Classifying an opponent’s behaviour in robot soccer

This paper illustrates the prediction of opponent behaviour in a competitive, highly dynamic, multi-agent and partially observable environment, namely RoboCup small size league robot soccer. The performance is illustrated in the context of the highly successful robot soccer team, the RoboRoos. The project is broken into three tasks; classification of behaviours, modelling and prediction of behaviours and integration of the predictions into the existing planning system. A probabilistic approach is taken to dealing with the uncertainty in the observations and with representing the uncertainty in the prediction of the behaviours. Results are shown for a classification system using a Naïve Bayesian Network that determines the opponent’s current behaviour. These results are compared to an expert designed fuzzy behaviour classification system. The paper illustrates how the modelling system will use the information from behaviour classification to produce probability distributions that model the manner with which the opponents perform their behaviours. These probability distributions are show to match well with the existing multi-agent planning system (MAPS) that forms the core of the RoboRoos system.

Achieving cooperation in a distributed multi-robot team

DMAPS (Distributed Multi-Agent Planning System) is a planning system developed for distributed multi-robot teams based on MAPS (Multi-Agent Planning System). MAPS assumes that each agent has the same global view of the environment in order to determine the most suitable actions. This assumption fails when perception is local to the agents: each agent has only a partial and unique view of the environment. DMAPS addresses this problem by creating a probabilistic global view on each agent by fusing the perceptual information from each robot. The experimental results on consuming tasks show that while the probabilistic global view is not identical on each robot, the shared view is still effective in increasing performance of the team.

A Helicopter named Dolly : behavioural cloning for autonomous helicopter control

This paper considers the pros and cons of using Behavioural cloning for the development of low-level helicopter automation modules. Over the course of this project several Behavioural cloning approaches have been investigated. The results of the most effective Behavioural cloning approach are then compared to PID modules designed for the same aircraft. The comparison takes into consideration development time, reliability, and control performance. It has been found that Behavioural cloning techniques employing local approximators and a wide state-space coverage during training can produce stabilising control modules in less time than tuning PID controllers. However, performance and reliabity deficits have been found to exist with the Behavioural Cloning, attributable largely to the time variant nature of the dynamics due to the operating environment, and the pilot actions being poor for teaching. The final conclusion drawn here is that tuning PID modules remains superior to behavioural cloning for low-level helicopter automation.

A modified particle filter for simultaneous robot localization and landmark tracking in an indoor environment

This paper presents the implementation of a modified particle filter for vision-based simultaneous localization and mapping of an autonomous robot in a structured indoor environment. Through this method, artificial landmarks such as multi-coloured cylinders can be tracked with a camera mounted on the robot, and the position of the robot can be estimated at the same time. Experimental results in simulation and in real environments show that this approach has advantages over the extended Kalman filter with ambiguous data association and various levels of odometric noise.

A global vision system for a robot soccer team

This paper describes the real time global vision system for the robot soccer team the RoboRoos. It has a highly optimised pipeline that includes thresholding, segmenting, colour normalising, object recognition and perspective and lens correction. It has a fast ‘paint’ colour calibration system that can calibrate in any face of the YUV or HSI cube. It also autonomously selects both an appropriate camera gain and colour gains robot regions across the field to achieve colour uniformity. Camera geometry calibration is performed automatically from selection of keypoints on the field. The system achieves a position accuracy of better than 15mm over a 4m × 5.5m field, and orientation accuracy to within 1°. It processes 614 × 480 pixels at 60Hz on a 2.0GHz Pentium 4 microprocessor.

Framework for the long-term operation of a mobile robot via the Internet

This paper describes an autonomous docking system and web interface that allows long-term unaided use of a sophisticated robot by untrained web users. These systems have been applied to the biologically inspired RatSLAM system as a foundation for testing both its long-term stability and its practicality. While docking and web interface systems already exist, this system allows for a significantly larger margin of error in docking accuracy due to the mechanical design, thereby increasing robustness against navigational errors. Also a standard vision sensor is used for both long-range and short-range docking, compared to the many systems that require both omni-directional cameras and high resolution Laser range finders for navigation. The web interface has been designed to accommodate the significant delays experienced on the Internet, and to facilitate the non- Cartesian operation of the RatSLAM system.

