Research On Intelligent Navigation Of Plastering Robots For Indoor Buildings

In the past three decades,my country’s construction industry has developed rapidly.However,the traditional construction industry still has problems such as labor shortage,high-risk working environment,serious environmental pollution,uneven construction quality,and low worker efficiency.Building industrialization is the inevitable road for the development of China’s construction industry.In this thesis,according to the application scenarios and specific construction requirements of indoor building plastering robots in the construction industry,a set of intelligent navigation systems suitable for indoor building plastering robots will be designed and developed.Its main contents are as follows:（1）This thesis proposes a model suitable for indoor construction robots to recognize unknown scenes-"BIM + Realism" model.After repeated comparison and analysis of traditional point cloud map modeling methods,and the method of repeated experiments,the depth camera is finally selected to model the scene.Combined with the BIM diagrams existing in the construction industry,a "BIM + Realism" model is created.Finally,through the experimental test,it is calculated that the error rate of the method for the real scene size estimation is 1.72%,which verifies that the model is a more suitable scene model for robots in the construction industry.（2）This thesis proposes a multi-sensor fusion localization algorithm for indoor construction robots.Through the analysis and comparison of the existing positioning algorithms,combined with the application scenarios and positioning requirements of indoor robots,the final choice is to integrate visual sensing technology,neural networks and laser ranging technology.And the correct rate of the algorithm is calculated by selfmade data set in the construction scene,which verifies the effectiveness and feasibility of the multi-sensor fusion positioning algorithm for indoor construction robots.（3）This thesis proposes a path planning and navigation algorithm based on A^*algorithm and suitable for plastering robot.Since the original A^* algorithm did not meet the direction of the final particle according to the actual business requirements.For a plastering robot that can be used for wall plastering,it is very necessary to strictly meet the direction requirements specified by the actual business requirements while the path planning is completed.Therefore,in order to solve the direction problem of the robot,this thesis improves the algorithm based on the A^* algorithm.After the robot completes the navigation,it rotates itself according to the business requirements,so that the robot can be in the appropriate direction after the navigation is completed.After several experiments in the simulation environment,the correctness and feasibility of the improved algorithm are verified.（4）This thesis designs and develops the intelligent navigation system of indoor building plastering robot.Including robot positioning,communication and data interaction,Unity visualization interface,and robot navigation based on path planning.Through a large number of experiments and debugging in construction scenarios,it is verified that the intelligent navigation system can meet the construction requirements of plastering robots in the construction industry.