Optimal Node Distribution in Wireless Sensor Networks Considering Sensor Selection

Local Positioning Systems are attracting the research interest for addressing the limitations of the Global Navigation Satellite Systems over the last few years. These ad-hoc deployments of sensors depend on the optimal disposition of the nodes for reducing the localization uncertainties. This requires the addressing of the Node Location Problem which has been considered as NP-Hard. Therefore, a heuristic solution to the problem is recommended. In this paper, we first introduce the Sensor Selection Problem for maximally reducing the error bounds of the entire coverage region during the Node Location Problem optimization. We propose a genetic algorithm optimization for a Time Difference of Arrival Local Positioning System architecture in which we compare the error bounds achieved for the employment of the five closest sensors to the target versus the best combination of sensors under coverage for calculating the target location which includes the Sensor Selection Problem. Results show that the consideration of the Sensor Selection Problem allows a reduction of the localization uncertainties by 16.7% to traditional Node Location Problem optimizations.