LAAP: A Learning Automata-based Adaptive Polling Scheme for Clustered Wireless Ad-Hoc Networks

In multi-hop ad hoc networks, besides collision-free transmissions, channel utilization should be also enhanced due to the scarce bandwidth. In this paper, we propose a learning automat-based adaptive polling scheme for medium access scheduling in clustered wireless ad-hoc networks to enhance the channel utilization. In this scheme, each cluster-head takes the responsibility of coordinating intra-cluster transmissions so that no collisions occur. Taking advantage of learning automaton, each cluster-head learns the traffic parameters of its own cluster members. Cluster members are prioritized based on these traffic parameters. Each cluster-head then takes the traffic parameters into consideration for finding an optimal channel access scheduling within its cluster. By the proposed polling scheme, each cluster member is assigned a portion of bandwidth proportional to its need (i.e., traffic load). The results show that the proposed channel assignment policy considerably improves the channel utilization. Simulation experiments also show the superiority of the proposed polling-based medium access scheme over the existing methods in terms of channel utilization, waiting time for packet transmission, and control overhead.