Research On The Methods Of Path Recognition And Path Planning

With the development of science and technology, a variety of robots showed up, that's the reason that more and more people pay attention to the robot. However, Robot Competition of Chinese college students is attracting a lot of fans after Robocup Open Competition. The robot can be regarded as mini-tour of China's tourism, which coincide the problem of a short holiday now, how can reasonably arrangements to visit more attractions. So it also greatly arouses our interest. In this paper, the robot tour of China is the method of studying the path planning and path recognition.The basic idea is that, firstly, according to the competition requirements, estab- lishing the basic interface platform of the competition is to store the image data and make prepare for the traverse. Then, It can follow the classic traversal methods to get a better travel routes through improving the method to the traversal of the spots. On the other hand,the marching speed of the robot is decided by the situation of scope of the robot. So extracting the situation of the road, processing the addition of noise which caused by the environment and redundant information is to get a better road goal. Lastly, extracting and analysis the several situations of road in front, using the method of corner detection to determine the number of branches, and make the robot choose a better route.Realization process is as follows:1. The storage of the map.First of all, establish the basic interface platform competition, then analyze and compare preponderance and defects of the several common storage methods.2. The traversal algorithm.Commonly used in artificial intelligence, first introduced several traversal algorithm, and the complexity of each algorithm and application analysis and comparison. According to the project needs to be improved, get a different traverse routes. Then the total distance is as the basis to judge the merits of methods, and ultimately determine a best path. Through experiments, the blindness of the high depth-first algorithm, for the establishment of the different nodes will be different; the Hamiltonian algorithm, although can walk, but the comparatively easy problems, in particular the increase in the node, the implementations increasing difficulty; and genetic algorithms, more suitable for this global search, better able to achieve. So obtain an improved genetic algorithm to determine the route to the game in the end.3. The image acquisition and processing of Road.It is the premise of identifying ahead of roads which to acquire the image info- rmation of the ahead of road .This paper considers sufficiently the environmental interference of the road while the robot goes forward when using the camera to obtain the image of the ahead of road. Through some methods such as median filtering, threshold segmentation and edge detection to capture a set of connecting way, we can process ithe mage of road.4. The identification of the road branch numberAccording to the offered court to certain the branch numbers while the robot running, we can make the robot judging which road. Using thoughts of corner detection, this paper did some research to determine how many branch numbers. Through the actual image and ideal graphics for identification, we find we can define the branch number basically.Through the above research, it settles the base for the path planning, road image processing and path recognition of the robot. It also offers a reference for practical application.