Jeong-Yean Yang

Affective interaction with a companion robot in an interactive driving assistant system

Driving assistant systems are becoming the attractive service tasks in the field of intelligent robotics. Humans meet a variety of situations while driving cars and robotic systems help humans to understand how surrounding situations change and how robot systems are aware of given situations. From the viewpoint of interaction performance, proper situation awareness by a robotic assistant in a car and the relevant determination of corresponding reactions are crucial prerequisites for long term interaction between a human driver and a cars robotic system. In this paper, we focus on preserving human-robot interaction for driving situations, considering how many types of cognitive situation occur and how affective interaction can be designed for the robotic driving assistant. The appropriateness of the driving situation and of the robotic reactions are testified in our experiment system, including the development of a virtual driving environment and tablet-based robotic agent system.

Interaction with objects inside a media space

In this paper, we describe media that can provide a physics engine with predefined object information. Once the object descriptors are embedded into an image as metadata via a technique that hides the information, the information can be decoded for the construction of a rigid body object during rendering. In this manner, a user can interact with and obtain sensory feedback from the object in a world constructed from a single proposed image.

Driving situation-based real-time interaction with intelligent driving assistance agent

Driving assistance systems (DASs) can be useful to inexperienced drivers. Current DASs are composed of front rear monitoring systems (FRMSs), lane departure warning systems (LDWSs), side obstacle warning systems (SOWSs), etc. Sometimes, DASs provide unnecessary information when using unprocessed low-level data. Therefore, to provide high-level necessary information to the driver, DASs need to be improved. In this paper, we present an intelligent driving assistance robotic agent for safe driving. We recognize seven driving situations, namely, speed bump, corner, crowded area, uphill, downhill, straight, and parking space, using hidden Markov models (HMMs) based on velocity, accelerator pedal, and steering wheel. The seven situations and global positioning system information are used to generate a situation information map. The developers of a navigation system have to tag driving events by themselves. In contrast, our driving assistance agent tags situation information automatically as the vehicle is driven. The robotic agent uses the driving situation and status information to assist safe driving with motions and facial and verbal expressions.

7-DOF horseback riding simulator based on a crank mechanism with variable radius and its inverse kinematics solution

This paper proposes a robotic simulator for horseback riding that incorporates a crank mechanism with a variable radius. In horseback riding simulations, it is necessary to efficiently generate vertical sinusoid motion. A novel crank mechanism is proposed to change the unidirectional rotation of a motor to sinusoid wave motion with an arbitrary magnitude and frequency. On the basis of the results of a kinematic analysis, the singularity and redundancy issues of the system are presented. To avoid the singularity problem, an inverse kinematics solution is presented and the result of the simulation shows stable solution of the inverse kinematics.

Development of emotional attachment on a cleaning robot for the long-term interactive affective companion

Emotional interaction has been widely applied to many robotic applications. Emotional behavior, which is complementary to limited cognitive ability in a service robot, is becoming the attractive service task even including its own therapeutic functions. This paper focused on the user study of an interactive affective companion by using a cleaning robot as a pet. The cognitive and emotional subsumption model is designed for the robust behavior selection with the response to humans commands. In addition, we testified the appropriateness of long-term interaction through tablet-based modality.

The modified inverse kinematics on variable crank for the XENMA pelvis rehabilitation robot

We are developing a XENMA pelvis rehabilitation robot for the gait motion which can generate periodic motion with specific variable crank structures. We are focused on trigonometric property of pelvis motion on gait. We designed robot for patient to ride it, and control his pelvis with two crank-slider mechanisms and variable crank structure that can control its crank length. We also designed modified inverse kinematics methods for a variable crank structures with both geometrical and resolved motion rate control solutions. Modified IK method is simulated and result shows our solution is valid and makes better following performance.

Episodic memory system of affective agent with emotion for long-term human-robot interaction

In long-term human robot interaction, memory of experience is important. This paper suggests episodic memory system for an affective robot which contains emotion. Storage, retention and retrieval of episodic memory are modeled with psychological bases. This modeled memory system will be implemented to pet-like virtual agent with reactive emotion generation model[2].

Human-Robot Interaction-Based Intention Sharing of Assistant Robot for Elderly People

Intention reading is considered as an important issue in the field of human-robot interaction (HRI). Specifically, for cooperation between a human and a robotic system, a robot should be aware of a given situation and human intention, which necessarily requires high-level knowledge and balanced interaction. For the disabled and elderly, who may not be familiar with manipulating robotic systems, intention sharing between two different agents is preferable for accurate intention reading. In this paper, we focus on the effect of intention sharing of a human user and a robot assistant to improve cooperation.

Laser scanner based foot motion detection for intuitive robot user interface system

In order to achieve effective interaction between humans and robots, user interface system is one of important factors. To achieve a valid and appropriate user interface system, we have introduced a new system that consists of a beam projector that can display images for interacting with users and a laser scanner that can detect users foot motion so that the system recognizes users input and intention. In this paper, we have focused on how to cause projected images to be rectangular so that users can interact with ordinary rectangular images instead of trapezoidal images that will be produced if the beam projector attached on the robot projects images sideling and how to distinguish users foot motions according to two classifications, drag motion and click motion. Finally, we have executed a test for checking the validity of the proposed system and have examined limitations and proposed further work for this system.

A walking horse dynamic model for generating sensations on a simulator at various walking speeds

A horseback riding simulator is developed to overcome challenges occurring on real horses. The goal of the simulator is to provide a real horseback riding-like sensation to the user in various situations with limited workspace. This paper focuses on generating various trajectories at different walking speeds with dynamic effects. A walking horse dynamic model is purposed based on a two-linked rimless wheel and is modified with 3D-captured data from a real horse. A controller is also designed for passing and compression-recoil modes, aiding in changing the models walking speed using speed change input. A dynamics simulation is performed and results show that the suggested walking model covers trajectory with measured data and shows more variation.