Fusion of point clouds for obstacle tracking during airport ground operations

Abstract

During airport ground operations, one of the responsibilities of the pilots is to look out for obstructions. In large commercial aircraft, it can be challenging to observe certain regions, as the outside view from the cockpit seats can be limited. This becomes even worse in low visibility conditions and at night when objects are less noticeable. Failure to notice obstructions and to maintain an adequate distance of separation can lead to collisions. This paper addresses this issue by proposing an obstacle detection and tracking technology to assist pilots during taxi. Data, in the form of point clouds, is acquired from a stereo-vision system and a LIDAR sensor. It is then processed and analyzed in order to detect obstacles. Detected obstacles are fused and tracked with Particle Filters and Occupancy Grids. Finally, obstacle information is represented on a bird’s eye view display. Experiments were carried out with the aircraft approaching a target obstacle and the performance of the system was assessed in different illumination and visibility conditions. The results obtained show that the system can reliably detect obstacles and represent them on the proposed display.