Asynchronous distributed searchlight scheduling

This paper develops and compares two asynchronous distributed scheduling algorithms for multiple controlled searchlights in nonconvex polygonal environments. A searchlight is a ray emitted by source location that (i) cannot penetrate the boundary of the environment and (ii) undergoes controlled slewing about its source location. Evaders move inside the environment along continuous trajectories and are detected precisely when they are on the searchlight ray at some time instant. The objective is for the searchlights to detect any evader in finite time and to do so using only local sensing and limited communication among them. The first algorithm we develop, called the distributed one way sweep strategy (DOWSS), is a distributed version of an algorithm described originally in 1990 by Sugihara et al [Sugihara, K., et al., 1990]; this algorithm may be slow in "sweeping" the environment because only one searchlight slews at a time. Second we develop an algorithm, called the parallel tree sweep strategy (PTSS), in which searchlights sweep concurrently under the assumption that they are placed in appropriate locations; for this algorithm we establish linear completion time.