A Density-Based Least Squares Formulation of the Sensor Selection Problem for Received Signal Strength Indicator Localization

The overall localization error performance in the received signal strength indicator (RSSI) localization depends on several factors that include the number of deployed sensors as well as their deployment positions within the deployment region. In this paper, we study the problem of selecting deployment positions from a set of candidate positions such that localization errors at some specific points of interest are satisfied in a least square error sense under the constraint of a specific number of sensors. To solve this problem, we propose a simple convex formulation of the sensor placement problem as a least squares problem. Instead of expressing error requirements, which can be difficult to evaluate in practice, we use the local sensor deployment density around a point of interest as a cost function to provide the simplified formulation. A numerical evaluation of the proposed approach shows its validity and usefulness.