Localization in NLOS environments using TOA, AOD, and Doppler-shift

In this paper, we consider localization in non-line-of-sight environments. We present algorithms to estimate the most likely position or the most likely trajectory of a mobile station given a sequence of time-of-arrival, angle-of-departure, and doppler-shift observations. Discretized positions are considered for computational feasibility of the proposed algorithms. The mobility model of the mobile station is defined by a Markov model. The most likely position and the most likely trajectory of the mobile station are efficiently computed using the forward-backward algorithm and the Viterbi algorithm, respectively given the whole sequence of observations. An online Bayesian recursive algorithm which estimates the most probable position of the mobile station is also proposed. It is shown, through simulations, that we can get a satisfactory performance given that the mobile station is not stationary during the course of the tracking time.