| As the most widely used type of robot, intelligent mobile robot is playing a more and more important role in aerospace, medical rehabilitation, industry, and other industries. Autonomous navigation, as an key manifestation of the intelligent mobile robot, has become a research hot spot in recent years. Therefore, study on robot autonomous navigation is significant to theoretical research and practical applications.Via deep analysis of technology of robot navigation, laser range finder becomes one of the most important sensors in robot navigation with the advantages of high precision, strong anti-interference capability. An overall framework of navigation system based on laser range finder and ROS is established using the SLAM(Simultaneous Localization and Mapping) navigation methods.For the problem of large amount of particles required and the inconsistency of the particle filter in mobile robot Fast simultaneous localization and mapping(FastSLAM), an improved unscented particle filter algorithm is proposed. The novel algorithm utilizes iterated sigma points particle filter to generate more accurate proposal distribution, which fuses the robot’s odometer information and laser information into sequential importance sampling routine through iterated update processing. The algorithm effectively improves the filter consistency and the state estimation accuracy, while it requires smaller number of particles. In addition, the precision of robot localization and map building is improved. Experiments prove that the improved SLAM algorithm can effectively enhance the real-time performance and robustness of SLAM.A smooth path algorithm based on priority strategy is proposed after the path planning is studied. The efficiency of path planning is improved through the design of the evaluation equation of the A* algorithm. The Child-node-generation strategy based on priority value is adopted to avoid robot collision with the obstacles. Key-point optimization process is applied to reduce the number of turning point of generated path. Smoothing A* model is established after initial path processed and mobile robot can thus track the path smoothly and reach the goal fastly. Searching in the global grid costmap established by means of the improved algorithm can make it possible to to plan out a optimal path without collision quickly. And the dynamic window approach is adopted to local path planning, so the mobile robot can effectively avoid the dynamic obstacles in the process of navigation.Finally, a navigation system that based on laser senor and ROS has been designed and implemented on Pioneer3- DX robot platform. Completed experiments including map building, dynamic obstacle avoidance and path planning in the actual environment reveal that the laser sensor based mobile robot autonomous navigation scheme is feasible and reliable. The success rate of robot navigation in different environment are above 90%, and the position offset of robot are less than 4cm, the deflection of direction are less than 2°. |
