An Efficient Technique for Locating Multiple Narrow-Band Ultrasound Targets in Chorus Mode

A basic problem in time of arrival (TOA)-based locating systems using narrow-band ultrasound (NBU) is how to improve the location update rate for tracking multiple targets. It is challenging because each ongoing ultrasound (US) signal occupies the channel for a rather long time because the slow propagation speed in air and it is hard to encode information in the narrow-band US. In this paper, we investigate to allow multiple NBU targets to transmit signals concurrently and to determine their locations by signal processing, which is called locating in chorus mode. The key observation is the signal interference characteristics of the concurrently chorusing targets. Based on it, the necessary and sufficient conditions for the receivers to determine the TOAs from multiple concurrently transmitting targets are investigated. However, because NBU cannot encode the target's ID, the detected TOAs are lacking labels of the target ID, causing ambiguities in location estimation. We exploited both historical consistence and self-consistence methods to narrow down the possible IDs of the TOAs and proposed probabilistic particle filter algorithm to disambiguate the motion trajectories of targets based on the targets' motion pattern consistency. A prototype of chorus-mode NBU locating system was developed. Extensive evaluations of both simulations and prototype experiments showed the effectiveness of the proposed theories and algorithms. In the testbed experiment, chorus locating provided 300% refreshing rate improvements compared with exclusive locating method, while the accuracy is still kept.