Joint location and parameter tracking of mobile nodes in wireless networks

Abstract

Achieving very high localization accuracy in wireless networks that measure time of arrival (TOA) is a challenging task, especially when low cost hardware is used. The local oscillators used in the wireless nodes will drift over time, which will result in frequency and time offset between different clocks. Synchronization between the clocks must be maintained to obtain highly accurate TOA measurements. The delay in the radio frequency electronics can also vary with time and environmental variation and for accurate localization this variation must be accounted for as well. Although calibrating these parameters prior to the operation of the network is one solution, it is not an option for networks that operate for longer periods of time or those that are rapidly deployed. In this paper we propose an algorithm that jointly tracks the frequency offset and radio delay of all the nodes in the network along with the location of the mobile nodes. The algorithm calculates the round trip delay measurements, which eliminates the need to estimate the time offset. We also derive the posterior Cramer Rao lower bound (PCRLB) for the joint estimation problem, which provides a bound on the maximum peiformance achievable. Through simulations we show that the peiformance of the proposed algorithm is in close agreement with the PCRLB for both the non-kinematic and kinematic state estimation.