On Topology Dependent Mitigation of NLOS Ranging Errors in LSE Positioning

In this work, effects of the anchor nodes' geometry on the positioning error are evaluated for a least squares estimation (LSE) of the linearized multilateration problem. It is shown that the linearized solution can produce an unbiased average position estimate, if all involved range estimates are biased (NLOS). A theoretical discussion contains the requirements on the reference topology and range error biases in order to obtain an unbiased solution. Several positioning scenarios are simulated in which such a behaviour occurs to verify the results of the theoretical discussion. In the final part of the work the properties of the range bias in the NLOS scenario are analyzed for measured UWB channels.