Research On Wall Travelling Control For Ship Rust Removal Wall Climbing Robot

The special operation environment of wall climbing robot has attracted much attention in field of robot technology. And the rust removal wall climbing robot is an important developing direction in ship rust removing application because of its non-pollution and high rust removing efficiency. Rust removal wall climbing robot is suitable for heavy workload and automation, but the study on location, path planning and automatic control of climbing robot does not meet the requirements of application. The load of rust removal wall climbing robot is changed in work condition, which takes difficulty to control the robot’s movement. The barrier shape and distribution in different ship wall is different. The problem of self-localization, obstacles localization and path planning in various environments need to be solve. The key technology of the ship rust removal robot is studied in depth on location, path planning, control system and control algorithm and problem.In this dissertation, the location data fusion model is proposed to solve the robot location problem. The first level fusion is used to solve the robot’s self-localization problem, and the second level fusion is used to solve the obstacles localization problem. This fusion model can solve the position of the robot accurately, as well as obstacle problem. The obstacle detection sensor is used in to the biologically inspired neural network to solve the dynamic path planning problem. The mathematic model of obstacle detection sensor is used to build the local environmental information dynamically, this method can plan the path for robot movement in real time and it can solve the environment inaccuracy and uncertainty problem better. The template matching and searching of motion path planning method is proposed. It is first time to use the limited motion direction of the wall climbing robot in path planning, which can realizes the direction and angle judgment while the robot is moving to the next position on the base of the path planning. This method extends the path planning application of a mobile robot in the limited conditions.The thesis is outlined as following.In Chapter 1, the aim and significance of the studies in the dissertation are discussed. Then, the research on ship wall cleaning robot at home and abroad is investigated. Robot location and path planning are reviewed. At last, the main research subjects in this dissertation are presented.In Chapter 2, the static and dynamic model of rust removal wall climbing robot is built, and the permanent magnet absorption force and driven force were designed by simulation. The key components of rust removal wall climbing robot mechanical system were analyzed. Finally, the load capacity of the robot was tested through experiments. All of the above lay the foundation for the application study of the next chapter.In Chapter 3, this chapter studies the wall location problem of rust removal robot while it is in operation. Use the environment perception sensors to solve the position problem, and establish a mathematical model of the kinematic model of the robot and a mathematical model of various sensors so that the foundation for the fusion algorithm is established. Fusion model used two levels of information fusion, positioning algorithm used the extended Kalman filtering algorithm and weighted average to fuse the location of the robot, and the positioning accuracy of the algorithm was analyzed.Chapter 4 focuses on the point to point path planning that the rust removal robot moves from its finishing point of the rust removal area to another starting point of the rust removal area. To solve this problem, the biologically inspired neural network path planning method is used. On the base of the algorithm, the obstacle detection sensor is integrated into the algorithm to solve the dynamic path planning problem. Besides, the motion template is used to judge the motion of the robot direction and turning angle. Finally experiments prove that the improved algorithm is feasible and can solve the problem that the point to point path planning ignores the movements of the rust removal robot.Chapter 5 studies complete coverage path planning of area-covered rust removal. On the ground of the total planning of the complete coverage path planning, the integrated application of heuristic search algorithm and template method realizes the area covered and obstacles-avoided operation of the rust removal robot. The experiment shows that the covered efficiency is too low, so the dynamic inversion control method is introduced, which can optimize the path planning on analyzed theoretically. Finally, the experiment study on comprehensive path planning, result shows that the rust removal wall climbing robot autonomous localization and path planning function are achieved.Chapter 6 summarizes conclusions in this dissertation and offers some future research proposals.