Research and application of rbf neural network in cone picking robot

In order to raise the efficiency of the cone picking robot and release the worker from heavy manual labor, a new control system of the RBF neural network is researched in this paper. The position and the object input voltage are taken as the input data of the RBF neural network model, and a combination learning algorithm is adopted to train the neural network. The sample data are gotten from a three-dimensional laser-scanner and some other sensors located on the cone picking robot. The test result shows that the new control system of the RBF neural network can automatically control the robot to pick cones accurately and quickly, and the efficiency of the robot is about 30-35 times than that of a worker who climbs up the tree to pick cones by hand with some special tools.