Andrew Chiou

Urban Search and Rescue Robots in Test Arenas: Scaled Modeling of Disasters to Test Intelligent Robot Prototyping

This paper presents the framework to provide for the prototyping of autonomous search and rescue robots test arena. Test arenas are scaled down physical representation of disaster areas. The arena is embedded with sensors to measure the performance matrices of intelligent autonomous robot prototypes as it navigates across the test bed.

P-Expert: A Prototype Expert Advisory System in the Management and Control of Parthenium Weed in Central Queensland

This paper discusses the problem of Parthenium weed infestation in Queensland and describes P-Expert, an expert advisory system designed to provide expert knowledge in control and management strategies of Parthenium weed. P-Expert is fundamentally a hybrid fuzzy expert system incorporating technologies from fuzzy logic, relational database and multimedia systems. The primary topic of this paper will be a description of the framework of P-Expert, focussing on three main areas: (1) Layered component architecture — each component of the expert advisory system is designed as modules to facilitate maintenance, adaptability and flexibility, (2) Discourse semantics (explanatory capabilities) — provision for explanation and justifying outputs given by the expert advisory system, and (3) Meta-consequent — mapping final aggregated output from a fuzzy If-Then rule onto a finite database.

Hierarchical fuzzy logic control for multiphase traffic intersection using evolutionary algorithms

In this paper, a hierarchical fuzzy logic traffic controller is constructed for a real intersection of fourteen vehicle lanes and two pedestrian crossings controlled by signals with seven light phases. The hierarchical fuzzy controller has seven inputs as queue lengths of the seven light phases, and one output as green time of the selected phase. In the hierarchical fuzzy controller, there are six layers of fuzzy sub-controllers with two inputs and one output. The sub-controllers in the first five layers have identical structure that has two inputs of queue lengths and one output of combined queue length employed as one input of next layer. The sub-controller in the last layer has two inputs, combined queue length obtained from the fifth layer and queue length of the selected phase, and one output as green time of the selected phase. Using the developed fuzzy controller, the best fuzzy rule base is obtained based on real traffic data of the intersection by employing evolutionary algorithm. The performance of this controller is simulated and compared to that of a controller that is currently employing in the intersection. The results show that the developed fuzzy controller shortens more than 38% of the vehicle waiting time.

A game AI production shell framework: generating AI opponents for geomorphic-isometric strategy games via modeling of expert player intuition

One of the most challenging aspects of game AI implementation in computer games is in the production of artificial opponents that can demonstrate human-intuitive strategic skills. In this paper, the need and requirements for a quick and efficient method to capture these skills are identified. Based on these requirements, a framework for a game AI shell is proposed. A case study is presented to demonstrate the workability of the framework by employing a prototype shell to generate a game AI middleware for a tabletop strategy-base game.

Industrial Decision Support System (IDSS) in Weed Control and Management Strategies: Expert Advice Using Descriptive Schemata and Explanatory Capabilities

This paper details the implementation of an industrial decision support system (IDSS) application to provide expert advice on weed control and management strategies. Due to the qualitative and subjective nature of the strategies, output based intensively on conventional and numerical data poses a level of complexity that can hinder stakeholders from interpreting the data and applying the results effectively to the problem area. To overcome this, the IDSS application employs descriptive schemata and self-generated explanations.

Multiple Format Search and Rescue Robot as a Competitive Arena

Current robot competition provides both students and researchers a platform to experiment with related technology and skills to solve challenging problems. However, the entry point for participation requires very demanding prerequisites of the participants. Even though robot competition is gaining popularity as a form of digital entertainment, it still remains a very specialised domain catering to academia. This paper presents on-going work in specifying, designing and developing a multiple format search and rescue robot to provide continuity for primary age participants to postgraduate researchers.

Pan stage steady state flow model for integration within a knowledge based supervisory support system

This paper covers the development of a core industrial process model of an expert advisory system and its integration within a knowledge based supervisory support system (KBSSS) for the control and management of a sugar mill crystallization stage. The pan stage steady state flow model predicts the long term steady state flow rates and purities of process materials within the pan stage. The primary topic of this paper will be a description of the steady state flow model for the knowledge based supervisory support system. This paper will focus on: (1) design features, (2) implementation, and (3) results of the proposed model.

Player Adaptive Entertainment Computing (PAEC): Mechanism to model user satisfaction by using Neuro Linguistic Programming (NLP) techniques

The concept of Player Adaptive Entertainment Computing (PAEC) is introduced to provide personalized experiences when interacting with the entertainment media. Two of the important areas in PAEC are to create specific targeted strategies to cater for individual user, and to perform personalization. To accomplish this, a suitable mechanism to model user experience is required. This is explored in a case study using neuro linguistic programming (NLP) that models the NLP spatial anchors of experiences of a sample group of users during game play. These NLP anchors are reconstructed during game play of a second group of users allowing them to experience the same level satisfaction as the first sample group.

A novel decision support framework for industrial processes

In this paper, a novel industrial decision support framework is discussed. This framework takes consideration of special features of industrial processes and makes use of intelligent and complex systems technologies to facilitate quality decision-making in industrial setting. The structure of the framework is described in details with some case studies used to illustrate its use.

Framework for educational robotics: a multiphase approach to enhance user learning in a competitive arena

Educational robotics involves using robots as an educational tool to provide a long term, and progressive learning activity, to cater to different age group. The current concern is that, using robots in education should not be an instance of a one-off project for the sole purpose of participating in a competitive event. Instead, it should be a sustainable long-term progression spanning the primary school to pre-university level. This article presents a framework for robotics using a multiphase approach to enhance user learning in a competitive arena.