Joint Sensor Selection and Placement in Partially Controllable Localization Networks

Abstract

This paper investigates the joint sensor selection and placement (JSSP) problem in a time difference of arrival (TDOA)-based partially controllabel localization networks, which consists of the existing network (E-Net) and the supplementary network (S-Net). The quantity of localization-enable nodes (LENs) should be well designed to save energy so that the sensor selection in E-Net and sensor placement in S-Net are worthy of study. Therefore, we introduce a Boolean vector to formulate the JSSP optimization problem that minimizes the localization error for the source under the constraints of LENs quantity and placement area of S-Net. Since the problem is highly non-convex to the decision variables, two heuristic algorithms, block enumerative comparison (BEC) algorithm and iterative swapping greedy (ISG) algorithm, are proposed to approach the sub-optimal JSSP solutions. The simulation shows that the localization accuracy of the proposed algorithms is always close to the benchmark algorithm with the varying TDOA measurement noise strength and the quantity of the LENs.