Controllable Impedance Variable Flexible Joint Control And Safe Path Planning Of Intelligent Arm

Intelligent arm usually works within human-computer interaction environment,which regards safety as its primary requirement,therefore it should meet a certain degree of control accuracy and response speed.As the front end input commands are mostly discrete and relative fuzzy instructions rather than precise given path,the intelligent arm controller should have independent autonomous path planning ability,which reflects its feature of "intelligent".As a result,the goal of the intelligent arm is to carry out movement intention of users safely,accurately,quickly and intelligently under discrete commands such as "go left","go right" or "grab" in complicated daily environment.The intelligent arm used in this thesis is a kind of newly developed 6 dofs lightweight robotic arm.First of all,the mechanical design of intelligent arm with many aspects of innovation is introduced,a kind of cable-driven differential-actuated joint is designed,and with the use of small-volume high-torque motors combined with lightweight aluminum structure,the requests for weight,output torque and response speed are met.Furthermore,the mathematical model of intelligent am is established using improved D-H method.Under the condition of joint coupling actuation,the rigid body kinematics and dynamics models consisting of tranformation among actuation space-joint space-Cartesian space are constructed,also a kind of close-form inverse kinematics algorithm is proposed considering the special joint configuration.This part is the hardware foundation of controllable impedance variable flexible joint(CIVFJ).And the establishment of mathmatical models is theoretical basis of safe path planning and CIVFJ control.Based on the work of the first part,flexible joint control of the intelligent arm is studied.In the thesis,a kind of virtual CIVFJ is introduced along with its realization and dynamics model,several kinds of wildly used motion control methods are introduced and compared.Eventualy,a kind of gain self-tuning controller based on Sigmoid function is proposed to achieve ideal position accuracy and dynamic response characteristics in the process of joint flexibility and load variation.This part of the study ensures the accuracy and dynamic response characteristics of intelligent arm,while laying a good foundation for passive safe path planning based on CIVFJ.The last part is about safe path planning of the intelligent arm,which could be divided into three parts: active method,passive method and combination of both.Among which,active method is based on Rapidly-exploring random tree(RRT),a kind of improved RRT* algorithm is used to realize autonomous obstacle avoidance with visual supervision.Passive safe path planning method is used when visual supervision is unavailable or unreliable,low potential damage could be maintained by adjusting joints' flexibility.Finally,the combination of two methods in Cartesian space could enhance the safe operation ablity in complex and dynamic environment.This part of the study reflects the safety and intelligence of intelligent arm.Among all the chapters,the kinematics and dynamics models,close-form inverse kinematics algorithm,CIVFJ control method and safety path planning method are simulated based on Matlab and Robotic Toolbox.Further more,experiments based on the intelligent arm experimental platform are carried out in order to veirfy the rationality of constructed models as long as the effectiveness of proposed methods.