Prolonged loneliness in older adults increases the risk of dementia, cardiovascular diseases, and premature death. However, timely detection and behavioral pattern analysis are challenging due to the slow progression of loneliness and diverse individual lifestyles. To address these barriers, we propose SMILE, an end-to-end framework for sensor-driven detection and counterfactual pattern analysis of loneliness. Our approach uses non-intrusive sensors to passively observe behavior and build a personalized profile over time. When SMILE detects meaningful changes in activity patterns, it employs generative models to produce counterfactual sensor profiles—synthetic representations of behavior predicted to result in lower loneliness scores. Suggested pattern changes are identified as the difference between observed and counterfactual profiles, highlighting the behaviors most associated with loneliness. We conducted a 6-month feasibility study with 18 participants from the United States and 15 from Japan to develop and evaluate SMILE. Our framework integrates a Time-series Transformer (TST) model for loneliness detection, which achieved 96.25% accuracy, and a diffusion model for generating behavioral explanations, which outperformed a baseline by up to 76.69%. These findings highlight the potential of SMILE to support sensor-driven, personalized loneliness detection and behavioral explanation in aging populations.

Dementia and mild cognitive impairment (MCI) represent significant health challenges in an aging population. As the search for noninvasive, precise and accessible diagnostic methods continues, the efficacy of electroencephalography (EEG) combined with deep convolutional neural networks (DCNNs) in varied clinical settings remains unverified, particularly for pathologies underlying MCI such as Alzheimer's disease (AD), dementia with Lewy bodies (DLB) and idiopathic normal-pressure hydrocephalus (iNPH). Addressing this gap, our study evaluates the generalizability of a DCNN trained on EEG data from a single hospital (Hospital #1). For data from Hospital #1, the DCNN achieved a balanced accuracy (bACC) of 0.927 in classifying individuals as healthy (n = 69) or as having AD, DLB, or iNPH (n = 188). The model demonstrated robustness across institutions, maintaining bACCs of 0.805 for data from Hospital #2 (n = 73) and 0.920 at Hospital #3 (n = 139). Additionally, the model could differentiate AD, DLB, and iNPH cases with bACCs of 0.572 for data from Hospital #1 (n = 188), 0.619 for Hospital #2 (n = 70), and 0.508 for Hospital #3 (n = 139). Notably, it also identified MCI pathologies with a bACC of 0.715 for Hospital #1 (n = 83), despite being trained on overt dementia cases instead of MCI cases. These outcomes confirm the DCNN's adaptability and scalability, representing a significant stride toward its clinical application. Additionally, our findings suggest a potential for identifying shared EEG signatures between MCI and dementia, contributing to the field's understanding of their common pathophysiological mechanisms.

Service robots have developed considerably in recent years. This study aims to propose two approaches for controlling the motion of a service robot as it pours liquid precisely from an unknown container into another unknown container without the need of any external tools. To realize this task, we must resolve two sub-tasks, which are determining the poured volume and controlling the pouring action. Our first proposal concentrates on the target container. The poured volume is calculated using a model of target container and the height of liquid in the target container. The action is controlled using a proportional-derivative controller, which considers the angular speed of the pouring container as a process variable and the poured volume as a control variable. Our second method concentrates on the pouring container. The poured volume is calculated using the relation between the angle of the pouring container and poured volume. The action is controlled with a simple proportional controller that takes the angular speed of the pouring container as a process variable and target angle as a control variable. A point cloud is used to model the two containers. These two methods were implemented in a dualarm robot system for testing, and the results show that both methods are effective for controlling precise pouring tasks.

This paper proposes a motion-planning method for a wrapping-with-fabric task. To wrap, robots require not only offhand paths (which represent the movement of a grasping point on fabric) but also actual robot motion. When a robot handles fabric, it must consider the range of motion and collisions. To overcome such limitations, the robot must pass fabric inter- and intra-hand. The proposition is based on the reliability of its movements. The method may be used to plan a combination of inter- and intra-hand passing motions to generate the basic movements for wrapping along an offhand path. Experiments show how, by using the proposed planning method, a dual-arm robot can wrap an L-shaped pipe.

