Visual Localization And End Effector Design Of Mobile Robot Unlocking Operation

With the rapid development of machine vision and robot control technology,more and more robots are applied to the extreme operating environment such as aerospace,deep sea and nuclear power plant.When the nuclear power plant by the external factors lead to nuclear leakage,the human living environment will be catastrophic impact,and disaster relief operations will also harm the health of the rescue workers,so the robot with mobile ability to become a nuclear power plant disaster relief operations in urgent need of equipment.Using a robot to open the lock of a fire door is a very important and difficult task in the relief operations of a nuclear power station.In this paper,through the analysis of the unlocking process and the visual positioning error,a kind of unlocking end effector with higher flexibility is developed,and some key technologies are studied deeply.Through in-depth analysis of the lock operation task requirements,the end effector functions to be defined,based on the relevant constraint analysis of the key insertion keyhole process,the kinematic model of the key and keyhole matching target is established,and then presents the parametric position calculation method to catch the capture area.The compliance parameters required by the end effector are obtained,which provide the theoretical basis for the design of the end effector.In order to realize the automatic unlocking task of visual guidance under the condition of nuclear power station,the method of recognizing the target image under the limited resolution of machine vision is studied,and the HOG-SVM method is used to identify the keyhole and door handle.And on the basis of that,a new method of identifying the physical characteristics of the target is proposed,which improves the accuracy of the target keyhole recognition.Study on dynamic tracking of the keyhole recognition through the target image.A new tracking method that a fusion of maximally stable extremal region features is proposed to meet the shortcomings of the traditional Meanshift tracking method,and the mean of tracking error is reduced.After the dynamic tracking,the target hole is intercepted and the location and orientation of the keyhole are determined on the basis of post-processing.Compared with the laser tracker position data,visual positioning error values are evaluated.Based on the softening parameters and the visual positioning error of the unlocking process,the overall design scheme of the under-actuated adaptive unlocking end effector with high compliant performance is proposed.The important components are optimized by simulation software.After the simulation analysis of the unlocking operation,the unlocking test platform is set up.On the basis of mechanical test’s evaluation,an automatic unlocking strategy is proposed.The results show that unlocking actuator can complete the task in the tolerance range.