Design And Implementation Of Mobile Charging Car Based On Binocular Vision

In daily production and life,there are often many devices which use rechargeable batteries.Due to the limitation of the battery capacity,these devices require charging regularly.Unfortunately,the charging work is usually done manually,which increases the burden in people's life and work.In order to solve this problem,a mobile charging small car system is designed in this thesis.The system can automatically find the power outlet on the work site to charge its own battery and then provide the transfer charging service for the electrical consumption,which can replace human to complete the charging work for the equipment,bring convenience to people.Usually,the mobile charging car is parked in its own base station.When it is found that there is a device need to be charged,the car starts to search the target.The system calibrates the binocular camera,corrects the polar line and stereo matches the images of the left and right cameras.The car shoots the surroundings through a calibrated binocular camera and transmits the video to the host computer via Wi Fi.The host computer processes the image to detecte and track the obstacle,and according to this information,an optimal path is planned for the charging car.The calculation result is also sent back to the car controller.The controller controls the car to reach the designated position according to the calculation result.When the charging car approaches the target,the binocular camera can identify the mark on the outlet and find the location of it.Then the manipulator holds the plug to connect it to the outlet and charge the device.The main work of the thesis includes(1)The development of a control module for a mobile charging car was completed.Functions such as PWM speed control,car steering control,and manipulator control were realized.(2)Using the checkerboard calibration method provided by Zhang Zhengyou,Bouguet algorithm and BM algorithm to realize the calibration of the binocular camera,polar line correction and image matching.For images and videos captured by a calibrated binocular camera,Harris corner detection is used to extract feature points.It is also used to implement obstacle detection and tracking,with the combination of LK optical flow method.The distance is calculated based on the obtained results and the perspective is computed by the method which combines space vector with Hough transform.(3)Modeling a five-degree-of-freedom manipulator and using the D-H parameter to solve the forward and inverse kinematics equations of the manipulator to control the trajectory of the manipulator arm.The template matching method is used to identify the mark on the socket,and the space coordinates of the socket are calculated to guide the manipulator arm to complete the docking with the socket.This article also designed a magnetic suction plug which can improve the charging system's fault tolerance and reliability.(4)The optimization of the classic ant colony algorithm improves the speed of the algorithm,and solves the search stall problem of the traditional ant colony algorithm.On the basis of analyzing the kinematics of the mobile charging trolley,the calculation formula of the turning radius is deduced,the turning direction is judged by the angle information,and the obstacle avoidance function of the car is realized.