Predictive-TDMA: A Markov Chain Based MAC Protocol for Mesh Networks

One of the most frequently concerned metrics for evaluating the performance of MAC protocol is throughput. In this paper, we propose a novel Markov Chain based TDMA MAC protocol which is called Predictive-TDMA (P-TDMA) to achieve high performance of throughput for Mesh Networks. In order to give a more effective time slot assignment strategy rather than even assignment, we use Markov Chain to predict the possibility of an end-user device´s future working condition, whether it´s more likely to be idle or busy. Then based on the ratio of all end-user devices´ possibilities of working condition, we assign a proper number of time slots to each end-user device to maximize throughput. Through simulation, the results show that under the scenario that the amount of data is randomly requested by each end-user device, throughput of our design is around 10% better than even assignment strategy. And under the scenario that some end-user devices request much more amount of data than others, which happens more frequently in realistic situations, throughput of our design can be 200% better than even assignment strategy.