• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.2
• /
• pp.28-37
• /
• 1999

Differentially Responsible Adaptive Critic Learning technique is proposed for learning the control technique with multiple control inputs as in robot system using reinforcement learning. The reinforcement learning is a self-learning technique which learns the control skill based on the critic information Learning is a after a long series of control actions. The Adaptive Critic Learning (ACL) is the representative reinforcement learning structure. The ACL maximizes the learning performance using the two learning modules called the action and the critic modules which exploit the external critic value obtained seldomly. Drawback of the ACL is the fact that application of the ACL is limited to the single input system. In the proposed Differentially Responsible Action Dependant Adaptive Critic learning structure, the critic function is constructed as a function of control input elements. The responsibility of the individual control action element is computed based on the partial derivative of the critic function in terms of each control action element. The proposed learning structure has been constructed with the CMAC neural networks and some simulations have been done upon the two dimensional Cart-Role system and robot squatting problem. The simulation results are included.

• Proceedings of the KIEE Conference
• /
• 1999.11c
• /
• pp.812-814
• /
• 1999

In this paper, we propose a new control algorithm based on the neural network(NN) feedback compensation with a desired trajectory modification. The proposed algorithm decreases trajectory errors by a feed-forward desired torque combined with a neural network feedback torque component. And, to robustly control the tracking error, we modified the desired trajectory by variable structure concept smoothed by a fuzzy logic. For the numerical simulation, a 2-link robot manipulator model was assumed. To simulate the disturbance due to the modelling uncertainty. As a result of this simulation, the proposed method shows better trajectory tracking performance compared with the CTM and decreases the chattering in control inputs.

• Journal of Institute of Control, Robotics and Systems
• /
• v.17 no.2
• /
• pp.131-134
• /
• 2011

In this study, we propose an optimal design and new fabrication method for a slim tactile sensor. Slim tactile sensor can detect 3-axial forces and has suitable flexibility for intelligent robot fingers. To amplify the contact signal, a unique table-shaped structure was attempted. A new layer-by-layer fabrication process for polymer micromachining that can make a 3D structure by using a sacrificial layer was proposed. A table-shaped epoxy sensing plate with four legs was built on top of a flexible polymer substrate. The plate can convert an applied force to a concentrated stress. Normal and shear forces can be detected by combining responses from metal strain gauges embedded in the polymer substrate. The optimal positions of the strain gauges are determined using the strain distribution obtained from finite element analysis.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.18 no.12
• /
• pp.111-116
• /
• 2019

With the recent acceleration of industrial technologies and active research, wearable robot technologies have been applied to various fields. To study the utility of wearable robots, basic research on kinetic mechanisms of the human body, bio-signal analysis, and system control are essential. In this study, we investigated the basic structure of a wearable system and the operating principles of a driving mechanism. The control system and supporting structure, which comprise the driving mechanism, were designed and manufactured. Motion and load analyses were performed simultaneously for the design of the kinematic drive, and the driving mechanism was constructed by analyzing walking motion. The operating conditions of the cylinder were verified by stride via driving experiments. Further, the accuracy and responsiveness of the system were confirmed by comparison with actual motion, and the system safety was validated by applying loads.

• Journal of Institute of Control, Robotics and Systems
• /
• v.8 no.11
• /
• pp.935-940
• /
• 2002

For robot navigation and visual reconstruction, structure from motion (SFM) is an active issue in computer vision community and its properties arc also becoming well understood. In this paper, when using stereo image sequence and direct method as a tool for SFM, we present a new method for overcoming bas-relief ambiguity. We first show that the direct methods, based on optical flow constraint equation, are also intrinsically exposed to such ambiguity although they introduce robust methods. Therefore, regarding the motion and depth estimation by the robust and direct method as approximated ones. we suggest a method that refines both stereo displacement and motion displacement with sub-pixel accuracy, which is the central process f3r improving its ambiguity. Experiments with real image sequences have been executed and we show that the proposed algorithm has improved the estimation accuracy.

