Norman E. Gough

Social Virtual Worlds for Technology-Enhanced Learning on an Augmented Learning Platform.

Virtual worlds have been linked with e‐learning applications to create virtual learning environments (VLEs) for the past decade. However, while they can support many educational activities that extend both traditional on‐campus teaching and distance learning, they are used primarily for learning content generated and managed by instructors. With the evolution of internet technology, social virtual worlds (SVWs) are now able to facilitate more social interaction, efficient visual communication, integration of rich media and sharing of student‐generated content. They offer the prospect of lively interactive virtual communities in which users interact through their emotional avatars in a 3D virtual world. SVWs are being increasingly embedded into e‐commerce and e‐learning, and challenge our ideas about the next generation of VLEs. This article outlines the impact of emerging social networking technologies on the internet, reveals the convergence between social networking and virtual worlds for technology‐enhanced learning (TEL), and examines the way in which SVWs are transforming the nature of learning as a social practice. The design and implementation of an innovative social interactive learning platform is presented, which augments SVWs and other social networking services with conventional learning management and student support systems. Practical experiments are described that have been prototyped on this platform, including e‐tutoring and student‐led exhibitions. The results demonstrate that an SVW can greatly enhance student‐centred active learning experience on the augmented learning platform in comparison to traditional VLEs. It is shown that the platform has the potential to support both formal and informal learning, as well as facilitating social interaction, self‐motivation, active engagement and creative thinking in TEL.

A new animation approach for visualizing intelligent agent behaviours in a virtual environment

Animated intelligent agents play an important role in dynamic virtual environments. They are able to perform behaviours based on internal states and external stimuli. Animation techniques are crucial to agent behaviour visualization in order to achieve effective interaction between agents and human users. This paper proposes a new animation procedure to visualise realistic intelligent agent behaviours. The animation system consists of three parts, namely synthetic vision, FzFSM with memory and an animation library that uses mesh skinning techniques. The agent is capable of exhibiting non-repetitive behaviours to reflect change from both its internal states and its environment. The system is implemented on a PC platform with modern 3D acceleration hardware. The result of this work can be applied to real time 3D graphics applications in the field of entertainment and multimedia.

Microcontroller based genetic algorithms

Abstract This work is concerned with the implementation of evolutionary based computer algorithms, genetic algorithms (GAs), on microcontrollers. Microcontrollers are low resource platforms not normally associated with GAs, which are typically resource intensive. This implementation will add to a suite of tools, based on different soft computing techniques, which will solve a range of optimization and decision-making problems that are essential in order for an autonomous guided vehicle (AGV) to carry out complex and highly involved tasks. Two examples are presented: path planning and the T-sequencing problem to demonstrate that the GA on an 8051 microcontroller is capable of carrying out necessary optimization and decision-making for AGVs.

Modelling and Scheduling Control of FMS based on Stochastic Petri-Nets

Abstract Recent results are given on a computer-aided analytic/simulation methodology for designing a class of realizable FMS scheduling controls. It is implemented by means of a PC-oriented program package, PETSIM, written in C+. It encompasses a bounded iterative sequence of dual-model building, simulation and performance evaluation based on Stochastic Petri Nets and Queueing Networks so that a sub-optimal plant activity sequence is obtained, and a Petri-net controller model derived. Analytic formulae are derived indicating lower and upper bounds on the desired solutions, together with simulation results.

A Genetic Algorithm for Path Planning

Abstract This paper details work on the development of an path planning system for an automated guided vehicle. An evolutionary approach, using a genetic algorithm, is taken which uses an interpretation scheme to ensure that all generated solutions are legal. Results are presented which show that this approach has the ability to plan a route between two points that is optimal in terms of both distance travelled and time. The algorithm is designed so that it could be embedded within a microcontroller and provide some intelligent on board control.

Decentralized Control in Multivariable Convolution Systems

Abstract This paper considers decentralized control of linear multivariable systems in a discrete convolution framework. Definitions and theorems are presented that indicate whether a decentralized system is stable, minimal phase and robust, and guide-lines are presented for choosing the loop interconnections. Existing steady-state results are extended to the dynamic case in terms of characteristic patterns and examples are given that implement the methodology using the simulation package CBSL.

Design and Implementation of an Automated Industrial Reheating Electrical Furnace

Abstract The design, implementation and on-site testing of an automated, programmable reheating electrical furnace, controlled to achieve and sustain three separate thermodynamic steady-state operations are presented The resulting automation based on a model of two-level, Partially decentralized control with overlapping between local controls, ensures good performance in both normal and safety-critical states

Iterative Learning Identification and Design of Industrial Furnace Process Control

Abstract An applications oriented control systems engineering approach for a class of well- posed thermal systems, e.g. industrial furnaces and ovens, that is consistent with most of theoretical results in systems and control sciences has been elaborated and tested in designing controls for several industrial-scale furnaces. It provides a methodology for iterative learning and resolving process identification and control design for multi-variable systems within a discrete convolution framework and using truncated k-time sequence matrices of characteristic input-output modes as well as their characteristic patterns and singular characteristic patterns, starting with standard non-parametric process time-domain models identified under operating conditions. Within computer process control environment and for practical engineering and maintenance reasons, digital implementations are sought in terms of partial steady-state decoupling and two-term laws or combination of certain MIMO and SISO controls. The pusher furnace in Skopje Steelworks is used to illustrate this methodological approach

Multivariable Control System Studies in the Time-Domain Using WCBSL

Abstract WCBSL is a Windows Convolution-Based Simulation Language for time-domain analysis and design of linear feedback systems based on convolution algebra The theoretical basis is outlined and the language and methodology are illustrated. Characteristic patterns and vectors are defined and the effect of delay is considered-Dynamic criteria for choosing between decentralized and multivariable control are illustrated using the model of an industrial air compressor.

Computer-aided graph-network path-motion recognition in robotized MIMO FMS structures

A computer-assisted methodology that permits convenient and simple CAD procedures for FMS-like systems has been developed. It is aimed at modeling, simulation and analysis of FMS (flexible manufacturing system) functional models with respect to the design of scheduling coordination controls. The methodology is based on the digraph or Petri network formalisms and situation or event-status recognition techniques using adjoined sets of knowledge for the system studied.<<ETX>>