This paper proposes an architecture with a two-layered structure for implementing an air-hockey robot that performs tactically in the deployment of entertainment robots. The two-layered architecture is comprised of two components: a tactical component that governs tactics and a skill component that deals with individual shots. The skill layer shoots a puck autonomously by selecting an appropriate way of shooting while considering what is physically feasible. The tactical layer controls the behavior of the skill layer by adjusting a probability function for selecting shooting methods. In this study, we implemented the architecture and conducted experiments to demonstrate its usefulness.

This paper describes emotional body expressions by flying robots for displaying emotions. Subjective evaluation experiments were performed to clarify the relation between movement parameters and subjective feeling of the expressions. Emotional body expressions were designed by eight people. The expressions were evaluated by another group of twelve people, who answered a questionnaire. The elements of emotional body expressions were investigated using the Lasso method. The results indicate movements that are most effective for displaying a particular emotion to a human. Such expressions are expected to enhance communication between the flight robots and humans.

This paper describes the implementation of a twisting skill to a multi-DOF end effector for manipulating linear deformable objects (LDOs). When manipulating of LDOs such as rope knotting, an important factor is to control torsion. However, controlling torsion is difficult using only the robot arm movement because of the range limit of a manipulator and the danger of collision with the environment such as a table and objects to be tied. On the other hand, implementing the movement into a robot hand as in-hand manipulation means that the robot-arm movement can be compact. In this study, we extract what is needed in a “twisting skill”, implement them into multi-DOF end effectors mounted on a dual-arm robot, and evaluate the skill. Then, we demonstrate rope manipulation including a knotting task and verify the effect by using a twisting skill. Our results show that the proposed method can be used to control the twisted angle of the rope with the required precision. Furthermore, using the proposed twisting skill, when placing a rope on a table top, the accuracy and reproducibility are improved compared with the case when the skill is not used. Finally, using the proposed method, clove hitches are successfully tied.

In the present paper, we propose a method for in-air knotting by a dual-arm, multi-finger robot. Most previous studies about knotting by robots assume that the rope is placed on a table. However, in-air knotting requires more skillful hand manipulations, such as swapping ropes between the right and left hands. Therefore, we herein focus on hand motion and first extract the essential hand motions for knotting, referred to herein as skill motions, by observing human knotting procedures. Next, hardware with the capability of executing these skill motions is developed. Finally, experiments are conducted to confirm that several types of knots can be tied in the air using the proposed method.

This paper describes a method for generating hand trajectories for a robot to enable it to place a rope on a table in a target shape. In many aspects of rope-work, e.g., making knots, hitches, and bends, it is important to control not only the topological relations (e.g., a tangle of string), but also to control the shape (e.g., the position, size, and so on). The shape of the rope on a table is the starting stage of the rope-work in that it affects the result of the rope-work, e.g., the position and size of a knot. Thus, it is necessary to generate hand trajectories automatically according to the target rope-shape when executing rope-work with a robot arm. To generate trajectories from a tabletop target shape, it is necessary to consider a positional relationship between the robot hand and the touchdown point, which is the point that the rope is placed on the table. In this study we propose a rope-shape model which can be used in limited situations, and derive hand trajectories from the model in a simple manner. The parameter needed for the model is easily identified. An experiment of making a loop was conducted to verify the effectiveness and limitations of the proposed method. As further verification, the method was applied to making a clove hitch using a dual-arm robot. In the starting stage of tying a clove hitch, two loops were created in a specified shape, and the clove hitch was successfully tied.

