3-D Path Planning Using Neural Networks for a Robot Manipulator

Abstract

It is complex to find a good path when the robot is in a complex dynamic change condition, we are developing a trajectory prediction system is to reliably predict the moving obstacles. Algorithms for the device to perform path planning and trajectory prediction are described. We used neural network whose retraining is automatically triggered if major changes in the target behavior pattern are detected. The path planner uses super quadratic potential fields and incorporates a height change mechanism that is triggered where necessary and in order to get over short massif or pass by the taller obstacle.