A multiple candidate time of arrival algorithm for tracking nodes in multipath environments

Tracking systems based on the measurement of time of arrival (TOA) assume that the measured signal travels directly between transmitter and receiver. In complex radio propagation environments this direct path signal is often weak compared to multipath signals and there is a tradeoff in the TOA algorithm between sensitivity to the weak direct path and false detection due to noise, sidelobes, and other artifacts. Conventional TOA algorithms return a single TOA, which can be early due to false detection or late due to an undetected direct path, and these errors degrade tracking performance. In this paper a novel approach to this problem is proposed in which tracking performance is improved using multiple candidate TOA values. In particular a set of TOAs are extracted from the channel impulse response for each received signal and converted to a set of range values. A decision as to which among the set of range values is due to the direct path is deferred to the tracking algorithm, which uses a probabilistic soft-decision approach. Experimental studies conducted using a wireless network demonstrate that the 90% percentile absolute position error is reduced from 3.3 m to 1.3 m and the relative position error is reduced from 1.3 m to 0.5 m.

► A multiple candidate TOA method is proposed for tracking in multipath environments.
► An algorithm is presented to extract multiple candidate TOAs from the CIR.
► Candidate TOAs are used in a soft-decision filtering method to track mobile nodes.
► A method to calculate probabilities to combine multiple hypotheses is presented.
► Experimental results show effectiveness of the approach in real world scenarios.