Recursive filtering approach to MS locating using quantized to a measurements

A mobile station (MS) location estimation and tracking approach based on the recursive filtering of quantized time-of-arrival (TOA) measurements is presented in this paper. Two techniques for estimating random variables from quantized measurements are investigated first. The TOA measurements may present biases which are unobservable and can determine the location filter to obtain a biased estimate while the location error covariance will be underestimated. The proposed location filter, GF-SKF, combines a Schmidt-Kalman (1960) filter (SKF) with the Curry's (1970) "Gaussian fit" (CF) scheme. The measurement biases are considered in the SKF model only through the effect of their covariances on the location error covariance. The GF scheme is meant to accommodate the quantized measurements in the filter model. A measurement noise covariance estimation method is described and a simple technique for alleviating the influence of the non-line of sight (NLOS) range errors over the state estimate is specified. Simulations with GF-SKF show a reasonable accuracy which depends on the quantization step.