• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.5
• /
• pp.601-607
• /
• 2006

This paper proposes a novel balance control scheme of a biped robot to predict the next position of ZMP using Kalman Filter. The mathematical model of the biped robot is generally approximated by 3D-LIPM(3D-Linear Inverted Pendulum Mode), but it cannot completely express the robot's dynamics. The stability of the biped robot depends on whether the ZMP(Zero Moment Point) position is in the stability region or out of. And the internal error between the robot mechanism and its model could affect the stability of a robot. Therefore, the proposed balance control not reduces the internal error, but also timely generates the proper control. The experiment of the proposed balance control is simulated on the virtual workspace where the biped robot may encounter with various difficulties.

• The Journal of Korea Robotics Society
• /
• v.17 no.3
• /
• pp.264-272
• /
• 2022

Human-robot cooperative tasks are increasingly required in our daily life with the development of robotics and artificial intelligence technology. Interactive reinforcement learning strategies suggest that robots learn task by receiving feedback from an experienced human trainer during a training process. However, most of the previous studies on Interactive reinforcement learning have required an extra feedback input device such as a mouse or keyboard in addition to robot itself, and the scenario where a robot can interactively learn a task with human have been also limited to virtual environment. To solve these limitations, this paper studies training strategies of robot that learn table balancing tasks interactively using deep reinforcement learning with human's facial expression feedback. In the proposed system, the robot learns a cooperative table balancing task using Deep Q-Network (DQN), which is a deep reinforcement learning technique, with human facial emotion expression feedback. As a result of the experiment, the proposed system achieved a high optimal policy convergence rate of up to 83.3% in training and successful assumption rate of up to 91.6% in testing, showing improved performance compared to the model without human facial expression feedback.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2005.05a
• /
• pp.559-562
• /
• 2005

When we analyze dynamics of a multi body system, a compliance of joints must be considered. If the virtual model for CAE(computer adied engineering) analysis is not considered compliance, the result of CAE analysis will be very different from the actual result. Especially in a biped walking robot, a compliance can be caused in joints of a walking robot, and the robot may lose walking stability. This paper proposes a compliance modeling method and the effectiveness of the compliance model is verified through experiments.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1995.10a
• /
• pp.315-318
• /
• 1995

In this paper, an optimal trunk trajectory for stable walking of biped robots is expressed as a simple differential equation, which is then solved by numerical methods. We used ZMP (Zero Moment Point), the virtual total ground reaction point within the region of the supporting food, as the criterion of stability of biped robot walking. If the ZMP is located outside of the stable region in dynamic walking, biped robots fall down. The biped robot considered in this paper consists of two legs and a trunk. The trajectories of the two legs and the ZMP of the biped robot are determined such that they are similar ti those of a human. Based upon those trajectories, the trunk trajectory is solved by numerically integrating differential dynamic equations. Leg motions are controlled by the computed torque control method. The effectiveness of control algorithm as well as the trajectories is confirmed by computer simulations.

• Proceedings of the KIEE Conference
• /
• 2006.10c
• /
• pp.527-529
• /
• 2006

In this paper, we find a robot's path using a Virtual Reality Modeling Language and overlay vision. For correct robot's path we describe a method for localizing a mobile robot in its working environment using a vision system and VRML. The robt identifies landmarks in the environment, using image processing and neural network pattern matching techniques, and then its performs self-positioning with a vision system based on a well-known localization algorithm. After the self-positioning procedure, the 2-D scene of the vision is overlaid with the VRML scene. This paper describes how to realize the self-positioning, and shows the overlap between the 2-D and VRML scenes. The method successfully defines a robot's path.

• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.8
• /
• pp.845-850
• /
• 2009

Localization which estimates a robot's position and orientation in a given environment is very important for mobile robot navigation. Although low-cost sensors are preferred for practical service robots, they suffer from the inaccurate and insufficient range information. This paper proposes a novel approach to increasing the success rate of low-cost sensor-based localization. In this paper, both the previous and the current data obtained from the IR sensors are used for localization in order to utilize as much environment information as possible without increasing the number of sensors. The sensor model used in the monte carlo localization (MCL) is modified so that the accumulated range information may be used to increase the accuracy in estimating the current robot pose. The experimental results show that the proposed method can robustly estimate the robot's pose in indoor environments with several similar places.

• Proceedings of the KIEE Conference
• /
• 2001.11c
• /
• pp.111-114
• /
• 2001

In this paper, we supposed teller following algorithm that using sound localization for developing dialog type humanoid robot. A sound localization is studied for develop the techniques of an efficient 3-D sound system based on the psychoacoustics of spatial hearing with multimedia or virtual reality. When a robot talk with human, it is necessary that robot follow human for improved human interface and adaptive noise canceling. We apply this algorithm to robot system.

• The Journal of Korea Robotics Society
• /
• v.2 no.3
• /
• pp.255-261
• /
• 2007

We propose a optimal fusion method for localization of multiple robots utilizing correlation between GPS on each robot in common workspace. Each mobile robot in group collects position data from each odometer and GPS receiver and shares the position data with other robots. Then each robot utilizes position data of other robot for obtaining more precise estimation of own position. Because GPS data errors in common workspace have a close correlation, they contribute to improve localization accuracy of all robots in group. In this paper, we simulate proposed optimal fusion method of odometer and GPS through virtual robots and position data.

• Proceedings of the KIEE Conference
• /
• 1996.07b
• /
• pp.1208-1211
• /
• 1996

In this paper, an efficient and robust robot path planning technique is discussed. Concentric Ripple Edge Evaluation and Progression( CREEP ) algorithm[1] has been elaborated and expanded to carry out 3-D path planning. Like the 2-D case, robot can always find a path, if one exists, in a densely cluttered, unknown and unstructured 3-D obstacle environment. 3-D space in which the robot is expected to navigate is modeled by stacking cubic cells. The generated path is resolution optimal once the terrain is fully explored by the robot or all the information about the terrain is given. Path planning times are significantly reduced by local path update. Accuracy and efficiency of wave propagation in CREEP algorithm are achieved by virtual concentric sphere wave propagation. Simulations in 2-D and 3-D spaces are performed and excellent results are demonstrated.

• Proceedings of the KIEE Conference
• /
• 2004.05a
• /
• pp.83-85
• /
• 2004

Personal Robot System being developed is designed to module-based system architecture for supporting internal devices have variable system platforms, environments, networks. For supporting development environment, integrating remote services, managing the interoperation among internal modules on this system design, a well-designed object oriented middleware is needed significantly. There are already some middlwares like DCOM, CORBA, UPnP, JINI, OSGi, etc. But they have some limitations to applying to Personal Robot in variable side view. We are researching for a suitable design scheme to require low system resources, to guarantee realtime services as possible, and to implement easily. In this paper, we suggest a middleware architecture have goals that are simple, light, and object oriented, so that can be used at diversity devices in Personal Robot System.