Conventional automatic doors cannot distinguish between people wishing to pass through the door and people passing by the door, so they often open unnecessarily. This paper describes an intelligent door system that observes people near a door so that the door opens at the proper time and only for people who have the intention of passing through it. To achieve this, we developed a novel laser range scanner that scans the surrounding environment three dimensionally, and formulated lightweight algorithms for the sensor. The experimental results show that the sensor can control a door according to the speed of people within the required specifications and that it correctly identifies people passing by the door.

This paper describes a localization technique for a mobile robot with unknown obstacles, such as pedestrians. It is important that a robot can run in environment freely in order to perform goods conveyance efficiently. In order to run in environment and deliver goods to a target position, self-position presumption is important. However, In real environment, robot may be unable to run to a target position, because the robot cannot localize the self-position correctly with unknown objects, such as people and goods. In this paper, this problem is solved by utilizing the robust position localization technique based on the observed space information to a delivery robot's navigation. Furthermore, by building the system which can set a robot's run path easily, flexible path creation and autonomously running are realized. we verified the validity of the proposed method by the localization experiment and the goods delivery demonstration in real environment. The system which localizes self-position with 0.1m error even in the environment where an unknown object exists, and performs goods delivery to a target position was realized.

This paper describes a localization technique for a mobile robot with unknown obstacles, such as pedestrians. Many robust localization technnique with unknown obstacles by using a particle filter and laser scanner have been propsed. In particular, Tomizawa's method(Space observation model) is effective. However, this method has problem that it does not consider the fixed object occlusion. Thus, in the environment where thin wall (this is example), the robot position cannot be localized well. In this research, we solve this problem by rebulding the space information which is observed intrinsically in predicted position by considerinmg the fixed object occlusion. From above, we realize robot position localization with more high accuracy than previous method, by solving the problem of previous method and improving the previous method. We evaluate the validity of the proposed method by a simulation and an experiment in a real environment.

Humanoid robots have discrepancies between the postures simulated on models and those of actual humanoid bodies. To correct this problem, this paper proposes a method in which a robot observes its own posture in a mirror. This method does not need an additional time-consuming setup, such as calibration of external cameras. We attach a reference point to a known position on the robot body. By observing markers and the reference point in the same view, we can estimate the position of the markers accurately based on the geometric relationship between the mirror, an object, and its mirror reflection. We evaluate the proposed method by experiments of measuring and correcting the posture of an actual robot body to attain the target posture.

This paper proposes a new strategy for making a knot with a general-purpose arm. The strategy is divided into four stages: First, a human-performed knotting action is analyzed to acquire some motion primitives. Second, based on these databases, the robot is taught to tie a knot. Third, key points on the stored trajectories are extracted. These key points are considered in a general case, and then used to implement the manipulation skills. Finally, knotting task is generated by synthesis of the manipulation skills. An overhand knot was made by the actual robot for the validity of the proposed method. This paper also describes a system for manipulating deformable objects (e.g. knotting a rope).

In this paper, we propose a task skill transfer method using a bilateral teleoperation. We assume that a task skill for a complicated motion is described by the suitable trajectory of the impedance center. Thus, a important point is how to get the suitable trajectory of the impedance center. In our method, the suitable trajectory is got from results of bilateral teleoperation. In the bilateral teleoperation, an operator constructs a motion procedure able to execute the task. The suitable trajectory of the impedance center is calculated from the position and the force data in the teleoperation experiment. A basic element of the task skill is composed of an initial condition, a task skill motion and a final condition. The initial and final conditions are described by values of both sensor and tip position. The task skill motion is described by a hybrid impedance/force control. To demonstrate how to use of our method, we choose the nut attachment task as an example

In this paper, we propose RT-middleware for robot system integration. "RT" means "robot technology", which is applied not only to industrial field but also to nonindustrial field such as human daily life support systems. RT-middleware which is proposed in this paper is a software platform for RT systems. We have studied modularization of RT elements and have developed RT-middleware, which promotes application of RT in various fields. Robotic system development methodology and our RT-middleware concepts are discussed. The RT-component, which is a basic software unit of RT-middleware based system integration, is derived from this discussion. A methodology of system development by using RT-components, and a framework for component development are proposed. Evaluations of some RT-component based systems are performed. Finally, conclusions and future work is described.

