A particle filter for robust calibration of RF ranging systems

This paper describes a robust method for calibrating radio-frequency ranging systems in indoor environments. An approach for outlier rejection is proposed, based on modeling the distribution of the distance-measurement results as a Gaussian-uniform mixture. Using this model, a batch maximum-likelihood and a recursive particle filtering estimator are implemented. The accuracy and robustness of the approach are evaluated by numerical simulations and by comparison with the Cramér-Rao Lower Bound. Finally, the approach is validated on experimental data obtained from a in-house developed Ultra-Wideband indoor ranging system.