Classification of human movements using Micro Doppler features in foliage environments.

Abstract

The focus of this work is measuring and classifying human motion in foliaged environments. A baseline study is performed using a vector network analyzer in open-space and foliaged environments for a variety of frequencies, activities, and formations. The results serve as a justification for a radar prototype’s parameters and a metric to gauge the prototype’s accuracy. The prototype is developed at 2.45 GHz to serve as a low cost and portable alternative to commercially available radars and to the vector network analyzer. Experimentation is repeated in open space and foliaged environments for a more thorough activity list and more participants. Data is post-processed to extract Doppler features and for classification. Results indicate that a low cost radar is capable of distinguishing human motion via human backscatter despite the attenuation, multipath, and blockage due to foliage.