An MoD/DSTL funded project through the Autonomous Systems Underpinning Research (ASUR) programme
An increasing amount of autonomy is built into the computerised systems that pervade our lives and this is ASUR Proposal Template changing the way we perceive and interact with these systems. Humans are beginning to learn to devolve responsibility to trusted autonomous systems, even accepting instructions from such systems, ultimately collaborating to achieve otherwise unattainable goals. We are beginning to develop empathy for our autonomous resources because we are increasingly aware of their strengths, weaknesses, constraints and motivations. At the same time there is an opportunity to tailor, or personalise, these systems to better complement individuals. We draw inspiration from the complex relationships encountered between military working dogs and handlers in goal--oriented environments, where through extensive training both dog and handler become sensitised to behavioural cues and reciprocal goals, combining as individuals to create a highly effective operational team. The project investigates the concept of personalisation and aims to formalise what it means in the context of autonomous systems. We simulate a goal--oriented task in a virtualised gaming environment and measure the performance of a human working with an autonomous system to meet the goals set within the game. We capture the decision--making process by closely monitoring physical movements of the human participant and deconstruct observations to identify opportunities to modify the autonomous functions within the system that would result in an enhanced combined performance. The approach allows us to measure the impact of personalisation and examine how the degree and nature of personalisation changes across a range of human volunteers. The ultimate aim will be to deliver more effective and adaptable autonomous solutions that humans can intuitively work with in pursuit of new and challenging goals.
Personnel Involved (QUB)