Fault tolerant sensor placement optimization with minimum detection probability guaranteed

Due to the potential of various applications of wireless sensor networks (WSNs) in emergency and security monitoring fields, fault tolerance has become one of the major fundamental concerns in network design. Previous studies in fault tolerant coverage have solely focused on k-coverage as a measurement for the degree of fault tolerance. Unlike the previous works, we have adopted the minimum detection probability at interest points in the entire sensing field after sensor node (SN) failures as the evaluator for fault tolerance in sensing coverage. The objective of our sensor placement optimization is to employ the minimum number of SNs to achieve the desired minimum detection probability even when some SNs have failed. Since our objective function is nondifferentiable, combinatorial and the size of the problem can get fairly large, we have selected Genetic Algorithm (GA) as our optimization tool. Results obtained from GA show that we can guarantee the required minimum detection probability in the worst cases of one and two SNs failures with reasonably low number of SNs.