| Deep space communication has characteristics of large transmission distance and large link loss, resulting in energy loss and inefficient feedback. To guarantee the equity and efficiency of transmit signal should take some technology or ways. The high-gain channel coding technology is one of ways. Deep space communication research has always been focus on excellent performance channel coding. The fixed code is widely used in deep space communication channel coding so far. Because of influence of environment, the deep space channel is dynamic time varying. The fixed code could get better capacity, but it can’t fit change of channel.However,rateless code can change the code rate according to channel change by itself, and code achieve the maximum utilization of channel capacity. Spinal code is proposed in 2012,a newly rateless code which essentially achieves capacity over both BSC and AWGN channels. In deep space communication, Spinal code can get super performance in a wide range of SNRs compared to the fixed code LDPC and rateless code Raptor code and Strider code.So,Spinal code is perfect for deep space communication.First, I introduce the theory of Spinal coding, and analysis the capacity of Spinal code through comparing the capacity of LDPC code and Raptor code.Then,to improvement the capacity or induce the complexity, I study Spinal code from two aspects: one is Spinal code’s encoding and decoding mechanism; the other is punctured coding.To reduce the complexity of decoding, Spinal code cut decoding tree when built the tree. Then sending more symbols of last of encoding section could make up the performance lost of decoding tree. The research shown that send more symbols of the last one of encoding section could improve about 9% capacity performance and reduce about 20% complexity when SNR?5 dB. And the number of symbols should be less of five, the number of encoding section should be less of four.The value of k plays an important role in code performance and complexity. From some research, performance and complexity of Spinal code have relationship of the value of message section. Then, I take different values on encoding from large to small.The closer different values, the better performance can get. And the performance closed to performance of minimum value of different values. The complexity is lower than complexity of maximum value of different values. Therefore, taking different values on message section could reduce complexity with a less lost of performance,and overcome the limit of the value of k.The characteristic of Spinal code mechanism limit code rate. So, punctured codingcould break the limit and refine rate step. From the study, refining symbol pass could improve code performance, but the complexity also is increased. Then Take the puncture scheme of four subchannel could get better capacity performance and low complexity. |
