Ingeol Chun

Faults and Adaptation Policy Modeling Method for Self-adaptive Robots

Owing to the proliferation of robots in the ubiquitous world, it is imperative to develop methods and theories to integrate robots with their operational environment for creating highly reliable systems. However the ubiquitous world where robots are operated has much uncertainty and uncontrollable conditions, so that it is impossible to make robots suitable to all situations. To achieve user satisfaction and overcome abnormal situation of the ubiquitous world, robots must be dependable, safe, scalable and adaptive. Especially robots could be more intelligent in the adaptation to deal with uncertainty and uncontrollable condition.

An agent-based self-adaptation architecture for implementing smart devices in Smart Space

Smart Space is a major currently challenging domain that includes ubiquitous, grid, and pervasive computing to provide intelligence, insight, and vision for emerging world of intelligent environment, products, services and human interaction. Smart Space consists of various objects (devices and applications) and, their extremely tight integration of and coordination between information world and physical resources. In Smart Space, people are allowed to acquire useful information and control devices anytime and anywhere with various smart devices. However the physical world where devices are deployed has much uncertainty and uncontrollable conditions, so that it is impossible to make devices suited to all situations. To achieve user satisfaction and overcome the system failure, devices in Smart Space must be dependable, secure, safe, and efficient, and operate in real-time. In addition, they must be scalable, cost-effective and adaptive. Especially, to deal with uncertainty and uncontrollable condition, devices in Smart Space could be more intelligent in the adaptation.In this paper, we propose an agent-based self-adaptation architecture for implementing smart devices in Smart Space. A self-adaptive smart device reasons about its state and environment, and adapts itself at runtime automatically and dynamically in response to changes. Also we present an adaptation mechanism used to implementing a self-adaptive system.

ES-debugger : the flexible embedded system debugger based on JTAG technology

In the past the Embedded Software was used for controlling industrial apparatus but now is used in control systems for military, home appliances and automatic sensor systems and so on at present. Embedded systems are different from conventional computer system and this distinction makes it much difficult to debug as well as develop the embedded systems because of the lack of resources and sensitiveness in target system such as memory, power, timing constraint, etc. Especially huge integrated system like SoC (System-on-a-Chip)-that has memory, I/O port, etc.-has no space to locate debugging gadget so that the debugging of the embedded system is more difficult. To solve above problem, Boundary scan testing (also named JTAG) is appeared and we proposed the embedded system debugger based on JTAG technology in this paper

WiP Abstract: From Design to Operation of a Large-Scale CPS

CPS has broad spectrum of system features. A LCPS (Large-scale CPS) has a number of computing entities and physical systems including various sensors and actuators. The design and implementation of a LCPS must be more difficult and different from those of stand-alone CPS because it has more heterogeneity, more unreliability, more complexity and more unpredictability. Therefore, a LCPS must be developed under the systematic method in order to make fault-free control software at design time. In addition, LCPS requires control middleware supporting different or orthogonal QoS (Quality of Service) requirements of the subsystems composing LCPS.

Intelligent Service Robot and Application Operating in Cyber-Physical Environment

Despite the fact that intelligent service robots are increasingly needed for the convenience of human beings, there is a lack of development of interface and applications which enable humans and robots to communicate. This paper proposes a user-based application system in order to overcome this problem and contribute to popularizing intelligent service robots. User application analyzes and controls information from each system, conveys user’s commands through bluetooth communication, and shows the robot’s location and video information to users. The proposed system will enable users to control robots easily without specialized knowledge about robots while at the same time enabling managers to control and monitor robots from remote places, making system management convenient.

Metamodel-Based CPS Modeling Tool

Cyber-Physical Systems (CPS) has been attentioned for the needs of integration among computing, networking, and physical processes. After the advent of CPS, reliability, real-time and intelligence of software for preventing errors and situations that we can’t expect became major issues. In this paper, we introduce CPS Meta Modeler (CMM) to approach designing complex and large scale systems in a way of modeling in hybrid and meta-models and showing the method of modeling using ETRI CPS Modeling Language (ECML).

A GDB-based real-time tracing tool for remote debugging of SoC programs

Since embedded systems based on System-on-a-Chip(SoC) have limited resources, debugging programs in such systems requires a remote debugging system that has enough resources. However, existing JTAG based remote debugging system that uses GDB in Linux environment does not provide tracing function, so it is hard to monitor the executions of SoC program in real time. This paper adds a tracing facility to existing GDB remote debugging system to provide a real time monitoring tool. To demonstrate a real time tracing of synthetic program, Intel a Xscale PXA series processor based target system is used.

I-SEE: an intelligent search agent for electronic commerce

With the recent advent of widespread, cheap, reliable, and easy access to the Internet via the World Wide Web, information-processing is moving away from an expensive, document-based, paper-oriented approach to a faster, cheaper, electronic, Internet-based approach. Electronic commerce on the Internet is increasing at an explosive rate. Nonetheless, the spread of E-commerce is hindered by the Webs huge and diffusive extent, its distributed and heterogeneous markets, and the heavy user load entailed in finding and purchasing products. As the amount of electronic data increases, it becomes burdensome for users to find what they want. There are many search systems designed to solve these problems, but none is able to adapt and filter information for the individual user. The authors have designed a new product-finding agent named I-SEE (Intelligent SEarching AgEnt for Electronic Commerce). It is an autonomous product-finding and -filtering system that supports users from requirement acquisition to supplying specific product information.

Model-based Autonomic Computing Framework for Cyber-Physical Systems

In this paper, we present the model-based autonomic computing framework for a cyber-physical system which provides a self-management and a self-adaptation characteristics. A development process using this framework consists of two phases: a design phase in which a developer models faults, normal status constrains, and goals of the CPS, and an operational phase in which an autonomic computing engine operates monitor-analysis-plan-execute(MAPE) cycle for managed resources of the CPS. We design a hierachical architecture for autonomic computing engines and adopt the Model Reference Adaptive Control(MRAC) as a basic feedback loop model to separate goals and resource management. According to the GroundVehicle example, we demonstrate the effectiveness of the framework.

DVML: DEVS-Based Visual Modeling Language for Hybrid Systems

This paper proposes a visual modeling language for a large-scale hybrid system based on discrete event system specification formalism. Within theproposed language, basic model diagrams define the hybrid behavior of systems, whilethe structure can be represented through coupled model diagrams. The language can visually define large-scale hybrid systems without a sacrifice of formal semantics, and does not require the users to have programming skills. A prototype of the modeling and simulation environment based on an extended version of the language has been implemented.