Design And Implementation Of Robotic Indoor Cruise System Based On ROS

In recent years,the application demands of indoor robots have continuously increased,and new requirements have been put forward for robots' ability to explore the unknown environment and their ability to locate and move.In addition,the diversity of robotic platforms and robotic applications requires high compatibility,portability,and scalability on robotic systems.Based on the requirement of indoor robots to explore unknown environments and autonomous navigation and patrols,an autonomous robotic cruise control system within the framework of robot operating system ROS is designed and implemented in this thesis.The system is mainly divided into global-profit POMDP active SLAM algorithm module and patrol control module.First,the historical background and development status of active SLAM algorithm and ROS framework are introduced,and the principle of SLAM algorithm based on particle filter and the existing active SLAM algorithms are elaborated.On this basis,an global-profit active SLAM algorithm is designed based on partially observable Markov decision process POMDP,with heuristic Monte Carlo search,a value function containing global exploration profit,as well as the model of confidence state update,observation model and state transfer function in the framework.And the global-profit active SLAM algorithm is implemented based on the ROS framework.Then a ROS-based indoor robot patrol and control system is designed and implemented.First,an improved dynamic window approach is proposed to solve the inefficiency of the original one.Then an algorithm for extracting the topological structure from the occupancy grid map is designed,and the robot autonomous patrol module based on the map topology is implemented.And this thesis also implements modules of the addition of name tags and the setting of custom patrol routes based on ROS visualization tool.Finally,the robot control extension interface is designed and implemented on the Android system.In the testing process,the robotic indoor autonomous cruise system designed in this thesis has reached the expected target.In the performance test of active SLAM algorithm,the global-profit POMDP algorithm designed in this thesis reduces the average coordinate error by 15.9%~21.9% and the average map key points error by 18.7%~25.7% compared with the frontier-based method.Compared with the mutual information POMDP method,it achieves the same level of accuracy with an average reduction by 33.8%~44.9% of the average single-step decision time and 19.4%~25.2% of the length of the exploration path in some cases.Furthermore,the global-profit POMDP algorithm has more robust performance in huge room with no obstacle.In the functional test of the patrol and control module,each sub-module can operate correctly and stably,achieving the design goals.