Principle and Limit for Guaranteeing the Mean Access Delay in S-ALOHA

Abstract

The access delay of Slotted ALOHA (S-ALOHA) is intricately coupled with backoff intervals and the number of backlogged users, while new arrivals continuously join the system according to the input rate. Consequently, guaranteeing a mean access delay constraint is a challenging problem since the system must control these interdependent variables. This work proposes an optimization framework that determines the maximum allowable input rate to satisfy the access delay constraint, where a throughput-maximizing backoff algorithm based on the extended Kalman filter (EKF) is developed. To solve the optimization problem, extensive analysis of S-ALOHA is conducted with access delay characterization. The results demonstrate that while the input rate is constrained through the proposed framework, the proposed EKF-based backoff algorithm successfully guarantees the access delay without the knowledge of the exact backlog size.