Research On Digital Fountain Coding For File Delivery In Deep Space Communications

Deep space exploration has great scientific and economic significance. Thenormal operation of the communication system is one important guarantee ofsuccess in the deep space exploration mission.However, the transmission distanceof the communication is so long that the receive end has a high loss rate and hugepropagation delay, on the other side, the high frequency band exploited by deepspace communication is highly vulnerable to fluctuating weather condition, whichcause the further aggravation of the signal energy, and high BERs will degrade thereliable delivery on the scientific data, due to frequent retransmissions on themissing data packets. Above all, the specificity that deep space communicationlinks own challenges the reliability of channel modeling, the file delivery protocoland modulation and coding technology.In this paper, a channel model on transport layer over Ka-band links isproposed, which apparently reflects the characteristics of rain attenuation andequivalent noise temperature from the earth’s atmosphere. The proposed model isbuilt on a four-state Markov chain, which considered the different features ofvarious packet types during CCSDS File Delivery Protocol (CFDP) transactions.We provide the exact criterion on how to divide different weather conditions ateach state, and give the analytical results on transition probability matrix. Inparticular, we take Guangzhou in China as an example to find the probabilitydistribution function of rain attenuation, and thus the transition matrix betweendifferent states, based on ITU-R rain attenuation model. The simulation resultsshow that, compared with two state channel model, the proposed model couldefficiently describe the error characteristics of different packet types during the filedelivery, which leads to obvious degradations on CFDP performance.Under low signal to noise ratio (SNR) condition, the Feher Quadrature PhaseShift Keying (FQPSK) demodulation is significantly susceptible to frequency orphase offsets. This paper proposes a serially concatenated coded modulationsystem of FQPSK and low-density parity-check (LDPC) codes, which couldefficiently resist residual frequency offset by an intended compensation algorithm.The designed compensation function based on maximum-likelihood estimation(MLE) is embedded into the iterative demodulation-decoding process. The jointdemodulation and decoding scheme employs soft-input-soft-output (SISO)-basedmaximum a posteriori probability (MAP) algorithm to recover the originalcodeword. In particular, the transmitted codeword sequence is rearranged at thesender in a different order from original codeword, in order to degrade the influence of frequency offsets on the decoding process. From the simulation we canfind that the Bit Error Rate(BER) of the proposed scheme can be improvedefficiently and especially be10-3to10-4under the frequency offsets from100ppmto700ppm. The system provided thus solved the problem that residual frequencyoffset has a great influence in deep space communication.