"Teaching by demonstration" is a method to generate a robot program that makes a robot do the same task as the task that a human operator demonstrates. We have developed a "teaching by demonstration in VR" system (TbDinVR) which automatically generates a robot program to work in the real world after a task is demonstrated by an operator in the virtual world. In this paper, we have explained necessary and sufficient condition for states in a task description, advantages of using state-based TbDinVR system, and a simulator for determining parameters of skill primitives by using sensor simulation in VR and a random search technique in detail. We have shown an experimental result and visualization of that result.

"Teaching by Demonstration" is a method of generating a robot program that makes a robot do the same task as the task that a human operator demonstrates. We have developed a "teaching by demonstration in VR" system that automatically generates a robot program to work in the red world after a task is demonstrated by an operator in the virtual world. This system generates a program that detects each contact state extracted from demonstration in VR and produces it by skillful motion primitives. We find that some types of motion that are included in the recorded motion given by a human operator in the VR in teaching stage cannot be realized in the execution stage if the robot tries to do the same task by using the human demonstration data even on a contact state level. In this paper, based an our observation that a human has the skill wherein contact state transitions are made non-deterministically, we propose a method of dealing with one of the important non-deterministic motions, "non-deterministic search type motion" Experimental results show the feasibility of the method.

In this paper, we present a method for analyzing parallel mechanism stiffness including elastic deformations in the structure. In our previous works, we developed a compact 6-DOF haptic interface as a master device which achieved an effective manual teleoperation. The haptic interface contains a modified Delta parallel-link positioning mechanism. Parallel mechanism is usually characterized by a high stiffness, however, it is reduced by elastic deformations of both parts and bearings. Therefore, to study such a parallel mechanism, we analyzed stiffness in a parallel mechanism, including elastic deformation of both parts and bearings. Key points of our method are a modeling of bearings, a modeling of a elasticity coefficient of a passive axis in a bearing and a modeling of a parallel mechanism. Finally, we show that our stiffness analysis method of a parallel mechanism is high reliability and precision.

Analysis of human hand grasp provides information that can be used to develop a grasping algorithm of robot hands that takes advantage of human knowledge and experience. In analysing the human hand grasp, not only motion of the fingers but also contact force acting on the fingertips should be measured. This paper presents a new master hand which can be attached to a human hand and measures motion of the human fingertips and contact force acting on the fingertips. Our purpose in developing the master hand is to clarify the strategy of a human hand grasp, which is subjected to the same physical constraints as those of a robot hand, and use the obtained information to develop a grasping algorithm of robot hands. Concepts of the development, mechanism and features of the master hand are described.

Manipulator tasks such as assembly and disassembly can generally be divided into several motion primitives. We call these "skills" and explain how most manipulator tasks can be composed of skill sequences. Skills are also used to compensate for errors both in the geometric model and in manipulator motions. There are dispensable data in the shapes, positions and orientations of objects when achieving skill motions in a task Therefore, we can derive simplified geometric models by considering both dispensable and indispensable data in a skill motion. We call such robust and simplified models "false models." This paper describes our definition of false models in planning and visual sensing and shows the effectiveness of our method using examples.

Describes a unified method for integrating different kinds of sensor-based control. We introduced "control space", which is smoothly related to actuator space. Each sensor-based control law is described in the control space. The integration is done automatically by mapping the control space to the actuator space based on local linear relations between them. We also proposed the multi-agent-based implementation of the control system. Simulation and experiment show the effectiveness of this new method.

This paper describes the network implementation for SELF, Sensorized Environment for Life, which the authors are developing. SELF is an experimental sensorized environment to support healthcare. All sensors are embedded in the environment cleverly so human in SELF may not care for being monitored from the environment. To construct such an environment and make use of the information from sensors, every sensor should be connected with the others and, consequently, they should form a sensor network. In this paper, the outline of SELF itself is described briefly but not discussed in detail. The focus of this paper is the aspects of this sensor network in SELF.

