ResQuad: Toward a semi-autonomous wilderness search and rescue unmanned aerial system

In a search and rescue scenario, an expeditious response is critical since the radius of the search area increases by roughly 3 km per hour while the life expectancy of the lost individual decreases. To find the object of the search as quickly as possible, the incident commander creates an overlay probability map identifying high priority search areas and develops a search plan that allocates search assets to each area to maximize the probability of success. Conventional aerial searches have the ability to cover a large area quickly, but are costly and require highly trained personnel to operate. In this work we present results of a proof of concept Search and Rescue Unmanned Aerial System, ResQuad, that capitalizes on the advances in drone industry while allowing for reduction in cost and retaining the advantages of aerial perspectives. Given the probability map from the search commander and the system constraints, a near optimal path is determined by the Path Planner in our system ensuring the coverage of high probability areas. While the Unmanned Aerial Vehicle (UAV) follows the path, an on-board computer with a camera processes and collates the data using an anomaly detection algorithm, saving the results to its memory for the base to verify in real-time.