Autonomous Search and Rescue (ASAR) with Multimodal Identification for Large Scale Operations

This proposal is for autonomous search and rescue (ASAR) with multimodal identification in large scale operations. In disasters such as Katrina, Tsunami, bushfires, etc., many people die.  Find and extract in urban areas is currently very human-intensive and slow.  The technical problem is in knowing where to look. Even with lots of vehicles there is too much area to cover and too much reliance on humans for sensor interpretation. Current state of the art is mostly R/C vehicles with human coordination.  Advances in sensing technologies and in flying cars will provide the opportunity.  Our innovative contribution is RF localization with a Bayesian particle filter coupled with computer vision based on unsupervised online learning. The result is automated rapid location and extraction of those in need. This is a game changer for rapid disaster rescue and for other applications as well including emergency supply delivery.