Research On Key Technology Of Flatcar Unhooking Robot

At present,in the coal transfer workshop,the main use of the tipper is for the quick unloading of open freight rail cars carrying coal.However,due to space limitations,only one to four carriages can be unloaded at a time,so the connected open cars need to be unloaded prior to unloading in order for the freight open cars to enter the tipper.Currently there are many types of hooks used to connect railway open wagons,but China’s railway freight open wagons are mainly connected to the open wagons using collision-type hooks,and the wagons are unloaded by means of a hook removal operation.However,the hook removal operation is mostly carried out manually,with various safety hazards and a low level of automation,resulting in low production efficiency.In order to eliminate safety hazards and automate production,this paper proposes a solution that uses a hook removal robot to replace manual hook removal operations.The solution applies computer vision technology to recognise the type of hook,segment the hook handle and detect the distance information,and use an optimised control algorithm to complete the hook removal operation.Compared to manual operation,the robot hook picking operation is standardised and requires no human involvement,improving production safety;at the same time,the use of a robot instead of a human to perform the repetitive and tedious task of hook picking can improve production efficiency and enhance the automation of the hook picking process of the tiller.The main research elements of this paper are as follows:Firstly,this paper conducts a detailed study on the principle of unlocking the hook of an open car,and designs a scheme for an orbital hook picking robot system according to the robot workflow.The system mainly includes three parts: actuator,sensing mechanism and control system.Solid Works software is used to model the hook picking robot,and the three main functions of hook detection,hook handle recognition and distance detection,and robot hook picking control are analysed in relation to the hook picking process.Secondly,in order to address the problem that different types of open wagons lead to different ways of unlocking open wagon hooks,this paper uses a hook classification algorithm based on open wagon code recognition to identify open wagon codes on a self-built open wagon code dataset.The code recognition is used to determine the type of open car hook and then correlate it to determine the unlocking method of the hook.As the hook picking operation requires high accuracy for the hook handle segmentation and its position detection,this paper proposes the SE-Unet algorithm to segment the hook handle and to detect the position of the hook handle according to the principle of binocular vision detection in order to obtain the position information of the hook handle.Finally,in order to achieve high precision hook picking control of the open wagon hook picking robot,this paper uses the Harris Hawk optimized fuzzy control algorithm to realize the robot hook picking motion control.And to address the problems of the low optimization accuracy of the Harris Hawk optimization algorithm,a non-linear iterative escape energy update strategy is used to guide the hawk position update by using multi-individual information and adding perturbations to the system through the Corsi variation,so as to improve the optimization accuracy of the algorithm.