A Hierarchical Neuro-Fuzzy System to Minimum-Time Trajectory Planning of Redundant Manipulators

In this paper, the problem of minimum-time trajectory planning is studied for a three degrees-offreedom (3-DOF) planar manipulator using a hierarchical hybrid neuro-fuzzy system. A first preprocessing step involves two components. The first component is NeFIK (for Neuro-Fuzzy Inverse Kinematics), a neuro-fuzzy network designed to learn and solve the inverse kinematics problem. The second one is an optimal time control problem involving robot dynamics and constraints. This latter is solved using Augmented Lagrangian (AL) technique. These pre-processing steps allow generating a large dataset of minimum-time singularity-free trajectories, covering mostly of the robot workspace. Upon this dataset, an objective neuro-fuzzy inference system is built to learn and capture the robot minimum time dynamic behavior. Once this system is trained and optimized, it is used in a generalization phase to achieve online planning. Simulation results showing the effectiveness of the proposal are presented and discussed.