A Multi-Layer Self-Healing Algorithm for WSNs

The implementation of Wireless Sensor Networks (WSNs) is a challenging task due to their intrinsic characteristics, e.g., energy limitations and unreliable wireless links. Considering this, we have developed the Disjoint path And Clustering Algorithm (DACA) that combines topology control and self-healing mechanisms to increase the network lifetime with minimum loss of coverage. Initially, DACA constructs a tree that includes all nodes of the network by using the Collection Tree Protocol (CTP). This tree is an initial communication backbone through which DACA centralizes the information. Then, DACA builds a set of spatial clusters using Kmeans and selects the Cluster Heads (CHs) using Particle Swarm Optimization (PSO) and a multi-objective optimization (MOO) function. Subsequently, DACA reconstructs the tree using only the CHs. In this way, DACA reduces the number of active nodes in the network and saves energy. Finally, DACA finds disjoint paths on the reconstructed tree by executing the N-to-1 multipath discovery protocol. By doing so, the network can overcome communications failures with a low control message overhead. The simulations on Castalia show that DACA considerably extends the network lifetime by having a set of inactive nodes and disjoint paths that support the communication when active nodes die. Besides, DACA still maintains a good coverage of the area of interest despite the inactive nodes. Additionally, we evaluate the shape of the tree (i.e., the average number of hops) and the risk of connection loss of the network.