A Task-Oriented Self-organization Mechanism in Wireless Sensor Networks

In wireless sensor networks (WSNs), one attractive and challenging issue is how to allocate tasks efficiently. Most existing studies focused on mapping and scheduling tasks to multiple sensors to ensure the task can be completed before deadline. Nevertheless, one vital aspect is neglected, that is, self-organization of WSN in task allocation. In this paper, we consider the problem of complex task allocation in which a task requires different resources for execution. A task-oriented self-organization mechanism is proposed to guarantee real time in task assignment. Toward this end, the frequently accessed sensors in previous task allocation will modify their structural links that can achieve a better allocation of tasks in the future. Simulation results illustrate significant performance improvements with our proposed mechanism.