A cross-layer goodput enhancement considering CRC coding and ARQ dynamics

Time varying channel conditions with high bit error rate causes frequent packet corruptions in wireless communication networks. Cyclic Redundancy Checks (CRCs) and Automatic Repeat reQuest (ARQ) mechanisms are heavily used for effective error detection and recovery, respectively. In previous studies, the performance of CRC generator polynomials is evaluated solely based on Undetected Error (UE) probability without taking retransmission and overhead dynamics into account. This paper presents a new performance evaluation criterion (i.e., link layer goodput) which takes both CRC coding and ARQ dynamics into account. The proposed criterion is used as an objective function to compute the best polynomial and message length pairs corresponding to varying network conditions. We, then, propose an adaptive CRC coding mechanism which dynamically updates the CRC polynomial and message length pair at run-time in accordance with the time varying conditions. Simulation experiments show that dynamic selection of CRC polynomial and message length pair yields significant performance gains.