Classification Induced Distributed Sensor Scheduling for Energy-Efficiency in Underwater Target Tracking Sensor Networks

This paper develops a distributed sensor scheduling methodology that utilizes target classification decisions to govern the number of active sensors selected around the target in the deployment region. This approach utilizes a distributed supervisor on each sensor node to control the multi-modal operating state of the node. A distributed sensor selection method is proposed that dynamically adjusts the number of active sensor nodes based on the classification decision. The proposed method is simulated and validated to show that incorporating classification into the control loop significantly conserves energy reserves while still allowing for accurate target estimation.