How robots are helping scientists build the big picture across a team

What is the optimal way to generate and share contextual observations among team members? That's one of the interesting questions being explored by a DSTG-University of Adelaide team led by defence scientist Dr Andrew Walters.

The team conducts experiments by setting loose a team of autonomous land vehicles in an engineering laboratory "sandpit". In the most recent experiments the researchers have focused on exploring improvements in the way the robots generate and share contextual and actionable information amongst themselves to create a common operating picture and establish advanced situational awareness.

Dr Walters reports that the robots successfully demonstrated progress in efficiently creating a more-powerful 'hive mind'.

'As the robot scouts traverse the simulated environment they build up knowledge that is usable across the team,' says Dr Walters. 'The point of this research is to see how we can build up knowledge, including assessment and representation of the incoming information, and then execute actions based on that new knowledge on-board the robots.'

'Much like human expeditionary teams, the robots are given some pre-mission intelligence about the environment. In the military sense, they are loaded with intelligence mission data - a digital description of what is expected in that environment. They can then compare what they're sensing in real-time against what they were given, to determine new or updated information, and make inferences. In essence, this makes the traditionally static mission data, dynamic! This facilitates adaptability and agility against the evolving threat.'

'For our experiments, objects in the environment also had QR codes that were scanned by the scouts to gather extra descriptors of the objects. That's because our focus is on the knowledge pipeline, not the data gathering aspect. We're testing the ability of the system to assess all these bits of information and connect them in a way that builds relationships between the objects, their descriptors and the environment so that the decisions that the robots make are much more informed.'

'The relationship between objects is key; our information assessment algorithms that we've deployed in our knowledge pipeline are really being tested to see how the system can quickly make that assessment, and whether the scout team can infer higher levels of understanding about the environment.'

What is too much information?

Determining the optimal way to share mission intelligence across the scout team in this lab scenario is of vital importance to Defence. Efficient and effective information sharing across Defence’s different tactical data networks could make all the difference in otherwise chaotic situations. The results of this DSTG research, in terms of understanding and characterising the sharing of information, is key to meeting the challenge.

‘It is an interesting dilemma,’ says Dr Walters. ‘If I wrote facts on a piece of paper and I gave it to you, and you read it and then retell it to the next person in a chain, the way you retell it may be based on your prior learning, your prior understanding, those type of things.

‘However, if I give you the piece of paper, you absorb the information, then you just pass the piece of paper to the next in the chain, they'll get it unchanged. What we're trying to answer, using these robot scouts is “What is the optimal point to share the new knowledge, and in what form?”’

The DSTG team will take revelations from this unclassified proof of concept system and trial it in a Defence scenario by embedding it in lab representations of ADF aircraft systems. The next step is then to develop prototypes systems that can be trialled with Defence personnel.

Andrew praises the joint DSTG-University of Adelaide team and the benefits of being located on campus in the engineering faculty, noting that the opportunity to define projects that are helping to support real-world requirements is valuable for both the engineering students and Dr Walters’ team.

DSTG enhances its science and technology impact by collaborating with research and industry partners, nationally and globally.

We access and leverage world-leading science, technology, knowledge and innovation, through collaboration with industry, academia and international agencies. Learn more about our partnerships.