Semi-myopic algorithm for resource allocation in wireless body area networks

The key features of wireless body area networks are limited energy resources of the sensor nodes and the need for highly reliable packet transmission. Therefore, designing a suitable algorithm that schedules transmission of the nodes is very important. Although an optimal algorithm has been previously reported based on a partially observable Markov decision process (POMDP), its complexity is high. In this study, the authors propose a suboptimal algorithm for scheduling the transmissions, i.e. a semi-myopic algorithm, with a performance close to the optimal, albeit with much lower complexity. To this end, they modify the structure of the cost function of the optimal algorithm such that the scheduling decision at each superframe is made regarding its effect on only a few upcoming superframes. The authors' proposed algorithm is considered in two different incoming packet scenarios, deterministic and random arrivals. Simulation results show that for both types of arrivals, there is a negligible difference between their proposed algorithm and the optimal one, i.e. POMDP, in terms of energy consumption and reliability.