Adaptation of acoustic sensor orientation based on sensor characteristics for improving tracking performance

This paper addresses the adaptation method of acoustic sensor orientation for improving tracking performance. The orientation adaptation is based on an empirical reception characteristics of the sensor. The study assumes that an object is transmitting Gaussian noise-corrupted signals. In a 3-D space, the observed measurements by the sensor are projected to each of 2-D plane, and the projected measurement variance is significantly reduced by changing the sensor orientation. This paper demonstrates the derivation of the projected measurement variances, and compares to the original measurement variances. The performance of the proposed method is evaluated with sequential Monte Carlo (SMC) techniques.