Separation of Interfering Time-Frequency Signals in CNC Satellite Communication Using the Extended Viterbi algorithm

In recent years, the new technology of satellite communication based on CNC has received much attention due to the optimal bandwidth usage. In this communication, the Cooperative receivers can save up to half of the bandwidth despite co-band interference and without reducing efficiency. However, the analysis of this type of communication in non-cooperative receivers is associated with many challenges due to the nature of time-frequency interference. Additionally, in this problem, no information about signal parameters is not available. So far, methods have been proposed to separate these interfering signals. However, all of the proposed methods are performed with knowledge of signal parameters such as phase and frequency offset, delay, and signal amplitude attenuation. In this article, after the mathematical modeling of the received signal in the non-cooperative receiver, the Signal parameters estimation is conducted using higher-order statistics of the received signal. Next, the separation process is performed by a proposed structure called the "Extended Viterbi algorithm". In this structure, the Viterbi algorithm has been used with the aim of solving the ambiguity of estimating phase offsets as well as the final estimation of two sequences of interfereing time-frequency symbols. The comparison and evaluations for interfering signals with two modulations BPSK and QPSK show that, parameters estimation has a significant impact on the error rate of the detected symbol, such that at signal-to-noise ratios higher than zero, the symbol error rate (SER) decreases more steeply in the proposed method than in the Per-Survivor Processing (PSP) algorithm. The simulation results show that the symbol error rate in the proposed method at a signal-to-noise ratio of 30 dB for BPSK modulation is reduced by 30% and for QPSK modulation by 60%. However, in the proposed method, the complexity of the separation method increases by 2^2L with increasing the modulation order.