Jung-Hoon Hwang

Mobile robots at your fingertip: Bezier curve on-line trajectory generation for supervisory control

A new interfacing method is presented to control mobile robot(s) in a supervised manner. Mobile robots often provide global position information to an operator. This research describes a method whereby the operator controls a mobile robot(s) using his finger or stylus via a touchpad or touch screen interface. Using a mapping between the robots operational site and the input device, a human user can provide routing information for the mobile robot. Two algorithms have been developed to create the robot trajectory from the operators input. Information regarding numerous path points is generated when the operator moves his finger/stylus. To prune away meaningless point information, a simple but powerful significant points extracting algorithm is developed. The resulting significant points are used as waypoints. An on-line piecewise cubic Bezier curves (PCBC) trajectory generation algorithm is presented to create a smooth trajectory for these significant points. As the method is based on distance and not on time, the velocity of mobile robot can be controlled easily within its allowable dynamic range. The PCBC trajectory can also be modified on the fly. Simulation results are presented to verify these newly developed methods.

A novel interactive robot soccer system

A new robot soccer system is proposed to increase the interaction between the operator and the robot soccer system. The multiple robots are commanded by a teleoperation function. When a robot is commanded with a joystick, other robots are running by the pre-programmed cooperating strategies. To develop the interactive robot soccer system, robust position/orientation sensing techniques have been developed. In order to make the teleoperation of the soccer robot more intuitive it is proposed to use a joystick control with respect to ground-based coordinates. To allow wireless and isolated teleoperation for each playground an infrared communication system has been adopted. To minimize the maintenance a continuous power supply system from the ground has also been developed.

Efficient Object Tracking System Using the Fusion of a CCD Camera and an Infrared Camera

To make a robust object tracking and identifying system for an intelligent robot and/or home system, heterogeneous sensor fusion between visible ray system and infrared ray system is proposed. The proposed system separates the object by combining the ROI (Region of Interest) estimated from two different images based on a heterogeneous sensor that consolidates the ordinary CCD camera and the IR (Infrared) camera. Human`s body and face are detected in both images by using different algorithms, such as histogram, optical-flow, skin-color model and Haar model. Also the pose of human body is estimated from the result of body detection in IR image by using PCA algorithm along with AdaBoost algorithm. Then, the results from each detection algorithm are fused to extract the best detection result. To verify the heterogeneous sensor fusion system, few experiments were done in various environments. From the experimental results, the system seems to have good tracking and identification performance regardless of the environmental changes. The application area of the proposed system is not limited to robot or home system but the surveillance system and military system.

Physical contact of devices: utilization of beats for interpersonal communication

In this paper, an interpersonal communication method based on tactile beats when devices are established physical contact will be proposed. Tactile beats are a well-known phenomenon that describes frequency modulation occurs when vibrating objects with similar but not same operating frequencies are physically connected. Vibrating signals at each configuration (no contact and contact with same/different frequency) were measured using a laser vibrometer. Preliminary user study revealed that the induced physical tactile stimulus was perceived well by the subjects. As an application, social touch interaction using hand-held devices with differently assigned operating vibration frequencies was described. Mapping the frequency deviation between devices to quantifiable social information, it is expected that the proposed algorithm can be applied for interpersonal communication based on physical contact with enhanced emotional and social experiences.

Increasing the impedance range of admittance-type haptic interfaces by using Time Domain Passivity Approach

This paper proposes a method to increase the impedance range of admittance-type haptic interfaces. Admittance-type haptic interfaces are used in various applications that typically require interaction with high impedance virtual environments. However, the performance of admittance haptic interfaces is often judged by the lower boundary of the impedance that can be achieved without stability problem; in particular, minimum displayable inertia. It is well known that rendering the low value of inertia makes the admittance-type haptic interfaces unstable easily. This paper extends Time Domain Passivity Approach (TDPA) to lower down the minimum achievable inertia in an Admittance-type haptic interface. To use the well-developed TDPA framework, an admittance haptic interface should be represented in network domain with clear energy flows, which was not straightforward due to unclear causality. Therefore by introducing dependent effort and flow source concept, the admittance type haptic interface is represented in electrical network domain. This network representation allows us to have clear causality, and consequently allowing to implement TDPA. The proposed idea was experimentally verified, and found successful in bringing down the minimum inertia 10 times lower than without TDPA case.

Visual testing system for the damaged area detection of wind power plant blade

For the maintenance of wind power turbine blade, we have researched robot platforms and sensor systems. In this paper, we proposed VT (Visual Testing)-based damage detection system based on pan-tilt zoom camera system. VT is a basic but indispensable NDT (Non-Destructive Testing) for wind turbine blade. Additionally, this fixed-type damage detection system can be applied for robot localization on the wind power blade. Our experiments show that the sensor system using pan-tilt zoom camera can detect damage on the blades, localization algorithm is feasible, and the system can be deployed in wind power farm.

A Formal Method of Measuring Interactivity in HRI

In human-robot interaction (HRI), interactivity is a critical issue in terms of how they understand each other and what makes them share grounds to communicate. This paper proposes measures to evaluate interaction process using information theoretic approach. The measures are designed to quantify the level of shared ground among agents and their environment. With these measures, an experiment is carried out in which human subjects were asked to interact with a virtual robot to assign tasks with voice interface. The results of the experiments show that the measures are useful in measuring the interactivity of HRI.

A Portable Mediate Interface `Handybot` for the Rich Human-Robot Interaction

The importance of the interaction capability of a robot increases as the application of a robot is extended to a human`s daily life. In this paper, a portable mediate interface Handybot is developed with various interaction channels to be used with an intelligent home service robot. The Handybot has a task-oriented channel of an icon language as well as a verbal interface. It also has an emotional interaction channel that recognizes a user`s emotional state from facial expression and speech, transmits that state to the robot, and expresses the robot`s emotional state to the user. It is expected that the Handybot will reduce spatial problems that may exist in human-robot interactions, propose a new interaction method, and help creating rich and continuous interactions between human users and robots.

A Formal Model of Sharing Grounds for Human-Robot Interaction

In human-robot interaction, interactivity is a critical issue in terms of how they understand each other and how they make other party understand efficiently. In this paper, this issue is approached formally with respect to the mental model and shared ground. A formal model of human-robot interaction based on the information theory is presented in order to explain the mechanism of human-robot interaction and the role of shared grounds and the mental model of human-robot interaction. It is expected that the presented model of H-R interaction will increase the understanding of human-robot interaction, and may be useful in classifying, describing and predicting this interaction process

Multiple human body tracking based on Normalized Cross-Correlation of Average Histogram using the Fusion of Color and thermal image sensor

In this paper, a novel approach for tracking multiple human bodies is proposed by using sensor fusion with respect to color and thermal image sensors. In our human status monitoring system, a color and thermal image sensor is used together to make a robust human tracking. The Normalized Cross-Correlation of Average Histogram (NCCAH) Algorithm is used for analyzing between detected human bodies. Our strategy for detection and tracking is to analyze recursively the correlation between texture and thermal data. Our system tracks the moving human bodies by using the proposed NCCAH algorithm, which uses the static information of the detected bodies and the dynamic information of changes between previous and current frames. Our experiments show that the method yields good detection and tracking performance in dynamic environments, such as an occlusion, various electric heating appliances or complicated backgrounds.