| The appearance of the robot has changed the mode of production and way of life of humansociety. In recent years, with the development of robot technology, the robot is more widely usedin military and scientific fields. When the robot working in the dangerous enviroment which isinaccessible to human beings, it must have the ability to move completely independent, so thenavigation technology for mobile robot has gradually become the one of the hotspots of robotics.Robot navigation technology in statics environment is relatively mature, and has obtained manyachievements, but the research in the unstructured dynamic environment has not formed theperfect and unified theory of knowledge. Therefore, the research on the navigation in thedynamic environment based on common platform is of great significance.At present, domestic navigation system with independent research and development, whenthe algorithm structure, software, hardware and working environment changes, the system needsto be re-development and the code reuse rate is almost zero. According to the shortcoming, anopen-source robot operating system is used as software platform and a two-wheeled robot as theexperimental robot. This paper studied on two key problems of robot navigation in unknowndynamic environment: the simultaneous localization and mapping problem (SLAM) and the pathplanning problem. A hybrid method based on A*algorithms and genetic algorithms is proposedto solve the optimum path planning for mobile robots and the simulation results show that themethod has good convergence in the grid map. Dynamic obstacle avoidance algorithm usingDWA (Dynamic Window Approach) algorithm, and improve the abnormal situation. Finally,experiments were carried in ROS and the results proved the feasibility of the scheme. |
PDF Full Text Request