AN INFORMATION ENTROPY APPROACH TO SELF-POWERED SENSOR NODE OPERATION

摘要

Machinery failure prevention is vitally important in many industrial, commercial, and military systems. Continuous machinery monitoring for condition assessment can be effective in identifying faults before a machine failure occurs and permit safe machine shutdown or timely machinery maintenance. The deployment of wireless sensor nodes for condition monitoring is growing rapidly due to the relatively low cost of distributed sensor nodes and the ability to avoid the cost associate with running wires to remote sensors. An enabling technology is the use of power scavenging to harvest energy from the environment. A self-powered sensor node eliminates the need to run power connections to the sensor node and also eliminates the need for batteries. A self-powered sensor node provides additional challenges for reliable, unattended sensor node operation. An operating strategy has been developed for wireless self-powered sensor nodes based on an information theoretic approach to optimize sensor node functions. An integrated information and power management strategy is established to maximize the amount of useful information gained for a unit of energy with the constraint on insuring power for sensor node survival. An information-theoretic approach to optimize sensor node operation promises to provide maximum relevant information while operating in dynamic, unexpected environments.