Natural Environment Research Demonstrator

Objectives

The development of autonomous systems operating in large scale outdoor environments opens a broad range of research challenges in sensing, describing and negotiating complex dynamic terrains. The potential application of such systems ranges from automated mining, through forest and ecology management, to bush fire fighting and defence. The combination of innovative, world-leading, basic research and the large number of potential applications provides a compelling and exciting case for the development of outdoor autonomous systems in Australia.

The natural environment research demonstrator will be based at the ACFR field test facilities near Marulan NSW. This facility has already been established and is used extensively in current research projects. It provides 5,000 hectares of varied farming, forest and river-side terrain, has an established runway for aircraft, fully equipped hangers, computing, communications, GPS base station and living accommodation for up to ten researchers. The demonstrator will make use of a number of existing platforms including a fleet of five unmanned air vehicles (UAVs), two amphibious ground vehicles and a distributed electro-optic and radar ground sensor network.

Outcomes

The primary demonstrations will build pictures of terrain, targets and features of interest in real time from a network of intelligent autonomous air and ground platforms. The demonstrations will exercise fundamental research in environment modelling, uncertainty management, cooperative control of multiple platforms, and modelling of complex systems of systems. The key milestones for the demonstration programme are:

• Year 1: Acquisition of a broad range of sensor data from land and air platforms, from radar/laser range sensors and EO imaging sensors. Modelling analysis and interpretation of data. Multi-platform systems modelling, architecture capability definition. Definition of experimental scenarios in collaboration with industry partners.
• Year 2: Demonstration of single and multiple platform data fusion, building static environment and terrain pictures, identifying, learning and tracking features, modelling information acquisition. Demonstration of air and land platform control in unstructured environments, learning traversability and motion control strategies.
• Year 3: Demonstrations of groups of ground vehicles and groups of air vehicles building dynamic terrain pictures in real time. Evaluation of perception modelling and uncertainty management ideas. Demonstration of individual platform control methods and algorithms. Implementation of systems design ideas. Development of tools for visualisation of interacting autonomous systems. Focus on applications in defence and in mining.
• Year 4: Development and demonstration of data fusion and coordinated control capabilities for different combinations of sensors and platforms. Particular evaluation of information modelling, information flow and the dynamics of networks of autonomous systems in the field. Integration of human operators and related knowledge sources in system design and operation. Testing and evaluation of systems modelling and design software.
• Year 5: Demonstrations of combined ground and air sensors and platforms, especially management of environment model types, communication and information sharing between systems. Capabilities demonstrated in cooperative natural environment localisation and mapping, and in coordinated information gathering by groups of platforms. Focus on applications involving coordinated use of air and ground systems, especially defence, environment monitoring, search and rescue.

The natural environment research demonstrator benefits from the already leading international position of the partners in land and air field robotics. The proposed milestones describe an innovative research programme demonstrating new methods, algorithms and systems substantially in advance of anything yet contemplated in the field of autonomous systems.