Resource allocation with uncertain sensor models

A Type-2 fuzzy logic approach to model sensor uncertainty is developed. Part of a remotely deployed sensor system, an acoustic array used for both classification and localization was determined to have potential issues with its array elements´ locations. These issues can be caused by a number of factors including improper initial deployment, movement of the array elements by external forces, and acoustic obstructions. The uncertainty model used for both tracking and resource allocation itself has uncertainty. Such uncertainty has been modeled using fuzzy logic for the resource allocation effort. The model of the uncertainty was modified using a type-2 fuzzy logic to handle variations in the array placement. This model was then used in the resource allocation technique. The use of uncertainty in the modeling did have a noticeable effect on the system performance with the key difference occurring for mid-level battery power. Also, the initial consideration of the new model for the tracking system is discussed.