This paper describes an intelligent environmental system, SELF (Sensorized Environment for Life), that enables a person to maintain his or her health through "self-communication." The system externalizes a "self" by storing personal data such as physiological status, analyzing it, and reporting useful information to assist the person, in maintaining his or her health. To establish the methodology to create such a system, this study analyzes and re-defines the self as an object of information processing. Self-communication is nothing but understanding ourselves objectively. It is not easy to put into practice because it requires a visualized or numerized representation of oneself. As an example of an externalized self, this study focused on the physiological status of a person. The physiological status is obtained as follows: SELF observes the person's behavior with distributed sensors invisibly embedded in the daily environment, extracts physiological parameters from it, analyzes the parameters, and accumulates the results. The accumulated results are used for reporting useful information to maintain the person's health. The authors constructed a model room for SELF consisting of a bed with pressure sensor array, a ceiling lighting dome with a microphone, and a washstand with a display. Fundamental results of self-externalization using SELF are presented.

The authors describes a concept and an implementation of a networked and sensorized environment which they are developing. They call this environment SELF, named after Sensorized Environment for Life. The main objective of SELF is to construct an environment which can support healthcare of not only aged but healthy people with the information accumulated by surrounding sensors installed in it. To equip such capabilities with the system, all sensors should be embedded in the environment and systematically form a sensor network. This paper outlines the concept of SELF and shows its current implementations.

This paper proposes a new method for monitoring normal breath sounds in a daily environment where many kinds of noises exist. In a typical conventional breath monitor, a thermistor, an accelerometer, or a contact-type microphone must be attached directly to a person's body such as a part around the nose and the mouth, or a part over the chest wall and the trachea. This paper focuses on key issues stemming from a factor of the environment: extraction of breath sounds from sensor signals contaminated by environmental noise; and realization of a monitoring system suitable for our homes. Features of normal breath sounds are clarified by frequency analysis. Based on the features, a method for extracting normal breath sounds is discussed. This paper also describes a ceiling dome microphone which is a kind of a parabolic sound concentrator. It realizes a function of detecting normal breathing sounds with high sensitivity while keeping a room's appearance natural. The gain obtained by such a concentrator is analyzed theoretically from the acoustic point of view. Experimental results proved the feasibility of the proposed method.

Contact state information is indispensable for performing skillful tasks in which an end-effector should be in contact with the environment. In previous work, the authors introduced the concept of a pseudo contact point that is useful for tasks involving contact as one expression of the contact state. In this paper, an extended definition of the pseudo contact point is proposed. The pseudo contact point is defined as the foot of the perpendicular from the nearest point on a reference object to the line on which the contact point and external force vector lie. The system may estimate the contact state and its transitions by monitoring the pseudo contact point in spite of the fact that the pseudo contact point does not always denote the real contact point. Experimental results suggest the effectiveness of our method, which is particularly valuable in providing an intuitive positional description of a contact state.

This paper proposes the concept of "agent networks" to realize flexible and unified integration of multi-agent intelligent robot systems. As an example, we developed a hand-to-hand robot on the agent network. Its robust and flexible behaviour proved the effectiveness of the method.

When the desired motion of the grasped object is given or planned, the coordination of motion of the manipulator and manipulation with its gripper is needed in order to achieve the desired motion of the object. In this paper, we have defined "graspability", "manipulability", and "virtual gripper". These concepts are necessary to deal with the geometric relation between a dexterous hand and its grasped object in the planning process and designing process. Since the coordination of motion of the manipulator and manipulation of its gripper is needed in order to achieve the desired motion of the grasped object, we have developed an algorithm that generates these coordinated motions: the path of the manipulator and the manipulation path of the object with respect to the coordinate frame of the gripper. This paper describes "graspability", "manipulability", "virtual gripper", an algorithm that generates the coordinated motions, and estimation of a "manipulable region" of a dexterous gripper which is needed in a specific task.

