Research And Design On Robotic System Modeling Identification And Controller

With the development in the field of robotics in the world, the direction of robotics control and research received more attention. Robots are employed not only for simple task like assemble and welding. They are also used in more and more fields, for example, the rehabilitation robot in the medical field, robot receptionist in the services. Development of robot leads others forward. In recent decades, people witness the rise of humanoid robots. So people need the robots can provide with more functions which are different with simple action, but some advantage function which combined with many joints work together, such as walk, run, jump, interact with human, intel igent control, machine vision and virtual reality.This article focuses on the theoretical research of humanoid robot system with many degree-of- freedoms. In the first, the modeling work by D-H method for the 6 DOFs manipulator will be discussed. According to the relationship between the links which connect together, the recursive Newton-Euler formula will be utilized to analyze the modeling of manipulator, and then the inertia parameters will be identified. Considering the convergence property of the particle swarm optimization, the trajectory of the joint can be excited and the parameters of trajectory can be optimized. The excited trajectories will be well constra ined and smooth so that they are very suit to examine the control scheme. Then the inertia parameters of every joint will be identified by ridge regression algorithm.Secondly, the recursive adaptive method based control scheme will be designed to examine the effectiveness of identified parameters. The adaptive control scheme will be used to improve the tracking performance in the identification part. The completed manipulator model with inertia parameters will be proposed finally. Application of the model identification for robot system can reduce the complexity of computation. The parameters of robot system can be obtained with assistance of Simulink.As last, adaptive neural network based control scheme will be designed for the biped walking motion. In the part of robot balance control scheme, adaptive neural network will be utilized to deal with uncertain system. The state feed-back based control laws are designed, and the real time estimation is used to compensate the system uncertain parameters, such as inertia matrix, centrifugal and Coriols force term, and extern disturbance in motion.Simulations will be carried out to verify the result of the proposed modeling method and identify algorithm in final. The effectiveness of the control scheme can be il ustrated by comparing the tracking errors.