A Study on Robot Motion Control and Regeneration for Small-Scale Industries

Abstract

The industrial manipulators have been increasingly applied in rehabilitation applications such as welding and painting because of the capability of providing impeccable repeatability and high welding quality on the desired trajectory. Generally, in the industrial fields, the motion controls of manipulators have been known as one of the most challenge and hardest problems. The robot manipulators are usually treated as a whole rigid body system, therefore, the set of dynamic equations governing the motion dynamics of the robot system is set up for the whole system. Such models are very complex, and thus implementing these model-based control schemes is not easy and requiring heavy time for computation. On the other hand, the manipulators are nonlinear multi-input multi-output systems having to face various uncertainties such as payload parameter, internal friction, and external disturbances. In addition, they generally work in gravitational fields where gravity force is the major cause of positioning error. Many advanced control algorithms have been applied in motion control for manipulators. Some of the main difficulties in robot motion control can be listed as follows: • Extremely nonlinear and strongly coupled in the set of dynamic equations of the robot system. • Mathematical modeling or mathematical parameters identification of the robot system. • During operation, the mathematical model parameters vary due to payload gravity, internal friction, and external disturbances, etc. • High-grade hardware configuration and complexity in software implementation. • The instant requirement of reprogramming or teaching due to the change of working conditions. To solve these problems, a straightforward method for robot motion regeneration used in painting and welding industries to achieve efficient and practical control is proposed in this study. In this work, instead of modeling the entire dynamics motion of the robot system, the proposed method is based on the decentralized modeling and motion control strategy. Exactly speaking, it is a user-oriented method and no need complexity in computation. The robot system is considered as a combination of individual joints; in this sense, the mathematical model of each joint of the robot system is derived individually. For this reason, these models are very straightforward. After that, three kinds of control algorithms are studied to regenerate the motion of each joint of the robot system. To completely perform these tasks, the followings are done: Firstly, a 6-DOF robot system description and mathematical modeling of each joint are presented. The mechanical and electrical components of the robot system are developed for convenient use in small-scale painting industry under the fact that easy reprogramming and no need any robotic specialists. Secondly, the nonsingular terminal sliding mode controller working together with time-delay estimation technique (TDE) is presented. The problem is that dynamic parameters of mathematical models of joints change when operating conditions change. In TDE technique, each joint actuator is considered as a free-inertia system without modeling nonlinear terms such as dynamic couplings, Coulomb friction, and gravitational payload. Based on this fact, uncertainties, nonlinearities, and disturbances in real systems are estimated effectively and instantly by using time-delayed states and control inputs. When the TDE is used in the control system, it does not require the detailed information of the mathematical model, leading to a very simple control method. The experimental results are shown to verify the practical feasibility and effectiveness of the proposed method. Thirdly, although experimental results of the control system based on TDE technique showed the practical feasibility for the robot control system, the designed control scheme took a long period of time to achieve asymptotic tracking performance due to this control method based on the assumption that unknown nonlinear functions in robot dynamics change slowly. Under the view of controlling dynamic systems with uncertain parameters, the model reference adaptive control method (MRAC) is considered as one of the most effective control methods and has been developed in the last few decades. Hence, the model reference adaptive control method should be deployed not only to cope with the parametric uncertainties but also to achieve more quickly asymptotic tracking performance. In this case, in order to improve tracking control performance and ensure the robustness of the control system, the control scheme of model reference adaptive control with uncertainty estimation is presented. Experimental results are presented for affirming the feasibility of applying the controller to the real robot system. Finally, the control methods based on TDE technique and the MRAC approach have shown the acceptable control performances. However, to obtain good tracking control performances under the condition that the motions change quickly like in painting industries, the nonlinear control method based on backstepping technique is proposed. Experimental results are shown to confirm the simplicity and feasibility with not only good control performances but also robustness under operating conditions without and with external disturbances.