Dynamics Modeling And Intelligent Control For Humanoid Mechanical Arm

The dynamics modeling and intelligent control for a class of 6-DOF hybrid humanoid arm, whose layout is 3-2-1 configuration that the shoulder is 3-DOF spherical parallel mechanism, elbow is 2-DOF parallel mechanism and wrist is planar 4-bar mechanism, are studied.The nominal dynamics model based on orientation parameters of moving plant of shoulder is developed, the passive trajectory tracking controller is designed and validated by simulation; the dynamics model with parameter uncertainties is ulteriorly developed considering the errors brought by linearization of components, approximately computing in linear density etc., and adaptive trajectory tracking controller is designed considering parameter uncertainties and external disturbances. Considering the characteristic of repeated acting of shoulder, an adaptive ILC (Iterative learning controller) is designed.The nominal dynamics model of forearm is developed by Lagrange method and RBF (Radial basis function) NN (Neural network) Adaptive trajectory tracking controller is designed. For the reduction of calculation during uncertainty parameters learning and the realization of repeated trajectory tracking, the lumped parameter Adaptive ILC is designed and validated by simulation. The dynamics model of 6-DOF hybrid humanoid mechanism arm is developed. Considering the parameter uncertainties, frictions and external disturbances etc., a Fuzzy Adaptive trajectory tracking controller is designed and validated by simulation.Pseudo underactuated dynamics model and Pseudo underactuated intelligent control strategy are proposed, and the study of pseudo underactuated intelligent control for 6-DOF humanoid arm is done. The first step, the pseudo underactuated dynamics models of shoulder, humanoid forearm and 6-DOF humanoid mechanism arm is developed by Lagrange method and extending general variable. The second step, A NN Adaptive trajectory tracking controller is designed for the pseudo underactuated dynamics model of 6-DOF humanoid mechanism arm, and the feasibility of this controller is validated by simulation. The extended general variables is learned by NN algorithm, and the dynamics uncertainty induced by learning errors and the influences induced by frictions and disturbances are compensated by adaptive algorithm. The perfect tracking results are gained.The control system of 6-DOF hybrid humanoid mechanism arm is researched and manufacture. The trajectory tracking experiments of shoulder and forearm, which establishes the foundation of application, are done and the dates detected by sensors are gathered and plotted.