• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.9 no.1
• /
• pp.1-8
• /
• 2008

We propose how to program the off-line buffing robot path along shoes' outsole shape in the footwear buffing process by a 5-axis microscribe system like robot mechanism. The microscribe system we developed consists of a 5-axis robot link with a turn table, a signal processing unit, PC and an application software program. Itmakes a robot path on the shoes' upper in accordance with the movement of a microscribe with many joints. The developed system calculates the encoder pulse values for the microscribe arm's rotation and transmits the angle pulse values to the PC through a processing unit. Denavit-Hartenberg's(D-H) direct kinematics is used to make the global coordinate from microscribe joint one. Problems with the microscribe's kinematics can be solved efficiently and systematically by D-H representation. With the coordinate values calculated by D-H equation, our system can draw a buffing gauge-line on the upper sole. We obtain shoes' outline points, which are 2 outlines coupled with the points and the normal vector based on the points. By applying the system to the buffing robot in a flexible manufacturing system, it can be used effectively to program the path of a real buffing robot.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2001.01a
• /
• pp.59-64
• /
• 2001

The robot soccer simulation game is a dynamic multi-agent environment. In this paper we suggest a new reinforcement learning approach to each agent's dynamic positioning in such dynamic environment. Reinforcement learning is the machine learning in which an agent learns from indirect, delayed reward an optimal policy to chose sequences of actions that produce the greatest cumulative reward. Therefore the reinforcement learning is different from supervised learning in the sense that there is no presentation of input pairs as training examples. Furthermore, model-free reinforcement learning algorithms like Q-learning do not require defining or learning any models of the surrounding environment. Nevertheless it can learn the optimal policy if the agent can visit every state- action pair infinitely. However, the biggest problem of monolithic reinforcement learning is that its straightforward applications do not successfully scale up to more complex environments due to the intractable large space of states. In order to address this problem. we suggest Adaptive Mediation-based Modular Q-Learning (AMMQL)as an improvement of the existing Modular Q-Learning (MQL). While simple modular Q-learning combines the results from each learning module in a fixed way, AMMQL combines them in a more flexible way by assigning different weight to each module according to its contribution to rewards. Therefore in addition to resolving the problem of large state effectively, AMMQL can show higher adaptability to environmental changes than pure MQL. This paper introduces the concept of AMMQL and presents details of its application into dynamic positioning of robot soccer agents.

• Proceedings of the Korea Society for Simulation Conference
• /
• 2001.10a
• /
• pp.321-324
• /
• 2001

The robot soccer simulation game is a dynamic multi-agent environment. In this paper we suggest a new reinforcement learning approach to each agent's dynamic positioning in such dynamic environment. Reinforcement Beaming is the machine learning in which an agent learns from indirect, delayed reward an optimal policy to choose sequences of actions that produce the greatest cumulative reward. Therefore the reinforcement loaming is different from supervised teaming in the sense that there is no presentation of input-output pairs as training examples. Furthermore, model-free reinforcement loaming algorithms like Q-learning do not require defining or loaming any models of the surrounding environment. Nevertheless it can learn the optimal policy if the agent can visit every state-action pair infinitely. However, the biggest problem of monolithic reinforcement learning is that its straightforward applications do not successfully scale up to more complex environments due to the intractable large space of states. In order to address this problem, we suggest Adaptive Mediation-based Modular Q-Learning(AMMQL) as an improvement of the existing Modular Q-Learning(MQL). While simple modular Q-learning combines the results from each learning module in a fixed way, AMMQL combines them in a more flexible way by assigning different weight to each module according to its contribution to rewards. Therefore in addition to resolving the problem of large state space effectively, AMMQL can show higher adaptability to environmental changes than pure MQL. This paper introduces the concept of AMMQL and presents details of its application into dynamic positioning of robot soccer agents.

• The Transactions of the Korea Information Processing Society
• /
• v.7 no.11S
• /
• pp.3732-3838
• /
• 2000

As the Internet sites and users have rapidly been increased, the development for search engines has also been accelerated to satisfy users' expectations. As the result, not only the action of collecting documents through many search engines gave hosts workload, but also regular updating all the information is needed since information is newly added. With the circumstances, the necessity of the technology to collect massive information in hosts has been increased for the speed which is a basic requisite of search systems, and for more accurate collection of documents. Also, the role of search engines grows bigger for Internet users' various demands and flexible process through World Wide Web. In this paper, we design and implement a robot agent and a remote control system which doesn't give an excessive workload on a target server and makes the collection of documents done in a short period by considering an average workload rate on the target server and the rate of the workload that a robot experience in collection time, after we compare and analyze the existing Robot Agent Systems and supplement their weak points.