We have developed a teaching system based on the assumption that the assembly task is a series of operations for achieving a target contact state among objects through changing their contact states. The features of our system are as follows: 1) teaching data for fine motion is extracted from a demonstration by an operator in a virtual world, 2) the operator can edit intermediate expressions on a display, 3) intermediate expressions can be used by any robot system with a skill library which has the ability to achieve the corresponding contact states. This paper describes the configuration of the system, program generation using intermediate expressions and a skill library to achieve each contact state.

This paper proposes a multi-agent based implementation of robot skills. As an example, a skill motion for assembly tasks, "move-to-touch", has been implemented on a multi-agent robot control system developed on a transputer network. The skill motion is constructed as an integration of several agents, a motion control agent, event detection agents and a management agent. Our implementation has achieved high flexibility, reliability and extensibility of the skill control system.

Monitoring of a contact state is an important technique to achieve tasks in which an end-effector should be in contact with an object. This paper introduces a pseudo contact point that is useful for precise tasks including contact processes. When the position of a point contact between the end-effector and the object moves to another position, the pseudo contact point denotes the present contact point. The pseudo contact point is defined as the foot of the perpendicular from the previous contact point to the line on which the external force vector lies. Experimental results suggest the applicability of our method.

Generally, a manipulator task can be divided into several motion primitives called "skills". Skill-based motion planning is an effective way to execute a complicated task. Also when planning tasks, a technique of fine motion planning such as the backprojection method in configuration space is often used. In this paper we introduce the concept of skill into the backprojection method. A backprojection is shaped by a pattern of motion on each skill primitive, and a skill library is constructed. Motion planning based on backprojection is achieved by combining several skills in the skill library. Therefore, a general and skilful motion planning can be derived.

We present a novel design and implementation of a foveated wide angle lens. It is strongly motivated towards applications for active vision in complex and dynamic environments. The projection curve is designed so that it facilitates active vision algorithms for motion analysis, object identification, and precise fixation. The implementation achieves small size and light weight so that the lens can be swung at a high speed.

Presents a scheme for specifying and executing purposive navigation tasks for a behaviour-based mobile robot. A user specifies the robot's navigation task in general and qualitative terms using a graphical resource called the purposive map (PM). The robot navigates using the incomplete and approximate information stored in the PM as an aid to achieve the specified mission. The robot is able to augment the knowledge provided in the PM with environment information learnt by the robot. Using the augmented PM, the robot learns how to perform efficient obstacle avoidance. The authors present experimental results using a real robot to show their scheme is robust. The authors' robot can escape from dead-ends, can deduce that goals are unreachable and can withstand disturbances to the environment between missions.

We have been developing a multi-agent robot control system for an intelligent robot which is an integration of many kinds of functional modules. The control system consists of multiple transputers and workstations. To support flexible and high speed communication between agents on the network, we have developed a router software. Using the router, the system realizes a very flexible robot control system which supports throughout of robot control, from the low level control such as servo control to the high level control such as motion planning. This paper describes the multi-agent robot control system with the router software, estimation of its communication performance and a multi-agent based implementation of task coordinate servo control of a direct drive manipulator: ETA3.

