Coordinating Teams of Multiple Humans and Autonomous Systems
- Reference number
- FFL24-0016
- Project leader
- Wohlrab, Rebekka
- Start and end dates
- 250801-300731
- Amount granted
- 15 000 000 SEK
- Administrative organization
- Chalmers University of Technology
- Research area
- Information, Communication and Systems Technology
Summary
My goal is to conceive solutions that enable teams of multiple humans and multiple autonomous systems to coordinate themselves, through the novel paradigm of adaptive mission coordination. Recently, a lot of research has focused on human-in-the-loop aspects of autonomous systems. What is often forgotten is that there is not just "the human in the loop", but multiple humans are involved, e.g., regulators, security experts, and end users. Currently, there are no principled approaches for supporting coordination between multiple humans and autonomous systems. I will address the following objectives: (A) develop methods to establish shared models, so that humans and systems can get a common understanding; (B) design adaptive solutions for coordinating multiple humans and multiple autonomous systems, e.g., by switching between centralized and decentralized, or explicit and implicit coordination; (C) conceive algorithms and methods for trade-off analysis of coordination mechanisms; (D) create methods for human-system teams to detect and respond to cybersecurity threats. To understand how humans reason about coordination, I will perform interviews, surveys, and user studies. I will design software architectures, algorithms, and methods. I will develop prototypes of my contributions. Towards the end, I aim to develop an integrated proof-of-concept implementation. I will design realistic scenarios for evaluation and validate the proof of concept with practitioners.
Popular science description
My goal is to conceive solutions that enable teams of multiple humans and multiple autonomous systems to coordinate themselves, through the novel paradigm of adaptive mission coordination. Adaptive mission coordination is needed, for example, when a team of robots should work in a warehouse next to a team of humans. In recent years, researchers and companies have understood that it is important to create autonomous systems that consider humans. What is often forgotten is that there is not just "the human", but different humans are involved, e.g., regulators, business owners, bystanders, security experts, and end users. In situations where multiple humans and multiple autonomous systems shall collaborate, it needs to be arranged how they should communicate. For example, if a robot faces an unknown object, it can be good if it asks a human what to do. On the other hand, when an emergency occurs, decisions need to be made quickly and it is better if a robot decides automatically what to do. These rules are captured in so-called "coordination mechanisms". I propose that coordination mechanisms should not be set in stone, but shall be adapted depending on the situation and the involved humans and systems. To understand how humans reason about coordination, I plan to perform interviews, surveys, and user studies. I will design and implement prototypes of systems. Towards the end of the project, I will create an integrated proof-of-concept implementation. To validate the proof of concept with practitioners, I will design concrete scenarios that meet the needs of the industry.