Research On Some Algorithms Of Motion Planning For Autonomous Mobile Robot

Autonomous mobile robot is an important branch of robot. It involves artificial intelligence, intelligent control, information procession, image procession and detection and conversion. The development of autonomous mobile robot has imposing on the defense, society, economy and academy, and becomes the tactic research object of high technology of all countries. Motion planning for autonomous mobile robot is one of the most critical technologies in the autonomous navigation researches.The motion planning is to generate control input for the mobile robot to drive from an initial configuration to a goal configuration. Path planning and tracking control are two fundamental issues in the motion planning robotics.This dissertation is focused on the path planning and the trajectory tracking. Several improved methods and novel solutions are presented in order to improve computational efficiency, and additionally extend application domains. The main content of this dissertation include the following aspects:1. This dissertation introduced the second kind of partial differential equation in mobile robot path planning, has exploited the application of PDE method in mobile robot path planning problem, then new technology and method are provide for mobile robot path planning research. The algorithm overcame disadvantages of the other path planning algorithms, such as local minimization of classical potential field algorithm, unnecessary cost of RRT algorithm and so on. The mathematics principle of this algorithm and adaptive mesh has guaranteed this algorithm compatibility.2. To solve the limitations of Rapidly-Exploring Random Tree (RRT) algorithm, which only can be used in path planning when the environment is known, with combination of RRT algorithm and rolling path planning, this dissertation proposes an improved path planning. Only the local environmental map is calculated in the planning, thus the searching all environmental map is avoided, so that the RRT algorithm can be used in path planning not only when the environment is known but also unknown. By introducing the heuristic evaluation function, the exploring random tree can grow in the direction of target point. The regression analysis, which avoids local minima, enhances the capability of searching unknown space.3. In order to enhance the planning efficiency for mobile robot in unknown environment, this dissertation combined Bug algorithm and rolling path planning. Only the necessary sensing data instead of the analytical expression of the obstacle are calculated in planning so as to save memory, thus the planning in real-time is guaranteed. As one of the innate limitations of algorithm, the mobile robot tends to enter the trap situation due to local minima, so this dissertation solves this problem utilizing the concept of virtual obstacle. In order to maintain the planner's completeness, a global criterion is added in order to guarantee convergence to the goal.4. The cubic spiral for mobile robot path tracking is well studied here and some excellent characteristics are found. Then the cubic spiral curve is used to smooth the collision-free path by defining the cost functions of path for smoothness. Therefore the mobile robot is easy to track and application field of mobile robot is extended.5. A kind of tracked mobile robot model that is suitable for the design of controller is proposed through analyzing the force characteristics of the tracked mobile robot. Based on the dynamic model and kinematics constraints of tracked mobile robot, a trajectory tracking controller is designed. Parameters are introduced to describe the track-soil interactions. Considering the dynamics of mobile robot, the limited control strategy is introduced to guarantee the robot's smooth motion.6. The motion planning of nonholonomic wheeled mobile robot working in the unknown environment is researched. This dissertation combined the Bug algorithm and rolling path planning in the path planning for mobile robot. Based on the general dynamic model and non-slipping nonholonomic kinematics constraint of mobile robot, a trajectory tracking control system is designed by nonlinear feedback linearization. The angular velocities of the two back driving wheels are the input data so that the tracking errors are decreased. The simulation and experimental results verified the algorithm's effectiveness.A summary of the research conclusions and a discussion on the most promising paths of future research work are presented in the last chapter of this dissertation.