• Journal of Institute of Control, Robotics and Systems
• /
• v.3 no.6
• /
• pp.569-575
• /
• 1997

모델 불확실성이 있고 n축 자유도(degree of freedom)를 갖고 있는 로봇 동작부(manipulator)에 대해서 위치 정보만을 이용하여 가변 구조 제어기(variable structure controller)를 설계하였다. 모델의 불확실성이 존재하는 경우에도 제어기에 사용되는 속도를 잘예측하기 위해 고이득 관찰기를 사용 하였으며 고이득 관찰기를 사용할때 발생할 수 있는 상태변수의 피킹현상(peak phenomenon)를 적게 하게 하기위하여 제어기의 값을 제한 (globally bounded)하여 제어기를 설계하였다. 부하(payload)의 범위만 알고 있는 2축 자유도를 갖는 로봇 동작부에 대해서 제안된 제어 방법에 따라 제어기를 설계하여 그 성능를 확인 하였다.

• The Journal of Korea Robotics Society
• /
• v.16 no.2
• /
• pp.122-129
• /
• 2021

For the Simultaneous Localization and Mapping (SLAM) problem, a different path results in different SLAM results. Usually, SLAM follows a trail of input data. Active SLAM, which determines where to sense for the next step, can suggest a better path for a better SLAM result during the data acquisition step. In this paper, we will use reinforcement learning to find where to perceive. By assigning entire target area coverage to a goal and uncertainty as a negative reward, the reinforcement learning network finds an optimal path to minimize trajectory uncertainty and maximize map coverage. However, most active SLAM researches are performed in indoor or aerial environments where robots can move in every direction. In the urban environment, vehicles only can move following road structure and traffic rules. Graph structure can efficiently express road environment, considering crossroads and streets as nodes and edges, respectively. In this paper, we propose a novel method to find optimal SLAM path using graph structure and reinforcement learning technique.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.495-495
• /
• 2000

The Stewart platform manipulator is a closed-kinematis chain robot manipulator that is capable of providing high st겨ctural rigidity and positional accuracy. However, this is a complex structure, so controllability of the system is not so good. In this paper, it introduces a new robust motion control algorithm using partial state feedback for a class of nonlinear systems in the presence of modelling uncertainties and external disturbances. The major contribution of this work introduces the development and design of robust observer for the slate and the perturbation w.hich is integrated into a variable structure controller(VSC) structure. The combination of controller/observer gives rise to the robust routine called sliding mode control with sliding perturbation observer(SMCSPO). The optimal gains of SMCSPO are easily obtained by genetic algorithm. Simulation and experiment are presented in order to apply to the stewart platform manipulator. There results show highly' accuracy and performance.

• Journal of IKEEE
• /
• v.8 no.1 s.14
• /
• pp.96-107
• /
• 2004

In this paper, a new integral variable structure regulation controller(IVSRC) is designed by using a special integral sliding surface and a disturbance observer for the improved regulation control of highly nonlinear robot manipulators with prescribed output performance. The sliding surface having the integral state with a special initial condition is employed in this paper to exactly predetermine the ideal sliding trajectory from a given initial condition to origin without any reaching phase. And a continuous sliding mode input using the disturbance observer is also introduced in oder to effectively follow the predetermined sliding trajectory within the prescribed accuracy without large computation burden. The performance of the prescribed tracking accuracy to the predetermined sliding trajectory is clearly investigated in detail through the two theorems together with the closed loop stability. The design of the proposed IVSRC is separated into the performance design and robustness design in each independent link. The usefulness of the algorithm has been demonstrated through simulation studies on the regulation control of a two link manipulator under parameter uncertainties and payload variations, in view of no reaching phase, no overshoot, predetermined response with prescribed accuracy, easy change of output performance, separation of design phase, and so on.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1997.10a
• /
• pp.111-114
• /
• 1997

Usually, Evolutionary Algorithms are considered more efficient for optimal system design, However, the performance of the system is determined by fitness function and system environment. In this paper, in order to overcome the limitation of the performance by this factor, we propose a co-evolutionary method that two populations constantly interact and coevolve. In this paper, we apply coevolution to neural network's evolving. So, one population is composed of the structure of neural networks and other population is composed of training patterns. The structure of neural networks evolve to optimal structure and, at the same time, training patterns coevolve to feature patterns. This method prevent the system from the limitation of the performance by random design of neural network structure and inadequate selection of training patterns. In this time neural networks are trained by evolution strategies that are able to apply to the unsupervised learning. And in the coding of neural networks, we propose the method to maintain nonredundancy and character preservingness that are essential factor of genetic coding. We show the validity and the effectiveness of the proposed scheme by applying it to the visual servoing of RV-M2 robot manipulators.