Control issues for velocity sourced Series Elastic Actuators

The Series Elasic Actuator has been proposed as a method for providing safe force or torque based acutation for robots that interact with humans. In this paper we look at some outstanding issues in the implementation and control of Series Elastic Actuators. The study addresses issues in making the Series Elastic Actuator respond effectively in the presence of physical difficulties such as restriction, using a computation efficient controller. The improvement over previous implementations is achieved by treating the motor as a velocity source to the elastic element, rather than as a torque source.

Learning to avoid indoor obstacles from optical flow

Optical flow (OF) is a powerful motion cue that captures the fusion of two important properties for the task of obstacle avoidance − 3D self-motion and 3D environmental surroundings. The problem of extracting such information for obstacle avoidance is commonly addressed through quantitative techniques such as time-to-contact and divergence, which are highly sensitive to noise in the OF image. This paper presents a new strategy towards obstacle avoidance in an indoor setting, using the combination of quantitative and structural properties of the OF field, coupled with the flexibility and efficiency of a machine learning system.The resulting system is able to effectively control the robot in real-time, avoiding obstacles in familiar and unfamiliar indoor environments, under given motion constraints. Furthermore, through the examination of the networks internal weights, we show how OF properties are being used toward the detection of these indoor obstacles.

Demonstrating the safety and performance of a velocity sourced series elastic actuator

Actuators with deliberately added compliant elements in the transmission system are often described as improving the safety of the actuator at the detriment of the performance. We show that our variant of the Series Elastic Actuator topology, the Velocity Sourced Series Elastic Actuator, has well defined performance characteristics that make for improvements in safety and performance over conventional high impedance actuators. The improvement in performance was principally achieved by having tight velocity control of the DC motor that acts as the mechanical power source for the actuator. Results for performance are given for point to point transition times, while results for safety are based on empirical assessment of the Head Injury Criterion during collisions.

Semi-analytic method of contact modelling

An algorithm to improve the accuracy and stability of rigid-body contact force calculation is presented. The algorithm uses a combination of analytic solutions and numerical methods to solve a spring-damper differential equation typical of a contact model. The solution method employs the recently proposed patch method, which especially suits the spring-damper differential equations. The resulting semi-analytic solution reduces the stiffness of the differential equations, while performing faster than conventional alternatives.

Extending the dynamic range of robotic vision

Conventional cameras have limited dynamic range, and as a result vision-based robots cannot effectively view an environment made up of both sunny outdoor areas and darker indoor areas. This paper presents an approach to extend the effective dynamic range of a camera, achieved by changing the exposure level of the camera in real-time to form a sequence of images which collectively cover a wide range of radiance. Individual control algorithms for each image have been developed to maximize the viewable area across the sequence. Spatial discrepancies between images, caused by the moving robot, are improved by a real-time image registration process. The sequence is then combined by merging color and contour information. By integrating these techniques it becomes possible to operate a vision-based robot in wide radiance range scenes.

