With changing of the global structure and the nature of potential adversaries, US ARMY begin to place greater emphasis on fast action with light-equipped troop to sudden attack adversaries’ command center and to make the whole resist force collapse. But our enemies often defend themselves against attack by constructing many large minefields and biological-contaminated areas.  These “danger zones” always take a great deal of time to get identified and eliminated.  To ensure that our troops can safely and effectively find a detour around these zones and can employ a sudden attack, it is imperative to quickly and accurately detect the perimeter of those “danger zones”.

    Traditionally, defining the perimeter of a minefield or a contaminated area has been done by human hand sampling. This method has been proven to be tedious, dangerous and often inaccurate.  De-mining decreases the range of a minefield, and biological or chemical contamination is difficult to contain because wind currents shift harmful particles to adjacent geographical areas.  Thus the dynamics of a contaminated danger zone will be continuously changing while the time-consuming hand-sampling method takes place.

     Developing a system that can accurately determine and monitor the fluidity of the perimeter of a danger zone is important and necessary.  State-of-the-art technology allows this procedure to be executed by robots rather than humans.  Although it have not been verified by the reality data from the multi-robot’s action in the war field, most papers indicate that using multi-robot, or “swarm” of robots with desirable redundancy will yield more benefits than using a single robot in the dangerous area such like minefield. The entire conclusion is based on that the robots will be efficiently deployed on the working area and will cooperate each other to complete the task. However, some papers already verified that using a simple homogenous approach robots have difficulty in completing the task due to the high inter-robot interference.

The overall problem of dynamic perimeter surveillance for swarms of robots is complex and diverse. This work involves different scenarios, contaminated areas, agent types, and other various contributing factors. This proposed project will focus on the evaluation and the implementation of the computer simulation, the development of suitable communication infrastructure for the thousands of small robots exchange information, the cooperation among robot teams to accomplish the task of defining the perimeter of a danger zone as well as surveying any change in that perimeter. At the beginning of this project, the robots in the team are homogenous. However, in the later part of this project the robot team should be composed of various kinds of robots that can handle different situations and the robot team can stand loss of a portion of the robot team by developing robust task-allocation methods.

 In this project, we are not going to deal with any kind of sensing device. We will assume that the sensors on these robots can detect the degree of danger located within a given radius in any danger area. The main objective of this research is to achieve and to devise optimize cooperative strategy and algorithms of multi-robots for detecting and tracking the perimeter and to show through modeling, simulating and experimental validating that intervention task performance by intelligent small robots can be improved by their ability to cooperative with other peers, to gather, to learn and to use their working environment information.