A linear programming approach to NLOS error mitigation in sensor networks

In this paper, we propose a linear programming approach to the problem of non-line-of-sight (NLOS) error mitigation in sensor networks. The locations of sensor nodes can be estimated using range or distance estimates from location-aware "anchor" nodes. In the absence of line-of-sight (LOS) between the sensor and anchor nodes, e.g., in indoor networks, the NLOS range estimates can be severely biased. If these biased range estimates are directly incorporated into practical location estimators such as the least-squares (LS) estimator without the mitigation of these bias errors, this can potentially lead to degradation in the accuracy of sensor location estimates. On the other hand, discarding the biased range estimates may not be a viable option, since the number of range estimates available may be limited. We present a novel NLOS bias mitigation scheme, based on linear programming, which (i) allows us to incorporate NLOS range information into sensor location-estimation, but (ii) does not allow NLOS bias errors to degrade sensor localization accuracy