Road Understanding And Path Planning For Outdoor Vision-Based Mobile Robots

Vision-based navigation technology is a new rising research field along with the development of sensors technologies, computer and artificial intelligence which can be used widely in both civilian and military applications. As to Vision-based outdoor mobile robot navigating in non-structural environment, how to identify the road region and the non-road region correctly according to the image information and then plan an appropriate path to the local target is pivotal. This paper discuss upon the main research on outdoor environment road understanding and path planning.To settle the problem of shadow's disturbing in road understanding, some special color spaces are made. The image is segmented into many different sub-sorts following the way of analyzing first, segmenting second and then analyzing again in multi-spaces. The road in the shadow can be segmented into the same sort in this way and so the problem of shadow's disturbing can be solved. Canny edge detector and Hough line detector are used to find the line edges in the vision image and then the real road edges are identified from them by the knowledge on road's edges and color.Road understanding is finished by using the fuzzy knowledge on the road, the kernel region information and the road edges in this paper. The experiments have shown that the algorithm is better than the approach of road understanding using only fuzzy knowledge or kernel information. The algorithm in this paper for road understanding is efficient and robust.As to the problem of path planning, firstly, the target spot is made referring to the road direction and the request of safety, then, a path is planned using both line detecting theory and grid theory. The algorithm for path planning can find the best path on the base of collision avoidance; also, it's efficient and easy for accomplishment.