| Wall-climbing robots are automated mechanical devices that can move and perform tasks on vertical or inclined surfaces,playing a crucial role in high-altitude construction,inspection,maintenance,and repair.These robots need to achieve stable,efficient,and accurate positioning to move on various surfaces and complete high-altitude extreme tasks.However,the inherent errors of positioning sensors and the complexity of wall environments(such as occlusions,reflections,and multipath effects)impose certain limitations on the positioning of wall-climbing robots.In response to the needs of economic and social development,this paper,in collaboration with the Fujian Fuqing Nuclear Power Co.,Ltd.’s horizontal cooperation project "High-altitude Metal Wallclimbing Intelligent Nuclear Cleaning Robot R&D"(GY-H-21069),proposes a magnetic adsorption crawler wall-climbing robot using multiple sensors to improve positioning capabilities.The main research content and results of this paper include:(1)In response to the initial nuclear cleaning requirements for the interior walls of nuclear reactor buildings,this study,based on the design principles of crawler-type magnetic adsorption and ADAMS drive characteristic simulation,completes the walking and drive mechanism design of the wall-climbing robot and conducts safety analysis and simulation.Meanwhile,considering the possible slippage and skidding phenomena of the wall-climbing robot during its movement,a wheeled odometry model is established and combined with other sensor information for parameter optimization to improve the accuracy of the odometry.In addition,the hardware and software control system of the robot is developed,providing a foundation for subsequent research on multisensor fusion positioning.(2)This study analyzes the positioning sensors applied to wall-climbing robots and their operating wall environments,formulating a suitable sensor scheme.Due to the strong magnetic field generated by the magnetic tracks of the magnetic adsorption wall-climbing robot,which may affect the estimation of the magnetometer within its Inertial Measurement Unit(IMU),this paper proposes a correlation-based redundant magnetometer fusion method to mitigate external magnetic interference.To further address the common issue of low positioning accuracy or failure in wallclimbing robots,a multi-sensor fusion positioning solution based on IMU,Ultra-Wideband(UWB),and Li DAR is proposed.(3)In order to implement the proposed redundant magnetometer fusion method and multisensor fusion positioning scheme,a tilted redundant magnetometer platform was designed.Based on this platform,a triple 3-axis magnetometer fusion method was proposed,which is based on the dynamic weighting algorithm of normalized correlation coefficient analysis to filter the axes affected by external magnetic field interference,thereby reducing the external magnetic interference.Then,the particle filter algorithm and adaptive Monte Carlo(AMCL)positioning method based on particle filtering were studied.Finally,the mathematical models of various sensors were integrated into the multi-sensor fusion framework algorithm,thus designing a fusion positioning algorithm suitable for wall-climbing robots.On this basis,the position data provided by UWB was used to improve the problem that the AMCL algorithm could not accurately initialize particles.(4)The safety,stability,and positioning performance of the wall-climbing robot were experimentally tested in real-world scenarios.The experimental results demonstrated that the designed wall-climbing robot has good adhesion and turning capabilities.By introducing different magnetic field configurations,it was confirmed that the fusion algorithm based on an inclined redundant magnetometer platform has significant advantages in reducing the effect of the magnetic field on the estimation of the robot’s heading,achieving a root mean square error of 13.68° in heading angle estimation without using a gyroscope.Finally,through comparative experiments of different positioning methods,the results confirmed the accuracy and stability of the proposed fusion positioning algorithm,which significantly improved the positioning accuracy(within 7cm)compared to the single-sensor positioning method(over 10cm),providing strong support and guarantee for the practical application of the wall-climbing robot. |
PDF Full Text Request