• Journal of Internet Computing and Services
• /
• v.10 no.6
• /
• pp.27-36
• /
• 2009

PSTN (public switched telephone network) provides voice communication service, whereas IP network provides data oriented service, and we can use IP network for multimedia transport service (e.g. voice over IP service) with economic price. In this paper, we propose RoIP (robot on IP) service scenario, signaling call flow, and implementation to provide home automation and monitoring service for remote site users. In our scheme, we used a extended SIP (session initiation protocol) for signaling protocol between remote site users and home robots. For our bearer transport control, we implemented H.263 video codec over RTP (real-time transport protocol) and additionally DTMF (dual tone multi-frequency) transport for robot actuator control. We implemented our scheme on home robots and experimented with KTF operator network, and it shows good communication quality (average MOS = 9.15) and flexible robot controls.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.39B no.7
• /
• pp.479-490
• /
• 2014

A lower extremity exoskeleton is a robot device that attaches to the lower limbs of the human body to augment or assist with the walking ability of the wearer. In order to improve the wearer's walking ability, the robot senses the wearer's walking locomotion and classifies it into a gait-phase state, after which it drives the appropriate robot motions for each state using its actuators. This paper presents a method by which the robot senses the wearer's locomotion along with a novel classification algorithm which classifies the sensed data as a gait-phase state. The robot determines its control mode using this gait-phase information. If erroneous information is delivered, the robot will fail to improve the walking ability or will bring some discomfort to the wearer. Therefore, it is necessary for the algorithm constantly to classify the correct gait-phase information. However, our device for sensing a human's locomotion has very sensitive characteristics sufficient for it to detect small movements. With only simple logic like a threshold-based classification, it is difficult to deliver the correct information continually. In order to overcome this and provide correct information in a timely manner, a probabilistic gait-phase classification algorithm is proposed. Experimental results demonstrate that the proposed algorithm offers excellent accuracy.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.26 no.2
• /
• pp.314-321
• /
• 2002

An endoscope is usually inserted into the human body for the inspection of the gullet, stomach, and large intestine (colon) and this may cause discomfort to patients and damage to tissues during diagnostic or therapeutic procedures. This situation necessitates a self-propelling endoscope. There are many kinds of mechanism to move in a rigid pipe. However, these methods are difficult to apply directly to the endoscope. The main reason is that human intestine cannot be considered as a uniform, straight, and rigid pipe. This paper proposes a flexible loop wheel mechanism, which is adaptable to the human intestine. This mechanism is designed and fabricated by a simple modeling, and tested by an experiment. Finally, the actuator is inserted into the pig colon.

• Journal of Korean Institute of Industrial Engineers
• /
• v.21 no.1
• /
• pp.119-135
• /
• 1995

We consider problems of tool loading and scheduling in a flexible manufacturing system (FMS) in which tool transportation constitutes the major portion of material flows. In this type of FMSs, parts are initially assigned to machines and released to the machines according to input sequencing rules. Operations for the parts released to the machines are performed by tools initially loaded onto the machines or provided by an automatic tool transport robot when needed. For an efficient operation of such systems, therefore, we may have to consider loading and scheduling problems for tools in addition to those for parts. In this paper, we consider three problems, part loading, tool loading, and tool scheduling problems with the overall objective of minimizing the makespan. The part loading problem is solved by a method similar to that for the bin packing problem and then a heuristic based on the frequency of tool usage is applied for tool loading. Also suggested are part input sequencing and tool scheduling rules. To show the effectiveness of the overall algorithm suggested here, we compare it with an existing algorithm through a series of computational tests on randomly generated test problems.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1996.04a
• /
• pp.141-146
• /
• 1996

Hand posture and force, which define aspects of the way an object is grasped, are features of robotics manipulation. A means for specifying these grasping "flavors" has been developed that uses an instrumented glove equipped with joint and force sensors. The new grasp specification system is being used at the Pennsylvania State University (Penn State) in a Virtual Reality based Point-and-Direct(VR-PAD) robotics implementation. In the Computer Integrated Manufacturing (CIM) Laboratory at Penn State, hand posture and force data were collected for manipulating bricks and other items that require varying amounts of force at multiple pressure points. The feasibility of measuring desired grasp characteristics was demonstrated for a modified Cyberglove impregnated with FSR (Force Sensitive Resistor) pressure sensors in the fingertips. A joint/force model relating the parameters of finger articulation and pressure to various lifting tasks was validated for the instrumented "wired" glove. Operators using such a modified glove may ultimately be able to configure robot grasping tasks in environments involving hazardous waste remediation, flexible manufacturing, space operations and other flexible robotics applications. In each case, the VR-PAD approach improved the computational and delay problems of real-time multiple-degree-of-freedom force feedback telemanipulation.ck telemanipulation.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1996.11a
• /
• pp.318-322
• /
• 1996

This paper presents a technique to control a robot which has a flexible manipulator moving in a vertical plane. The flexible manipulator is modeled as an Euler-Bernoulli beam. Elastic deformation is represented using the assumed modes method. A comparison function which satisfies all geometric and natural boundary conditions of a cantilever beam with an end mass is used as an assumed mode shape. Lagrange's equation is utilized for the development of a discretized model. A control algorithm is developed using a simple PID control technique. The proportional, integral and derivative control gains are determined based on the dominant pole placement method and tuned to show no overshoot and having a short settling time. The effectiveness of the developed control scheme is showed experimentally. In the position control experiment, three different end masses are used. The experimental results shows little overshoot, no steady state error, and less than 2.5 second settling time in case of having an end mass which is equivalent to 45% of the total system weight. Also the residual vibration of the end point is effectively controlled.