Comparative Study of Joint TOA/DOA Estimation Techniques for Mobile Positioning Applications

In this paper, we present a comparative study of the performance of the joint time of arrival/direction of arrival (TOA/DOA) estimation techniques - estimation of signal parameters via rotational invariance technique (ESPRIT), joint angle and delay estimation (JADE) and space-alternating generalised expectation maximisation (SAGE) - in the context of a strong multipaths model. We focus on a number of performance issues that are of particular interest to the mobile positioning application, such as the estimation accuracy of the line of sight (LOS) component within a strong multipaths environment and their robustness to channel model errors. In addition, we also analyse the computational complexity of the algorithms with respect to the number of paths in the channel and the number of antenna elements. Through extensive simulations in MATLAB, we have shown that: 1) SAGE outperforms the other two algorithms in terms of TOA/DOA estimation variance in a strong multipaths environment, thus can give much more accurate position estimate provided the LOS component is relatively strong. 2) SAGE is more robust to channel modeling errors. 3) SAGE's computation complexity exhibits a linear relationship with the number of channel paths (while ESPRIT and JADE are independent) while being unaffected by the number of antenna elements.