Channel Decoding from Multiple Looks Under Receiver Impairments

Receiver impairments play a significant role in determining system performance, particularly in adversarial environments subject to intentional jamming that seek to impair operation by collapsing the dynamic range of the receiver. Clock jitter is one of the key impairments that limits analog-to-digital (A/D) converters, and it causes significant challenges for channel decoding as it distorts the received signal’s timing, making compensation difficult. Motivated by this challenge, we consider performance and design for the two-look channel, which utilizes two receivers with independent clock jitter. Using a statistical characterization of the jitter, we derive the degree to which mutual information is increased in the two-look scenario versus the case when a single receiver is employed. This analysis motivates the consideration of optimal decoder design, but this is complicated by the nonlinear timing jitter model. Hence, we turn to a machine learning (ML)-based decoding framework. Numerical results demonstrate the improvement in mutual information from having two looks, and simulation results demonstrate the improved channel decoding performance.