Research On Robot Navigation In Unknown Environment

Robot Navigation is an important and basic problem for researching the technology of robot controlling and it is always the researching focusing of mobile robot technology. Moreover, robot navigation in unknown global environment is the difficulty of the field. In such kind of environment, robot knows nothing about whether there exists any obstacle in the environment and the coordinate of any obstacle. Being lack of future knowledge of the environment, robot can’t plan a global path. Therefore, robot must get detailed information about obstacles in the environment through the sensor and build the environment model on the basis of which it can plan the navigation path. When the working environment of robot is far larger than the detecting range of its sensor, robot can only detect the information of the local environment in its vision through the sensor but not the global information. So robot can just plan a local navigation path in its vision instead of getting the global path. Further researches on the real-time detecting of the environment through a real sensor, environmental model establishing and the realization of global navigation using local paths are still needed. Besides, although there are substantial achievements on robot path planning and navigation in both the known and the unknown environment, few of those achievements has practical value as most of them are based on PC simulation experiment.Aiming at above problems, the robot rolling navigation based on the improved A*algorithm is studied firstly in this paper. Simulation experiment is also conducted and shows the simplicity and effectiveness of the improved A*algorithm. In order to detect obstacles in real environment and determine their coordinate and then build an environmental model, analysis of principles of laser sensors and researches on the realization of the algorithm and the programming of getting the information of obstacles in the real environment through laser sensors are conducted in the paper. These algorithms refer to the preprocessing of the information about obstacles and exchanges of coordinates through which the environmental model of robot navigation in real environment can be built. The algorithm of extend point (EPA) of obstacles and the concept of dynamic walk step are proposed in the paper considering that the laser sensor can only detect front edges and angles of obstacles without determining the size of the space occupied by obstacles. The wheeled-pattern robot global navigation algorithm based on the above environmental model is proposed in the paper to solve the navigation problem as the working environment of robot is far larger than the vision of it. In addition, to realize the autonomous navigation of robot, there are also motion control principles and control localization methods and specific programming methods in the paper.To verify robot navigation algorithm based on the laser sensor, the design plan and the experiment process of doing practical navigation experiment using MT-R robot as well as detailed introduction of the experiment’s key problems including steering control and dynamic step determination of robot are all listed in the paper. The experiment shows that robot can pass the area with obstacles and reach the target point safely and smoothly along a relatively optimal path using the navigation algorithm proposed in the paper.