Optimal secondary access in retransmission based primary networks via chain decoding

This paper investigates the design of secondary access policies which exploit the temporal redundancy of the retransmission protocol employed by primary users (PU) to improve the spectral efficiency of wireless networks. Secondary users (SU) perform selective retransmissions in order to optimize the potential of interference cancellation at the receiver. The corrupted signals are selectively buffered at the SU receiver, and then decoded via the successive application of chain decoding [1]. The optimal SU access policy which maximizes the SU throughput under a constraint on the maximum interference caused to the PU is derived, and its performance is found in closed form. It is shown that such policy optimally randomizes among three modes of operation of the SU: 1) The SU remains idle over the entire retransmission interval of the PU, to avoid interfering with the PU; 2) The SU transmits only after decoding the PU packet to leverage interference cancellation; 3) The SU always transmits over the entire retransmission interval of the PU, so as to leverage chain decoding. The optimal randomization is determined by the constraint on the maximum interference caused to the PU. It is shown numerically that chain decoding attains a throughput gain of 15% with respect to a state-of-the art scheme where the SU does not perform selective retransmissions.