An Anthropopathic Robot Arm-and-hand System Robust Nonlinear Control

Many areas, such as medical rehabilitation, aerospace, social services, and other fields require a variety of skillful, dexterous, and moving smooth anthropopathic robot arms-and-hands. In these specific areas, the anthropopathic robot arm-and-hand motion is not only to complete the specific actions and functions as the main target, but also in the movement process which can exercise and perform fast, smooth and precise actions according to the specific circumstances and needs, even as human multi-joint arm with a kind of skill and dexterity. Moreover, applications of the anthropopathic robot arm-and-hand is not only to replace or assist people to complete the duplication or special work, and want to make the anthropopathic robot arms-and-hands change their forms to accomplish desired tasks according to the surrounding environment flexibility, and also hope to reduce damage between the anthropopathic robot arms-and-hands and their own or the human, which requires some actuators can automatically transform the form of intelligent robot software instead of metal skeleton type robot complete these special tasks according to different requirements. Therefore, based on the multi-joint human arm movement mechanism, and combined with the mechanical characteristics of the robot arm-and-hand, construction based on special human simulated intelligent software driver of an anthropomorphic robot arm-and-hand,the realization of the structure and function of bionic, to create a centralized function, high efficiency, like the human multi-joint arm as smart and skilled anthropomorphic robot arm-and-hand is one of the effective ways. Based on the present situations and the urgent need to solve soft actuator modeling and advanced control challenges. This dissertation focus on researches on anthropomorphic robot arm-and-hand robust control, will discuss deeply some key technologies, consisting of the drive mechanism and properties of soft actuator and modeling,anthropomorphic controller design based on the human multi-joint arm movement mechanism, multiple-input multiple-output(MIMO) anthropomorphic robot arm-and-hand advanced control will be investigated. The main research contentsand results are as follows:1. Characteristic analysis and modeling of the smart soft actuator. The soft robot arms-and-hands can move freely in an unstructured environment with excellent security and flexibility. Although there are more advantages than the traditional rigid robot arms-and-hands, but the softness of the material makes the soft actuator movement mechanism become very complex, and brought to many difficulties in soft actuators or robot arms-and-hands modeling and controller design.In order to solve these practical problems, this dissertation discusses the basic structure, design principle, strong bending property and so on. The position control model based on the physical characteristics between the pressure and the bending angle is analyzed, and the force control model of the relationship between the pressure input and the force output is identified by the experimental data.2. An anthropopathic robot arm-and-hand robust nonlinear control system design. Based on the analysis of anthropopathic robot arms-and-hands research background, research significance and current development situation, a mixed anthropopathic robot arm-and-hand overall frame consisting of rigid and soft actuators is proposed, the anthropopathic robot arm-and-hand robust nonlinear control system design is investigated. First of all, the anthropopathic robot arm-and-hand nonlinear robust control system architecture is proposed in order to constitute an anthropopathic robot arm-and-hand consisting of two subsystems of major control and execution unit, namely the rigid robot arm and the soft actuator-based soft finger. Secondly, to control the end position of the anthropopathic robot arm, and the end position and the force of the soft fingers, two robust nonlinear control subsystems are designed based on an operator-based robust right coprime factorization approach. The operator controllers and an operator tracking controller are designed, and the robust stability and tracking performance of the anthropopathic robot arm are analyzed and evaluated, and the robust stability and tracking conditions are also discussed. Then, for the soft hand with three fingers,position and force control are investigated, and a robust control subsystem based on operator-based robust right coprime factorization techniques is designed, and the operator tracking controller is also designed. The robust stability and trackingperformance of the soft hand with three fingers are also analyzed and evaluated, and the robust stability and tracking conditions are discussed. Finally, the nonlinear robust stability and tracking performance of the whole anthropopathic robot arm-and-hand system are analyzed and discussed.3. An structure and function anthropopathic approach based on the structure and viscoelastic properties of multi joint human arm. According to how to use effectively human multi-joint arm viscoelastic properties in the design and control of the anthropopathic robot arm system, to explore the anthropopathic robot arm of the structure and function of anthropomorphic method, that is, through the analysis and research of the dynamic mechanical characteristics of moving human multi-joint arm viscoelasticity mechanism, and the anthropopathic controller based on human multi-joint arm viscoelastic properties is designed to control the end position of the anthropopathic robot arm.4. Compensation of uncertain plant based on the operator theory. Model uncertainty will bring about some challenging issues of tracking controller design.Aiming at these challenging issues, this dissertation presents a robust nonlinear control scheme using operator-based robust right coprime factorization approach to compensate the effect of model uncertainties, the operator observers based on the nominal model characteristics are designed to compensate the effect of model uncertainties. For the soft hand with three fingers with model uncertainties, the operator observers and the tracking controller are designed.