A self-stabilizing clustering algorithm with fault-containment feature for wireless sensor networks

In this paper, the clustering of wireless sensor network has been equaled to constructing maximal independent set in graph theory, and a clustering algorithm with “self-stabilizing” and “fault containment” properties, which is considered as a critical feature in the issue of fault tolerance for distributed systems, is proposed. Existing comparable methods include no fault containment and their design is based on assuming a “centralized scheduler”. The proposed algorithm is recovered from single fault configurations by space and time complexity of O(1), and work under policy of unfair distributed scheduler which has the maximum matching with operation environment of sensor networks. The “self-stabilization” and “fault containment” properties of algorithm will be proved by formal reasoning; Simulation results also show that regardless of the number and concentration of nods, the suggested method in addition to quick recovering against small-scale faults, will improve convergence time compared to previous methods. The creation of efficient clustering construction, reducing the numbers of updating messages and stabilizing by minimum change in the clustering topology structure, are other advantages of this algorithm.