expressions of system performance. Thirdly, this paper designs a new transmission frame format and discusses the control information in the frame in detail. The new frame can effectively improve the transmission efficiency. Finally, this paper puts forward an adaptive algorithm based on optimal solution and SNR estimation, and an adaptive algorithm based on decoding threshold table and probe mechanism. Both the algorithms can adjust the size of the new packages and incremental packets on time to match the changes of channel state so that the system can get the best performance.The two algorithms proposed in this paper can obtain good performance in four channel models. The average throughput reaches about 97.7% of the theoretical value. The average number of transmission reaches less than 1.2 times of the theoretical value and the average cumulative decoding complexity reaches less than 1.3 times of the theoretical value. The channel utilization can reach about 100% when the decoding capability is 1.3 times of symbol transmission rate. By comparison, the algorithm based on optimal solution and SNR estimation has better performance, but the other algorithm is simpler and not affected by SNR estimation.