Constrained particle swarm algorithms for optimizing coverage of large-scale camera networks with mobile nodes

Proper sensor placement is crucial for maximizing the usability of large-scale sensor networks. Specially, the total sensible area covered by a sensor network can be maximized if we optimally arrange all sensors. To address this coverage optimization problem, this paper studies a typical sensor network—camera network. In this network, both locations and orientations of the cameras can be adjusted. An interesting constraint is the moving distance limitation. It transforms the optimization into a constrained problem. To tackle this problem, we investigate as possible solutions three variations of the particle swarm optimization (PSO) algorithm, namely the absorbing PSO, the penalty PSO, and the reflecting PSO. They are tested against several benchmarks. The experiments show that the PSO can be effectively applied on optimizing the coverage of the constrained camera network. And it can be easily adapted for coverage optimization of general sensor networks. The statistical analysis shows that the performances of the above three algorithms are in descending order. The results further prove that the absorbing PSO is an optimal choice for improving the coverage of the aforementioned sensor network.