Wireless Sensor and Actuator Networks: Algorithms and Protocols for Scalable Coordination and Data Communication

9.3 SELECTING BEST ROBOT(S) WHEN COMMUNICATION COST IS NEGLIGIBLE

Most existing solutions referring to multirobot coordination for single or multiple events, single or multiple robot, single or multiple task to each robot and so on, are centralized. One of robots, or a central entity, gathers all the information from other robots and makes a decision. Communication cost for gathering information in case of multihop robot networks is rarely considered. The decision maker can theoretically make an optimal decision by gathering necessary information with little cost because communication cost is assumed negligible. Indirectly (since no details of communication protocols used are given), a complete graph (where each robot is within communication distance to any other robot) is assumed. Centralized solutions usually define coordination problem as an ILP problem. Main advantage of a centralized solution is that, theoretically, optimal solution can be found. However, centralized solution features high computation and communication overhead, bad scalability and slow responsiveness. Moreover, the actual cost for communicating is ignored, especially for large robot networks. It is further not clear how robots communicate if the graph is not complete one. Centralized solution also has low fault tolerance if leader is malfunctioning in any way.

For static actors, Melodia et al. (2007) stated a mixed integer nonlinear program (MINLP) formulation to maximize average remaining energy of (multi) ...