Contour estimation using collaborating mobile sensors

A mobile wireless sensor network may be deployed to detect and track a large-scale physical phenomenon such as a pollutant spill in a lake. It may be called upon to provide a description of a contour characterized by an isoline of a specific concentration value. In this paper, we examine the problem of tracing a contour of a particular concentration within a bounded region of varying pollutant concentration using a network of mobile sensors. Since controlled movement of sensors within a given region is known to improve the overall quality of measurements by reducing sensing uncertainty, we explore various ways of guiding a set of mobile sensors optimally so as to surround and trace the contour. We formulate the contour estimation problem as a nonlinear multi-extremal optimization problem. We use accuracy and latency as performance metrics and show that in majority of the cases our proposed strategy based on collaboration of sensors delivers the best performance