Collision-Aware Pattern Design for Uplink Complementary Coded Code-Hopping Multiple Access Systems

Code-hopping multiple access (CHMA) technology is an effective approach for improving the user capacity of direct sequence spread spectrum systems. However, the use of randomly generated code-hopping (CH) patterns in a typical CHMA system often causes code collisions that severely degrade the bit error rate (BER) performance. Our previous work has developed a CH pattern design method to improve the BER performance of complementary coded CHMA (CC-CHMA) by pre-generating well-designed CH patterns. However, this method is mainly applicable to single-path channels, while detecting collisions and taking appropriate actions in complex multipath channels remains challenging. In this work, we present a collision-aware algorithm and propose a CH pattern update algorithm based on multi-agent reinforcement learning for CC-CHMA in uplink multipath channels. The simulation results show that the proposed CH pattern design scheme can automatically adapt to the uplink multipath channels and effectively reduce code collisions, thereby achieving further improved BER performance compared with existing CC-CHMA systems.