Target Tracking in NLOS Environments Using Semidefinite Programming

Abstract

In this paper, the problem of target tracking in non-line-of-sight (NLOS) environments is investigated. Target tracking has many commercial, civilian, and military applications. The accuracy of target tracking is highly affected in indoor environments where the majority of connections are NLOS. A novel tracking estimator based on semidefinite programming (SDP) with ability to mitigate the NLOS propagation is derived. Requiring no statistical information about the NLOS propagation, the proposed SDP algorithm estimates the NLOS biases jointly with the location and velocity of the target. The performance of the proposed estimator is evaluated through computer simulations where ray tracing is used to simulate the NLOS biases. It will be shown that the proposed SDP estimator outperforms the classic extended Kalman filter as well as other recently proposed estimators in NLOS environments.