Study on Binocular Picking Robot Path Optimization
|School||Northeast Agricultural University|
|Course||Agricultural Electrification and Automation|
|Keywords||Picking robot Camera calibration Genetic algorithm Path optimization Kinematics model Matlab|
With the development of modern agriculture, the agricultural robots have been widely applied to the agricultural production and harvest. But in the aspect of crop harvest, the targets of harvest are not entirely regular but high random and the working environment is complex. So the agricultural robots are needed higher requirements, such as, the ability of judgment and intelligence. The key of achieving picking fruits by robots is obtaining the accurate spatial location of fruits. Therefore, the method of recognition and positioning based binocular vision has been widely used. In this paper, a positioning system according to binocular technology research is designed. In addition, the path of robot should be optimized for improving the efficiency of picking. We optimize the working path of robot with the genetic algorithm and find the optimal path.At first, taking the mature green pepper as the research object, the corresponding data are derived by means of the camera calibration experiment based on the principles of robot visual location. And then the location information of green pepper is obtained with the binocular location model, so as to achieve the purposes of identification and location.Secondly, taking the shortest distance as the performance index, the various points of path which robot passed are planned with genetic algorithm. The experiment results indicate that genetic algorithm achieving the purposes of making the picking path shortest is not only has better effect than the tradition methods of polynomial interpolation and tri-spline but also has a small amount of calculation.Thirdly, taking 6-DOF fruit picking robot as the object research, the positive solution and inverse solution are obtained according to the theory of robot kinematics, and then, the end actuator can pick the target.Finally, the robotics toolbox running under the Matlab environment is used to simulate. The feasibility of systems is proved by means of the simulation. This paper achieves the purpose of the research and provides a reference for the future research..