The authors present the assembly-plan-from-observation (APO) method for robot programming. The APO method aims to build a system that has the capability of observing a human performing an assembly task, understanding the task based on the observation, and subsequently generating a robot program to achieve the same task. This paper focuses on the task recognition module (TRM), the main component of a complete APO system. The TRM recognizes object configurations before and after an assembly task, detects a configuration transition, and infers the assembly task that causes such a configuration transition. We assume that each assembly task aims to achieve a face contact relation between an object that has just been manipulated and the stationary environmental objects. We prepare abstract task models that associate transitions of face contact relations with assembly tasks that achieve such transitions. Next, we implement TRM in order to verify two issues: 1) that such a contact transition can be recovered from the output of the object recognizer; and 2) that given these relation transitions, it is possible to use the abstract task models to generate robot motion commands.>

The authors have been developing a novel method for programming a robot, called the assembly-plan-from-observation (APO) method. The APO method aims to build a system that has threefold capabilities. It observes a human performing an assembly task, it understands the task based on this observation, and it generates a robot program to achieve the same task. This paper concentrates on the APO's main loop, which is task recognition. Using object recognition results, the task recognition module determines what kind of assembly task is performed. A previous system, recognizes assembly tasks which only handle polyhedral objects. The system reported here handles curved objects and other mechanical contacts as well. The authors define task models for these cases and show that task models are useful in recognizing assembly tasks, and that it is possible to generate robot motion commands for repeating the same assembly task.

A novel method for programming a robot, called the assembly-plan-from-observation (APO) method, is proposed. The APO method aims to build a system that has the capability of observing a human performing an assembly task, understanding the task based on the observation, and generating the robot program to achieve the same task. Assembly relations that serve as the basic representation of each assembly task are defined. It is verified that such assembly relations can be recovered from the observation of human assembly tasks, and that from such assembly relations it is possible to generate robot motion commands to repeat the same assembly task. An APO system based on the assembly relations is demonstrated.>

The authors propose a novel method to program a robot, the assembly-plan-from-observation (APO) method. The APO method aims to build a system that has the capability of observing a human performing an assembly task, understanding the task based on the observation, and generating the robot program to achieve the same task. Assembly relations which serve as the basic representation of each assembly task are defined. It was verified that such assembly relations can be recovered from the observation of human assembly tasks, and that from such assembly relations, it is possible to generate robot motion commands to repeat the same assembly task. An APO system based on the assembly relations was demonstrated.>

We have been developing a novel method to program a robot, an APO (assembly-plan- from-observation) method. A human performs assem- bly operations in front of the APO system's TV caimera. The APO system recognizes such assembly operations and generates an assembly plan to repeat the :assembly operations using its robot arm. We have developped a APO system which extracts two kinds of information from observed object config- urations: 1) face-contact relation and 2) motion pa- rameters necessary to move objects around. Under the presence of error, due to error-contaminated motion pa- rameters, the system may fail to perform an assembly operation, although a face-contact relation and thus an assembly operation is correctly recovered. This paper proposes a method to correct eirr~lneous motion parameters based on a face contact re)(ation. We assume that a given face contact relation reflects the actual face contact correctly. Motion paramleters are determined by solving face contact constraint equaitions, given by face contact relations, simultaneously )using the singular value decomposition method. We implement this method in the APO system, apply the method to several assembly examples, and verify the effectiveness of tlhe method. are the most promising. Yet, these methods are often in- convepient and impractical.

A model-based robot system is described. This system was developed to replace and improve upon the conventional robot control scheme of teaching-by-showing and replay. The key concept of this system consists of the combination of manipulation skills with an environment model. The model provides the geometric structure and the physical properties of the objects in the environment. Manipulation skills enable reliable task execution in the presence of unavoidable errors and uncertainties. Skills are implemented within the hybrid position/velocity/force control scheme and incorporate sensing procedures to detect the state to be achieved. An experiment involving the disassembly of a valve using this system is presented.>

A model-based robot system is described. This system was developed to replace and improve upon the conventional robot control scheme of teaching-by-showing and replay. The key concept of this system consists of the combination of manipulation skills with an environment model. The model provides the geometric structure and the physical properties of the objects in the environment. Manipulation skills enable reliable task execution, in the presence of unavoidable errors, and uncertainties. Skills are implemented within the hybrid position/velocity/force control scheme and incorporate sensing procedures to detect the state to be achieved. An experiment involving the disassembly of a valve using this system is presented.>