Acoustics based sensor positining system

The proposed acoustic positioning system addresses the challenge of accurately determining the locations of acoustic sensors in GPS-denied environments, achieving localization within 1% error. In this setup, sensors are deployed by an unmanned ground vehicle (UGV) without prior knowledge of their positions. The system employs an iterative acoustic technique where each sensor broadcasts signals detected by neighboring sensors. The receiving sensors calculate the signal's time-of-flight (ToF) and relay these data to a central computational node. The process repeats until all sensors have transmitted, allowing the central node to compute precise locations by aggregating ToF data. Relative distances between sensor pairs are derived from ToF measurements, and Multidimensional Scaling (MDS) techniques transform these distances into the exact spatial configuration of the sensors. To further enhance accuracy, the wind vector affecting sound propagation is estimated via least squares optimization. Finally, a nonlinear optimization procedure, solved using the Levenberg-Marquardt algorithm, refines both the sensor positions and the wind vector by minimizing discrepancies between measured and calculated acoustic delays.