Research and development has been conducted over a period of one year in the following areas:

 

• Multimodal sensor fusion and situational awareness
• Auditory localisation
• Development of a software framework for system integration
• Development of custom safety devices for the industrial robot arm
• Development of evaluation paradigms based on experimental psychology methods for human-

  machine interaction

The system requires the ability to detect human presence in order for it to achieve a level of responsivity to the environment. Key research has been conducted into the use of stereo vision integrated with monaural face tracking and gesture recognition in detecting, in the first instance, the presence of human interlocutors, and then inferring their intent. Audio also provides useful hints about the environment. Therefore research and development work was also carried out in developing a robust auditory localisation system that could filter most ambient noise that is usually present in an artistic installation scenario. As the robot requires adherence to catagory-4 safety standards as stipulated by Australian Standard AS 4024.1-1502-2006 Safety of Machinery: Design of safety related parts of control systems-Validation, and is also required to maintain visibility and the feel of proximity to users (requirement of the artist), the engineering team were required to develop a novel safety system for the robot in consultation with industry experts. A custom safety fence and associated interlocks were developed to meet this requirement.

The complete system was then integrated through a novel event driven software architecture. Such a software system also essential to accommodate non-traditional methods of controlling the industrial robot arm.

Similar in some sense to traditional market research, parts of the overall system have been evaluated through user studies conducted at MARCS Labs. Based on psychological methods, users were tested on the 'likeability' of the Prosthetic Head