| With the popularization of automatic parking technology and the improvement of high-power charging technology,the automatic charging technology of electric vehicles has developed rapidly.However,the current automatic charging process mainly uses the visual system to locate the charging port of the EV with cooperative characteristics,which requires the transformation of the EV and make the overall cost high and the promotion difficult.In order to promote the further popularization of EV automatic charging technology,it is urgent to develop a positioning system with low cost,high precision,strong generalization and good environmental adaptability.Therefore,it is very important to study the monocular visual positioning technology without cooperative characteristics.The camera model was built and the camera distortion was corrected.That includes center perspective projection model,camera distortion model,parameters of industrial camera,formula derivation and the method of camera internal reference calibration and hand-eye calibration.The internal reference calibration and hand-eye calibration experiment of industrial camera are completed.A combination recognition method of high and low exposure based on texture feature is proposed.The algorithm can successfully identify the charging port area of EV in the environment of different light intensity and complex background.The principle of optical flow method is presented,and the charging port tracking experiment is carried out based on the pyramid LK optical flow method.An algorithm for locating the charging port of an EV without cooperative features based on monocular vision is designed.The kinematics modeling of AUBO-i5 robot and the derivation of forward and inverse kinematics equations are studied.By extracting and calculating the characteristic points of the charging port and the visual servo control of the robot,the positioning algorithm can accurately locate the charging port of the EV.An experimental platform of automatic charging system for electric vehicles was built.Under the experimental platform,the experiments of recognition accuracy,positioning accuracy and the insertion were completed.The experimental results show that the recognition success rate of the charging port is 97%.The repeated positioning accuracy of the system can reach ±0.25 mm and ±0.25°.The positioning accuracy of the system can reach ±1mm and ±1°.With proper control strategy,the system can plug in charging port smoothly and realize automatic connection and charging. |
