Combining Multiple Autonomous Mobile Sensor Behaviors Using Local Clustering

Abstract

In crisis situations it may be necessary to re-establish communications via ad-hoc networks of communicating way-stations. Crisis, defence, or surveillance scenarios may require the distribution of sensor units over some region of interest. In both cases the use of communicating and sensing, autonomous, mobile robots will become prevalent in the near future. Existing literature describes various control rules for forming teams of robots into optimally distributed communicating or sensing grids. We show how these robot behaviors exhibit advantages and disadvantages related to the range between robots. Further we describe a new approach which enables several different behaviors to be combined, utilizing the best behavior for any particular inter-robot range. This method is demonstrated using computer model simulations and the improved performance of the combined behaviors, relative to that for individual behaviors, is graphed over important situational parameters