Task-Oriented Self-Organization of Ad Hoc Sensor Systems

An ad hoc network is a collection of perhaps wireless mobile nodes dynamically forming a network without the use of any pre-existing network infrastructure or centralized administration. Resource limitations of sensors deployed in an ad hoc setting necessitates careful consideration of the choice of sensors. In this paper, a distributed task-oriented self-organization algorithm for sensors in such a network to organize themselves according to the task being announced is proposed. The proposed algorithm uses appropriate attribute vectors to describe the specifications of tasks and sensors. When a task is announced, a distributed leader election algorithm is used to select the sensor that is best matched based on minimizing a distance measure between the announced task and sensor specifications. The selected sensor then uses a task decomposition method to generate, and then announce, residual tasks that 'cover' the matching error. In this manner, sensors are sequentially selected until all residual tasks are 'covered'. All sensors selected by the algorithm form a sensor group for the originally announced task. The proposed algorithm incorporates criteria to avoid picking sensors that 'over-satisfy' the announced task. To improve the performance of the algorithm and to more efficiently utilize the newly selected sensor capabilities, a residual task management method that adjusts the previously generated residual tasks as new sensors are selected is also incorporated.