Research On Key Technologies Of Intelligent Inspection Robot System Under Vehicle

Due to the large number and long mileage of domestic rail transit,the task of vehicle bottom inspection mainly relying on manual inspection is heavy,and there is an urgent need to replace manual inspection with robots for vehicle bottom inspection operations.The mobile manipulator has been widely used in various industrial on-site inspection tasks due to its ability of independent movement and flexible operation,among which the inspection task of rail vehicle bottom is one.At present,the vehicle bottom inspection robot has been gradually applied to the first level maintenance tasks of rail transit vehicles such as Multiple unit and subways,and can complete the functions of autonomous movement in the maintenance trench,fixed-point image acquisition,vehicle bottom fault image recognition,etc.However,the autonomous motion range of existing vehicle inspection robots is limited,and many can only move within a single trench,making it difficult to autonomously move from one maintenance trench to another;In addition,manual teaching methods are generally used for the collection of vehicle bottom images,which require the environment,vehicle condition,and teaching time to be identical,and have poor anti-interference ability.Based on the above issues,this article conducts research on the key technologies of the intelligent inspection robot system for the vehicle bottom,aiming to enhance the autonomy and intelligence of the vehicle bottom inspection robot,and lay the foundation for it to complete more complex vehicle bottom inspection tasks.The main research work of this article is as follows:(1)Analysis of vehicle bottom inspection issues and overall plan design.Based on the requirements of inspection tasks,this paper analyzes the two major problems of the current inspection robot at the bottom of the car,namely,it is unable to achieve cross trench inspection work,and the poor anti-interference ability of the image acquisition process at the bottom of the car.According to the inspection system of Multiple unit vehicles and the characteristics of the on-site environment of the inspection garage,this paper proposes a cross trench navigation scheme and an intelligent image acquisition scheme for the first level inspection task of the bottom of rail vehicles.(2)Design and Implementation of Cross Trench Navigation Scheme.A composite navigation scheme of laser SLAM navigation and magnetic stripe navigation is proposed based on the diversity of the maintenance warehouse environment.This article establishes a simulated maintenance library scenario in a laboratory environment,and conducts physical experiments on the proposed cross trench navigation scheme in this scenario.The experimental results show that the cross trench navigation scheme designed in this article can achieve robot operations across different trenches,and the navigation accuracy meets the requirements of vehicle bottom inspection tasks.The scheme also combines safety and stability.(3)Intelligent mapping and system testing.In this paper,the robot arm visual servo technology is introduced into the vehicle bottom image acquisition task,and an intelligent image acquisition scheme for the railway vehicle bottom is designed.This scheme enables the inspection robot to have a certain ability to correct the field of view,and can collect images of the vehicle’s bottom from an ideal perspective.Firstly,a scene of vehicle bottom bolt components was built in the laboratory,and a physical experiment was conducted on intelligent image acquisition of the vehicle bottom in this scene.The experimental results verified the effectiveness of the image acquisition scheme in this paper.Secondly,a bolt loosening fault image diagnosis method based on anti loosening line recognition is adopted to diagnose bolt loosening faults on the collected images.Finally,the hardware composition of the entire system and the completeness of system functions were introduced.