| With the continuous improvement of the requirements of airport security and the development of the technology of robot,inspection robots are used instead of manual inspection to improve the patrol efficiency and forecast the warning situation of airport,which has gradually become the trend of airport development.According to the inspection requirements of the airport,a control system of airport inspection robot is designed and developed,which mainly completes the following tasks:Firstly,the functional requirements of the control system of airport inspection robot are analyzed in depth.Based on the existing robot platform and requirements analysis for the control system of airport inspection robot,the hardware and software architecture of the control system is designed,and the selection of core components was completed.Secondly,aiming at the problem of the coexistence of multiple complex features that such as open areas,tall buildings,and people and cars in the environment of the airport,a robot positioning method combining binocular vision,IMU and RTK-GNSS is proposed to overcome the defect that single method of positioning would be failed in this environment.The data of binocular vision and IMU is tight coupled in a variable-length sliding window to obtain continuous pose data which is used as the initial pose of the robot because RTK-GNSS cannot output positioning data when be obscured heavily.The data of RTK-GNSS and the data of initial pose of the robot is fused by the method of adaptive Kalman filtering to eliminate the cumulative error in the initial pose and obtain an accurate pose of the robot.Thirdly,aiming at the problem that the speed of the robot is impacted largely when the flow of people inrush to the functional area after the airplane enters the port,the topological map of the functional area is established with the time of inspection as the weight,and the weight in the topological map is updated according to the time when the robot starts to inspect and the passengers stay in the functional area.An inspection mode that takes the separate area in the functional area as the inspection unit is designed,and the area of inspection is simplified into regional inspection points and auxiliary inspection points.The algorithm of Dijkstra is used to calculate the inspection path with the shortest time between pairs of inspection points,and then an improved algorithm of ant colony is used to plan a globally optimal path between inspection points.A rule to evaluate whether the area of spection is completely covered by the planned path is added to the algorithm of ant colony to solve the problem that the global path may not completely cover the area needs to be inspected.Fourthly,in order to reduce the running time of the robot at the right-angle turns in the planned path,the robot control method of front-wheel steering and rear-wheel driving is used.In order to improve the traceability of the right-angle turns in the global path,a cubic B-spline is used to smooth the planned path.The kinematics model and error model of the airport inspection robot are established,and the tracking control of trajectory is performed by the method of sliding mode control.A switching function of sliding mode is designed,and the single-power reaching law is used to reduce the chattering.Based on this,the control law of trajectory tracking is calculated to control the motion of robot,which makes the robot has a higher accuracy of the trajectory tracking.Fifthly,based on the robot operating system(ROS),the programming of the Ontological software of robot and the background software of robot is completed.And the reliability of the inspection robot's control system is verified by the actual test in Jining Qufu airport. |
PDF Full Text Request