Reinforcement Learning Based Geographic Routing Protocol for UWB Wireless Sensor Network

Utra-Wide Band (UWB) technology can provide high data rate and accurate localization at low energy cost. It is considered to be very useful for wireless sensor networks. We propose a reinforcement learning based geographic routing algorithm for UWB sensor networks. A comprehensive reward function is proposed in the learning algorithm to consider node energy, delay, routing failure, and network lifetime. The algorithm performance is evaluated in NS2 and compared with GPSR. Simulation results demonstrate that the proposed algorithm can improve network robustness and network lifetime to be 75% to 213% better than GPSR