Network flow formulation of optimal perimeter sensory coverage problem

Abstract

In this article, the perimeter detection optimization problem in field surveillance and target tracking are discussed. The detection range of sensors is assumed to be circular or elliptical. Sensors are also assumed to be associated with a cost factor reflecting their operational characteristics and power usage. We show that the problem of optimal sensor selection can be reduced to a network flow problem and can then be solved using any existing classical methodology. This significantly reduces the computational time of sensory selection problem which in many cases needs to be solved in almost real time basis, every time that the dynamics of the field changes. The field dynamics could change due to such events as wind direction change and sensor failures.