• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1992년도 한국자동제어학술회의논문집(국내학술편); KOEX, Seoul; 19-21 Oct. 1992
• /
• pp.436-440
• /
• 1992

As an approach to design the intelligent controller, this paper proposes a new FNN(Fuzzy Neural Network) control method using the hybrid combination of fuzzy logic control and neural network. The proposed FNN controller has two important capabilities, namely, adaptation and learning. These functions are performed by the following process. Firstly, identification of the parameters and estimation of the states for the unknown plant are achieved by the MNN(Model Neural Network) which is continuously trained on-line. And secondly, the learning is performed by FNN controller. The error back propagation algorithm is adopted as a learning technique. The effectiveness of the proposed method will be demonstrated by computer simulation of a two d.o.f. robot manipulator.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1990년도 한국자동제어학술회의논문집(국제학술편); KOEX, Seoul; 26-27 Oct. 1990
• /
• pp.878-883
• /
• 1990

There have been many approaches to solve the disturbance rejection problem in the control of LTI systems with state independent disturbances or possibly nonlinear state dependent disturbances. From the view point of each actuator, robot manipulators can be modeled as the second class of systems. With this model, M.Nakao et al. [1] introduced a decentralized control scheme based on interference estimation which is simple in its implementation and robust to the coupled dynamics and parameter variations. This paper systematically generalizes the control scheme to arbitrary finite dimensional LTI systems with disturbances. In doing so, we develop a disturbance observer theory for solving the disturbance rejection problem. We also present a discrete version of the theory with discussion of sampling and time-delay effects.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 2000년도 춘계학술대회논문집 - 한국공작기계학회
• /
• pp.171-175
• /
• 2000

Neural networks are in the framework of sensorbased tracking control of robot manipulators. They learn by practice movements the relationship between PSD (an analog Position Sensitive Detector) sensor readings for target positions and the joint commands to reach them. Using this configuration, the system can track or follow a moving or stationary object in real time. Furthermore, an efficient neural network architecture has been developed for real time learning. This network uses multiple sets of simple backpropagation networks one of which is selected according to which division (corresponding to a cluster of the self-organizing feature map) in data space the current input data belongs to. This lends itself to a very training and processing implementation required for real time control.

• Journal of Advanced Marine Engineering and Technology
• /
• 제26권5호
• /
• pp.581-588
• /
• 2002

A fuzzy learning control algorithm is proposed in this paper. In this method, two fuzzy controllers are used as a feedback and a feedforward type. The fuzzy feedback controller can be designed using simple knowledge for the controlled system. On the other hand, the fuzzy feedforward controller has a self-organizing mechanism and therefore, it does not need any knowledge in advance. The effectiveness of the proposed algorithm is demonstrated by experiment on the position and force control problem of a parallelogram type robot manipulator with two degrees of freedom. It is shown that the rapid learning and the robustness can be achieved by adopting the proposed method.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2003년도 추계학술대회
• /
• pp.1757-1762
• /
• 2003

For coping with various ship types easily, welding automation using CAD/CAM is demanded and and developed. In this paper we propose a dedicated CAM system which generates MACRO files to control robot manipulators. The paper contains CAD interface, virtual simulation, macro generation and job schedulling. At first, it defines and extracts weldline from CAD data, and generates proper MACRO programs. And it obtains optimum job schedules. This system removes the manual work, and consequently reduces the overall lead time. And it reduces costs and time for developing robot welding programs. This can be expanded to virtual factory simulation technology. Moreover, it is possible to apply this system for automation of Cutting and Painting

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1990년도 하계학술대회 논문집
• /
• pp.504-509
• /
• 1990

This paper presents a decentralized adaptive control scheme for multi-Joint robot manipulators based on the independent joint control scheme. The control object is to achieve accurate tracking of desired Joint trajectories. The proposed control scheme does not use the complex manipulator dynamic model, and each joint is controlled simple by a feedback controller which ensure stable and also a position-velocity-acceleration feedforward controller and also auxiliary signal, with adjustable gains. Simulation results are given for a two-link manipulator under independent control, proposed decentralized adaptive control of manipulator is feasible. In spite of a pay load variation and strong static and dynamic couplings that exist between the joints.

• International Journal of Precision Engineering and Manufacturing
• /
• 제6권2호
• /
• pp.61-71
• /
• 2005

The study reported deals with investigating the feasibility of control strategy for a serial rigid link manipulator that applies impulsive torques at the joints. The strategy is illustrated for a planar three rigid link manipulator. An impulse-based concept which uses successive torque impulses on rigid link as the controller for motion correction was introduced. This control strategy was tested over the entire trajectory to demonstrate that the tracking error could be reduced effectively. The best condition for minimizing the tracking error with the least impulse input at each joint is investigated by considering one design and one operating parameter. The first was the damping in the system, and the second was the sampling time during operation. The results show that this approach can provide useful guidance for the design and control of robot manipulators that require minimum impulse feedback for accurate tracking.

• 한국정밀공학회지
• /
• 제12권5호
• /
• pp.117-125
• /
• 1995

An improved robust control law is proposed for uncertain rigid robots. The uncertainty is nonlinear and (possibly fast) time-varying. Therefore, the uncertain factors such as imperfect modeling, friction, payload change, and external disturbances are all addressed. Based on the possible bound of the uncertainty, the controller is constructed. For uncertain flexible-joint robots, some feedback control terms are then added to the proposed robust control law in order to stabilize the elastic vibrations at the joints. To show that the proposed control laws are indeed applicable, the stability study based on Lyapunov function, a singular perturbation approach, and simulation results are presented.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 2000년도 추계학술대회논문집 - 한국공작기계학회
• /
• pp.108-112
• /
• 2000

Neural networks are used in the framework of sensorbased tracking control of robot manipulators. They learn by practice movements the relationship between PSD (an analog Position Sensitive Detector) sensor readings for target positions and the joint commands to reach them. Using this configuration, the system can track or follow a moving or stationary object in real time. Furthermore, an efficient neural network architecture has been developed for real time learning. This network uses multiple sets of simple backpropagation networks one of which is selected according to which division (corresponding to a cluster of the self-organizing feature map) in data space the current input data belongs to. This lends itself to a very training and processing implementation required for real time control.

• 한국산업융합학회 논문집
• /
• 제7권2호
• /
• pp.167-172
• /
• 2004

Neural networks are used in the framework of sensor based tracking control of robot manipulators. They learn by practice movements the relationship between PSD(an analog Position Sensitive Detector) sensor readings for target positions and the joint commands to reach them. Using this configuration, the system can track or follow a moving or stationary object in real time. Furthermore, an efficient neural network architecture has been developed for real time learning. This network uses multiple sets of simple back propagation networks one of which is selected according to which division (Corresponding to a cluster of the self-organizing feature map) in data space the current input data belongs to. This lends itself to a very training and processing implementation required for real time control.