부산대학교 도서관

Ultra-reliable low-latency communication (URLLC) and machine-to-machine (M2M) relaying are increasingly important. In-band full-duplex (IBFD) communication is desirable for low-latency communication because it can theoretically double the capacity of half-duplex (HD) communication, but is limited by self-interference (SI) in which the IBFD transmitter injects interference into its received signal due to large transmit and receive power differences. There are many methods to mitigate SI in IBFD, but we focus on multiple-input multiple-output (MIMO) precoding to spatially null SI, while also being careful to limit communication latency. Unlike works aiming to compute the highest achievable rate in IBFD relays with an essentially fixed precoder, we propose a strategy allowing the precoder to change on a per channel use basis to avoid rate bottlenecks at each hop. Our strategy generalizes to multiple relays. We frame finding the sequence of precoders to relay packets through $N$ IBFD MIMO relays with the lowest communication latency as a shortest path problem. Specifically, we design a quantized covariance precoder codebook at each transmitter based on limiting the maximal SI power each precoder produces. Then, an iterative algorithm is used to optimize the selection of precoders over channel uses to relay packets in the fewest channel uses possible.