Case-control study of early childhood caries in Australia

The aim of this case-control study of 617 children was to investigate early childhood caries (ECC) risk indicators in a non-fluoridated region in Australia. ECC cases were recruited from childcare facilities, public hospitals and private specialist clinics to source children from different socioeconomic backgrounds. Non-ECC controls were recruited from the same childcare facilities. A multinomial logistic modelling approach was used for statistical analysis. The results showed that a large percentage of children tested positive for Streptococcus mutans if their mothers also tested positive. A common risk indicator found in ECC children from childcare facilities and public hospitals was visible plaque (OR 4.1, 95% CI 1.0-15.9, and OR 8.7, 95% CI 2.3-32.9, respectively). Compared to ECC-free controls, the risk indicators specific to childcare cases were enamel hypoplasia (OR 4.2, 95% CI 1.0-18.3), difficulty in cleaning child's teeth (OR 6.6, 95% CI 2.2-19.8), presence of S. mutans (OR 4.8, 95% CI 0.7-32.6), sweetened drinks (OR 4.0, 95% CI 1.2-13.6) and maternal anxiety (OR 5.1, 95% CI 1.1-25.0). Risk indicators specific to public hospital cases were S. mutans presence in child (OR 7.7, 95% CI 1.3-44.6) or mother (OR 8.1, 95% CI 0.9-72.4), ethnicity (OR 5.6, 95% CI 1.4-22.1), and access of mother to pension or health care card (OR 20.5, 95% CI 3.5-119.9). By contrast, a history of chronic ear infections was found to be protective for ECC in childcare children (OR 0.28, 95% CI 0.09-0.82). The biological, socioeconomic and maternal risk indicators demonstrated in the present study can be employed in models of ECC that can be usefully applied for future longitudinal studies.

Disease prevention and control

Globally, the main contributors to morbidity and mortality are chronic diseases, including cardiovascular disease and diabetes. Chronic diseases are costly and partially avoidable, with around sixty percent of deaths and nearly fifty percent of the global disease burden attributable to these conditions. By 2020, chronic illnesses will likely be the leading cause of disability worldwide. Existing health care systems, both national and international, that focus on acute episodic health conditions, cannot address the worldwide transition to chronic illness; nor are they appropriate for the ongoing care and management of those already afflicted with chronic diseases. International and Australian strategic planning documents articulate similar elements to manage chronic disease; including the need for aligning sectoral policies for health, forming partnerships and engaging communities in decision-making. The Australian National Chronic Disease Strategy focuses on four core areas for managing chronic disease; prevention across the continuum, early detection and treatment, integrated and coordinated care, and self-management. Such a comprehensive approach incorporates the entire population continuum, from the ‘healthy’, to those with risk factors, through to people suffering from chronic conditions and their sequelae. This chapter examines comprehensive approach to the prevention, management and care of the population with non-communicable, chronic diseases and communicable diseases. It analyses models of care in the context of need, service delivery options and the potential to prevent or manage early intervention for chronic and communicable diseases. Approaches to chronic diseases require integrated approaches that incorporate interventions targeted at both individuals and populations, and emphasise the shared risk factors of different conditions. Communicable diseases are a common and significant contributor to ill health throughout the world. In many countries, this impact has been minimised by the combined efforts of preventative health measures and improved treatment of infectious diseases. However in underdeveloped nations, communicable diseases continue to contribute significantly to the burden of disease. The aim of this chapter is to outline the impact that chronic and communicable diseases have on the health of the community, the public health strategies that are used to reduce the burden of those diseases and the old and emerging risks to public health from infectious diseases.

Liking vs. wanting food : importance for human appetite control and weight regulation

Current train of thought in appetite research is favouring an interest in non-homeostatic or hedonic (reward) mechanisms in relation to overconsumption and energy balance. This tendency is supported by advances in neurobiology that precede the emergence of a new conceptual approach to reward where affect and motivation (liking and wanting) can be seen as the major force in guiding human eating behaviour. In this review, current progress in applying processes of liking and wanting to the study of human appetite are examined by discussing the following issues: How can these concepts be operationalised for use in human research to reflect the neural mechanisms by which they may be influenced? Do liking and wanting operate independently to produce functionally significant changes in behaviour? Can liking and wanting be truly experimentally separated or will an expression of one inevitably contain elements of the other? The review contains a re-examination of selected human appetite research before exploring more recent methodological approaches to the study of liking and wanting in appetite control. In addition, some theoretical developments are described in four diverse models that may enhance current understanding of the role of these processes in guiding ingestive behaviour. Finally, the implications of a dual process modulation of food reward for weight gain and obesity are discussed. The review concludes that processes of liking and wanting are likely to have independent roles in characterising susceptibility to weight gain. Further research into the dissociation of liking and wanting through implicit and explicit levels of processing would help to disclose the relative importance of these components of reward for appetite control and weight regulation.