Intelligent Sensing and Computing in Wireless Sensor Networks for Multiple Target Tracking

With sixth generation (6G) communication technologies, target sensing can be finished in milliseconds. The mobile tracking-oriented Internet of Things (MTT-IoT) as a kind emerging application network detect sensor nodes and track targets within their ranges cooperatively. Nevertheless, huge data processing low latency demands put tremendous pressure on the conventional architecture where is executed remote cloud short transmission distance 6G channels presents new challenges into design topology. To cope with above difficulties, this paper proposes resource allocation scheme to perform delicate node scheduling accurate tracking multitarget networks. dynamic problem formulated an infinite horizon Markov Decision Process (MDP), state space that considers energy consumption, system responding delay, important degree extended. A model-free reinforcement learning applied obtain satisfied actions by frequent iterations, which smart agents interact complicated environment directly. performance each episode evaluated action-value function search optimal reward. Simulation results demonstrate proposed shows excellent terms cost delay.