Research On Multi-model Autonomous Navigation Method Of Patrol Vehicle On Airport Enclosure Road

It's critical to research on autonomous navigation method for patrol vehicle completing autonomous patrol around the airport enclosure road. The airport enclosure road can not be described by a unified mathematical model because of its semi-structured complex environment. Traditional navigation methods based on single sensor are limited to semi-structured road. Therefore, the study on autonomous navigation method using multi-sensor fusion is significant to the application of vehicle in airport enclosure patrol.Firstly, for the demand of computer vision in application of visual navigation, a fast airport enclosure road detection algorithm based on region of interest and geometric constraints is proposed. In the algorithm, Hough Transformation is used to extract line segments from origin road image, and the road boundary line segments can be acquired from line segment candidates according to geometric characteristics and inter-frame continuity constraints; and then the regions of interest are used to reduce the complexity of algorithm;finally, inverse perspective mapping is utilized to project the road boundary line segments from camera frame onto the ground plane.Secondly, airport enclosure road boundary detection algorithm based on laser radar is raised to cover shortages of visual navigation. The algorithm performs clustering, fitting and combination procedures on range data of laser radar to extract road boundary line segments,and then the results are applied to vehicle navigation on airport enclosure road.Thirdly, a road navigation algorithm based on passable region is presented which combines target bearing with road boundaries. This navigation algorithm regards road boundaries as virtual obstacles, and local path planning method based on drivable area is employed to plan an optimal collision-less path according to environmental information. Then,vehicle can be controlled transversely and vertically under optimal path.Finally, the navigation software is programmed, and the global autonomous navigation experimental results show that the autonomous navigation method is effective and reliable.