A digital signal-processing technique for compensating ultrasonic sensors

A factor of great importance when assessing the accuracy of ultrasonic rangefinders is the accuracy in the knowledge of the speed of sound, necessary to convert temporal into spatial information. A digital signal-processing technique for making an ultrasonic transducer array capable of automatically compensating for variations in the speed of sound due to temperature or any other atmospheric conditions is proposed and discussed in this paper. The technique is based on an iterative linearized least-squares estimator, namely an extended Kalman filtering algorithm, for processing the time-of-flight measurements from a reference target whose location is only approximately known. In contrast to other well-known techniques, neither additional external sensors for monitoring the environment nor accurately positioned reference targets are required. A sensitivity analysis of the proposed algorithm is performed through a Monte Carlo simulation study. The theoretical analysis provides a clear-cut picture for understanding the merits of the technique under a variety of physical operating conditions. The level of the measurement noise and the correct calibration of the transducers are proven to be the crucial factors for obtaining estimates of the speed of sound at a prescribed level of accuracy, given a fixed temporal interval for collecting the measurements. The main conclusions of the simulation study are confirmed by some real-life results obtained using an experimental tracking sonar device.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>