Introduction

  The general setting of a pursuit-evasion problem involves some bounded environment which may contain one or more evaders that can freely move around at arbitrarily high speeds. The task is to to implement a movement strategy for one or more pursuers that will guarantee that all of the evaders are at some time in some way detected by the pursuer. Most of the suggested applications focus on the need to ensure that some area is free of unfriendly actors (such as criminals, or freely moving pieces of litter (??)), or that all friendly actors in a given area are found (such as injured people, or freely moving easter eggs (??)). In this project the case of a single pursuer is considered, with 360 degree line of vision evader detection capability.

Although the pursuit-evasion problem has experienced many successful results in recent years, most of them are related to the case where computed results require execution by a robot capable of precise movement and localization. In addition, the map of the environment is known prior to determining a solution.

In this project, the robot required to execute the solution requires only limited movement options, and does not need precise localization. Furthermore, an attempt is made to advance toward the solution of the pursuit-evasion problem when the map is not known.