• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.7
• /
• pp.893-898
• /
• 2005

Due to the merits that only a small amount of computation is needed for solutions and stochastic policies can be handled explicitly, the actor-critic algorithm, which is a class of reinforcement learning methods, has recently attracted a lot of interests in the area of artificial intelligence. The actor-critic network composes of tile actor network for selecting control inputs and the critic network for estimating value functions, and in its training stage, the actor and critic networks take the strategy, of changing their parameters adaptively in order to select excellent control inputs and yield accurate approximation for value functions as fast as possible. In this paper, we consider a new actor-critic algorithm employing an RLS(Recursive Least Square) method for critic learning, and policy gradients for actor learning. The applicability of the considered algorithm is illustrated with experiments on the two linked robot arm.

• Journal of the Korean Society for Railway
• /
• v.11 no.4
• /
• pp.364-370
• /
• 2008

In this paper, we proposed new sensor network architecture with autonomous robots based on beacon mode and implemented real time monitoring system in real test-bed environment. The proposed scheme offers beacon based real-time scheduling for reliable association process with parent nodes and dynamically assigns network address by using NAA (Next Address Assignment) mechanism. For the large scale multi-sensor processing, our real-time monitoring system accomplished the intelligent database processing, which can generate not only the alert messages to the civilians but also process various sensing data such as fire, air, temperature and etc. Moreover, we also developed mobile robot which can support network mobility. Though the performance evaluation by using real test-bed system, we illustrate that our proposed system demonstrates promising performance for emergence monitoring systems.

• Journal of KIISE:Computing Practices and Letters
• /
• v.13 no.2
• /
• pp.100-109
• /
• 2007

Object recognition of service robots is very important for most of services such as delivery, and errand. Conventional methods are based on the geometric models in static industrial environments, but they have limitations in indoor environments where the condition is changable and the movement of service robots occur because the interesting object can be occluded or small in the image according to their location. For solving these uncertain situations, in this paper, we propose the method that exploits observed objects as context information for predicting interesting one. For this, we propose the method for modeling domain knowledge in probabilistic frame by adopting Bayesian networks and ontology together, and creating knowledge model dynamically to extend reasoning models. We verify the performance of our method through the experiments and show the merit of inductive reasoning in the probabilistic model

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.640-645
• /
• 2004

This paper proposes a direct adaptive control method using self-recurrent wavelet neural network (SRWNN) for stable path tracking of mobile robots. The architecture of the SRWNN is a modified model of the wavelet neural network (WNN). Unlike the WNN, since a mother wavelet layer of the SRWNN is composed of self-feedback neurons, the SRWNN has the ability to store the past information of the wavelet. For this ability of the SRWNN, the SRWNN is used as a controller with simpler structure than the WNN in our on-line control process. The gradient-descent method with adaptive learning rates (ALR) is applied to train the parameters of the SRWNN. The ALR are derived from discrete Lyapunov stability theorem, which are used to guarantee the stable path tracking of mobile robots. Finally, through computer simulations, we demonstrate the effectiveness and stability of the proposed controller.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.10 no.2
• /
• pp.101-106
• /
• 2010

We propose an approach to estimate the real-time moving trajectory of an object in this paper. The object's position is obtained from the image data of a CCD camera, while a state estimator predicts the linear and angular velocities of the moving object. To overcome the uncertainties and noises residing in the input data, a Extended Kalman Filter(EKF) and neural networks are utilized cooperatively. Since the EKF needs to approximate a nonlinear system into a linear model in order to estimate the states, there still exist errors as well as uncertainties. To resolve this problem, in this approach the Kohonen networks, which have a high adaptability to the memory of the inputoutput relationship, are utilized for the nonlinear region. In addition to this, the Kohonen network, as a sort of neural network, can effectively adapt to the dynamic variations and become robust against noises. This approach is derived from the observation that the Kohonen network is a type of self-organized map and is spatially oriented, which makes it suitable for determining the trajectories of moving objects. The superiority of the proposed algorithm compared with the EKF is demonstrated through real experiments.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.12 no.7
• /
• pp.2913-2918
• /
• 2011

This paper proposed the friction model using Preisach algorithm with neural network based on experimental results. In order to apply the neural network algorithm, the back propagation update rule was used and the updated weighting factor of neural network was applied to distribute function of Preisach model. In order to implement the proposed algorithm, the LabView software was used to apply to the precision control of mechanical system. The evaluation of the proposed friction model was executed through experiments.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.18 no.1
• /
• pp.1-11
• /
• 1994

This paper presents an intelligent robotic control method for chamferless parts mating by integrating fuzzy control and neural network. The successful assembly task requires an extremely high position accuracy and a good knowledge of mating parts. However, conventional assembly method alone makes it difficult to achieve satisfactory assembly performance because of the complexity and the uncertainties of the process and its environments such as not only the limitation of the devices performing the assembly but also imperfect knowledge of the parts being assembled. To cope with these problems, an intelligent robotic assembly method is proposed, which is composed of fuzzy controller and learning mechanism based upon neural net. In this method, fuzzy controller copes with the complexity and the uncertainties of the assembly process, while neural network enhances the assembly scheme so as to learn fuzzy rules from experience and adapt to changes in environment of uncertainty and imprecision. The performance of the proposed assembly scheme is evaluted through a series of experiments using SCARA robot. The results show that the proposed control method can be effectively applied to chamferless precision parts mating.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.676-678
• /
• 2003

Microtunneling techniques play a crucial role in the construction of pipelines. This paper shows the automatic tunneling algorithm of microtunneling system using fuzzy logic technology to assist operators to assure the quality of microtunneling construction. To have effective output value of fuzzy controller, we slightly modified the conventional defuzzification methods. The proposed automatic tunneling algorithm shows good tunneling results comparable with those of experts.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.25 no.4
• /
• pp.97-109
• /
• 2000

This work proposes a neural network approach to solve vehicle routing problems which have diverse application areas such as vehicle routing and robot programming. In solving these problems, classical mathematical approaches have many difficulties. In particular, it is almost impossible to implement a real-time vehicle routing with multiple vehicles. Recently, many researchers proposed methods to overcome the limitation by adopting heuristic algorithms, genetic algorithms, neural network techniques and others. The most basic model for path planning is the Travelling Salesman Problem(TSP) for a minimum distance path. We extend this for a problem with dynamic upcoming of new positions with multiple vehicles. In this paper, we propose an algorithm based on SOM(Self-Organization Map) to obtain a sub-optimal solution for a real-time vehicle routing problem. We develope a model of a generalized multiple TSP and suggest and efficient solving procedure.

• 제어로봇시스템학회:학술대회논문집
• /
• 1993.10a
• /
• pp.150-155
• /
• 1993

The neural network MRAC system is presented. The purpose of this paper is applied to a plant that is to be controlled in a strongly nonlinear environment. The proposed system has a learning and adaptive ability in the varying environment by using the back-propagation learning algorithm based on Lyapunov stability theory. N.N. regulator is a part of overall system and is guaranteed to be stable in initial stage. Nonlinear terms of the varying mass, colilori, centifugal, and gravity are compensated for by feedforward N.N. regulator. And the feedback controller (adaptive mechanism) works to eliminate errors of position, velocity which the feedforward controller cannot compensate for. Finally, the proposed system will be demonstrated by simulation of a two d.o.f robot manipulator.