Authors
Christopher Wahl and Bertrand Hochwald
Abstract
Efficient spectrum sensor location selection is an important aspect of the radio mapping process. Well-placed spectrum sensors allow for more accurate radio map estimation and can reduce costs by lowering the number of needed sensors for a certain accuracy requirement. Previous work on sensor location selection looked at minimizing the Geometric Dilution of Precision for localization applications or assumes a correlation model for the propagation. In this work, we propose using a Maximum Entropy heuristic on the spectrum power measurements for selecting sensor locations. This approach is fast to run and provides sensor locations that improve radio map estimation performance.