• Journal of the Korea Computer Graphics Society
• /
• v.27 no.1
• /
• pp.9-17
• /
• 2021

In this paper, we present a simple and fast supervised learning framework based on model predictive control so as to learn motion controllers for a physic-based character to track given example motions. The proposed framework is composed of two components: training data generation and offline learning. Given an example motion, the former component stochastically controls the character motion with an optimal controller while repeatedly updating the controller for tracking the example motion through model predictive control over a time window from the current state of the character to a near future state. The repeated update of the optimal controller and the stochastic control make it possible to effectively explore various states that the character may have while mimicking the example motion and collect useful training data for supervised learning. Once all the training data is generated, the latter component normalizes the data to remove the disparity for magnitude and units inherent in the data and trains an artificial neural network with a simple architecture for a controller. The experimental results for walking and running motions demonstrate how effectively and fast the proposed framework produces physics-based motion controllers.

• Proceedings of the Korea Multimedia Society Conference
• /
• 1998.04a
• /
• pp.262-267
• /
• 1998

A hierarchical motion control system for animation of 3D human character is implemented using the motion database in realtime. The proposed motion control system consists of coordination controller for gait timing and balancing of walking motion, joint servo controller for realistic limb movement, and motion database for goal-directed character animation which makes time-consuming animation relatively easy task. As one example among the various applications of the proposed motion control system. We present a simple virtual reality system in which the motion control system plays a central role in generating realistic motion of virtual human character.

• Proceedings of the KIEE Conference
• /
• 1990.11a
• /
• pp.446-449
• /
• 1990

A Multi-Robot Control language (MRCL) is proposed to effectively control the multi-robot. MRCL has not only single-robot, command, but multi-robot command with multi-task OS, XINU. Concurrent motion, coordinate motion, and simple collision avoidance motion are implemented. This language is expected to act as a intelligent supporting tool for multi-robot system. To verify the effectiveness of the MRCL, a simple puzzle matching example is illustrated.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.117.3-117
• /
• 2001

This paper deals with a method for creating animation by interaction with animation operators. Operators are able to edit/transform any given motion data to more natural animations by the motion editing method proposed in this paper. The proposed technique is especially useful when some paris of character structure are changed. The system to be proposed is designed to fully utilize the experience of animation operators as well as to accomodate semi-automation process with spline interpolation. An example for retargeting a given motion data to a new character of dramatically changed kinematic structure.

• Proceedings of the IEEK Conference
• /
• 1999.06a
• /
• pp.403-406
• /
• 1999

Among the compression methods of moving picture information, a motion estimation method is used to remove time-repeating. The Block Matching Algorithm in motion estimation methods is the commonest one. In recent days, it is required the more advanced high quality in many image processing fields, for example HDTV, etc. Therefore, we have to accomplish not by means of Partial Search Algorithm, but by means of Full Search Algorithm in Block Matching Algorithm. In this paper, it is suggested a structure that reduce total calculation quantity and size, because the structure using Bit Plane select and use only 3bit of 8bit luminance signal.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 1989.10a
• /
• pp.46-50
• /
• 1989

Methods for the estimation of the coefficient matrices in the equation of motion for a linear multi-degree-of-freedom structure arc studied. For this purpose, the equation of motion is transformed into an auto-regressive and moving average with auxiliary input (ARMAX) model. The ARMAX parameters are evaluated using several methods of parameter estimation; such as toe least squares, the instrumental variable, the maximum likelihood and the limited Information maximum likelihood methods. Then the parameters of the equation of motion are recovered therefrom. Numerical example is given for a 3-story building model subjected to an earthquake exitation.

• 제어로봇시스템학회:학술대회논문집
• /
• 1992.10a
• /
• pp.736-741
• /
• 1992

The Stewart platform is one example of a motion simulator which generates 6 DOF motion in space by 6 actuators connected in parallel. The present SISO controllers are designed to track displacement command of each actuator computed from reference 6 DOF motion of platform by Stewart platform inverse kinematics. But this type of control can't cope with external load variation, geometric configuration of motion simulator, and different dynamic behavior of 6 DOF motion. In this paper, a multivariable controller using H- optimal control theory is designed for linerized simulator model with each actuator driving force as control input and platform 6 DOF motion as measured output. Nonlinear simulation result of the H$_{\infty}$ MIMO controller is not satisfied in steady-state characteristics. But the proposed H$_{\infty}$ + PI control scheme shows acceptable performance.e.e.

• Korean Journal of Computational Design and Engineering
• /
• v.16 no.4
• /
• pp.249-259
• /
• 2011

In this paper we present our work on the parameterized construction of virtual drivers' reach motion to seat belt, by using motion capture data. A user can generate a new reach motion by controlling a number of parameters. We approach the problem by using multiple sets of example reach motions and learning the relation between the labeling parameters and the motion data. The work is composed of three tasks. First, we construct a motion database using multiple sets of labeled motion clips obtained by using a motion capture device. This involves removing the redundancy of each motion clip by using PCA (Principal Component Analysis), and establishing temporal correspondence among different motion clips by automatic segmentation and piecewise time warping of each clip. Next, we compute motion blending functions by learning the relation between labeling parameters (age, hip base point (HBP), and height) and the motion parameters as represented by a set of PC coefficients. During runtime, on-line motion synthesis is accomplished by evaluating the motion blending function from the user-supplied control parameters.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1994.10a
• /
• pp.449-454
• /
• 1994

The Stewart Platform is one example of a motion simulator which generater 6DOF motion in space by six actuators in parallel. The presented control methrol of 6DOF motion simulator is generally classified into two types, one is SISO and the other is MIMO control type. The SISO control can't compensate for external load variation and different dynamic behavior of 6DOF motion, trerefore this type don's control motion precisely. On the other hand, the MIMO control compensates for a interference of 6DOF motion because MIMO controller is designed with 6DOF motion simulator synamics. But MIMO control of motion simulator has a complexity of 6DOF displacement feedback, because in oder to obtain feedback value we must solve the forward kinematics using measurement of cylinder length or design a state estimator, unless measurement of 6DOF displacement is possible. In this paper, a multivariable controller using H .inf. optimal control theory is designed to consider a interference of 6DOF motion and to obtain robust,precise control of system. Also in order to solve the mentioned problem of MIMO control, this paper presents a modified MIMO control model which control 6DOF motion by using feedback of measurement od cylinder length.

• ETRI Journal
• /
• v.37 no.4
• /
• pp.766-771
• /
• 2015

Robust motion estimation for human-computer interactions played an important role in a novel method of interaction with electronic devices. Existing pose estimation using a monocular camera employs either ego-motion or exo-motion, both of which are not sufficiently accurate for estimating fine motion due to the motion ambiguity of rotation and translation. This paper presents a hybrid vision-based pose estimation method for fine-motion estimation that is specifically capable of extracting human body motion accurately. The method uses an ego-camera attached to a point of interest and exo-cameras located in the immediate surroundings of the point of interest. The exo-cameras can easily track the exact position of the point of interest by triangulation. Once the position is given, the ego-camera can accurately obtain the point of interest's orientation. In this way, any ambiguity between rotation and translation is eliminated and the exact motion of a target point (that is, ego-camera) can then be obtained. The proposed method is expected to provide a practical solution for robustly estimating fine motion in a non-contact manner, such as in interactive games that are designed for special purposes (for example, remote rehabilitation care systems).