Sensor Location through Linear Programming with Arrival Angle Constraints

Abstract

In previous work, we established a linear programming framework to determine sensor location from measured link distances between neighboring nodes in a network. Besides providing greater accuracy compared to other techniques, linear programs in particular suit large networks since they can be solved efficiently through distributed computing over the nodes without compromising the optimality of the objective function. This work extends our framework to determine sensor location from measured arrival angles instead. An extensive simulation suite substantiates the performance of the algorithm according to several network parameters, including noise up to 15% the maximum error; the proposed algorithm reduces the error up to 84% depending on the noise level.