Research On Autonomous Polishing Mobile Robot For Building Environment

With the disappearance of the demographic dividend,China’s construction industry is facing enormous labor cost pressures,as well as high-risk,low production efficiency and other problems.The construction industry is one of the lowest digitized and least automated industries in the world.Among the existing modern technology systems,it is robotics that is most likely to take on the responsibility of the construction industry.Cleaning and polishing is an important process in the construction industry.With the large-scale construction of buildings,cleaning and grinding operations are facing more and more challenges,such as construction efficiency and personnel safety and health issues.Domestic research on cleaning and polishing robots for the built environment is still relatively rare.The related applications are limited to some mechanical transmission devices based on single sensors that need manual operation,which is difficult to adapt to complex and dangerous construction environments.To this end,this paper proposes a hardware architecture based on laser radar sensors and other sensors,a three-dimensional reconstruction map and a map-based full coverage path planning software architecture autonomous mobile robot system,and explore the laser SLAM,path planning and motion control algorithms.The main contents of this paper are as follows:(1)For the robot construction part,this paper proposes an inertial navigation aided laser radar 3D reconstruction system to provide high quality map information for the navigation part.The introduction of IMU information can not only predict the motion state,but also the motion distortion of the 3D laser point cloud,so that the pose of the observed 3D point is measured in the same coordinate system.The front-end NDT rough matching and GICP fine matching algorithm,combined with the loopback constraint and the virtual plane assumption as the back end of the constraint,greatly improve the positioning accuracy of the odometry.In order to further improve the accuracy of 3D reconstruction,this paper proposes a plane fitting of the ground points in the local submap by moving least squares method,and constrains the pitch,roll and shift in the Z axis direction by matching the current frame.Finally,the 3D reconstruction algorithm is validated on the actual scene and open source dataset,which proves the robustness and accuracy of the proposed algorithm and provides high-precision 3D map for autonomous navigation in full coverage navigation.(2)For the robot navigation part,this paper proposes a software framework for cleaning and polishing robots with multi-source information fusion.The proposed navigation system is equipped with lidar,sonar and infrared sensors,combined with the high-precision map proposed above,by reducing the dimensionality of the three-dimensional map to two-dimensional while retaining obstacle information at multiple levels,so that the robot locate itself while the surrounding obstacles are stereoscopically perceived,thereby achieving autonomous full coverage cleaning and polishing operations at the construction site.After many times of comparative experimental tests,the system can greatly reduce the safety hazards during the construction process while improving the efficiency of cleaning and polishing operations.(3)Summarize the full text work and draw a blueprint for future research.