An Investigation of a Dynamic Sensor Motion Strategy

This paper considers a dynamic sensor coverage problem in which a single mobile sensor attempts to monitor multiple sites. Sensor motion is modeled using a discrete time, discrete state Markov process. State dynamics at each site are modeled as a linear system. A stochastic simulation is used to demonstrate previously derived theoretical conditions under which a single sensor is or is not sufficient to maintain a bounded estimate of the state of every site. Observations are made about the relationship of sensor motion to system dynamics. A strategy is presented to find a good sensor motion model based upon the system dynamics and to determine the convexity of the solution set.