Collision detection scheme for unmanned aircraft using passive Panospheric(TM) Imaging sensor

With recent improvements to a PanosphericTM1 Imaging (PI) system at the Canadian Defence Research Establishment Suffield, there is now active consideration in using commercially available technology as a passive collision avoidance sensor for Unmanned Aircraft. This could provide the aerospace industry with a partial solution to the problem of piloted and remotely operated aircraft operating in the same airspace. The scheme presented in this thesis detects and tracks all targets that maintain a constant bearing relative to the observing platform over a period of time. Detection is performed by comparing successive digital frames, reducing noise, and isolating targets of interest. A tracking function further eliminates erroneous targets. When coupled with the guidance system of an aircraft, the algorithm would be capable of autonomous operation in an emergency situation. Under normal operation, the sensor and algorithm would provide additional situational awareness to the operator, significantly increasing the 'See and Avoid' capability of the Unmanned Aerial Vehicle. Simulation results are provided to evaluate the effectiveness of the collision avoidance scheme.; 1"Panospheric" is a registered trademark of Panoramic Viewing System Inc In Canada and the USA.