Parameterization of Acoustic Signatures in Ultrasound Images for the Detection of Human Presence by Autonomous Vehicles

We address the problem of human detection for autonomousvehicles. Instead of using an optical system, we propose toemploy an acoustic 2D array to reliably obtain an image of ahuman in a 3D spatial power spectrum which is independentof lighting conditions and uses cheap ultrasound sensors.We show that humans have a distinct acoustic signature andpropose to model the echoes from reflecting parts of objectsin the scene by a Gaussian-Mixture-Model. When the model is fitted to the acoustic image, it is straightforward toobtain geometric relations between the present echoes andrepresent the acoustic signatures in a low-dimensional parameter space. We present results based on real data measurements that demonstrate that different objects can be reconstructed from the data and discriminated. The obtainedparameter space can then be the basis for subsequent detection and classification of humans