Performance Improvement of Random Access by Prioritizing Collisions

The importance of unlicensed spectrum is highlighted in terms the flexibility network deployment for various services envisioned 5G and beyond. Since listen before talk mandatory channel access it causes an unavoidable waste resources due to collisions, efficient random backoff mechanism required. In existing schemes that impose waiting penalties on collided packets, a degraded fairness performance observed. this work, we analyze how prioritizing packets can improve compared schemes. To end, devise scheme called Collision Priority Backoff (CPB) under concept giving priority packets. We apply Bayesian optimization carefully determine parameters CPB maximize throughput. optimized require number stations network, also adaptive version Adaptive (ACPB). deal with environment where changes as switching bandit problem, employ variant upper confidence bound policy ACPB. Various simulation results validate proposed shows high throughput performance.