Abstract

This thesis is based on vision of the future where robotics and industry are centred around humans, emphasising collaboration between humans and robots rather than mere automation.

In this collaborative future, robots serve as active assistants, coexisting closely with humans and engaging in physical interactions to execute tasks. Such symbiotic systems leverage the unique abilities of both humans and robots, enhancing efficiency and prioritising human safety and well-being through personalised robotic assistance.

Realising this vision requires being able to quantify the diverse abilities of humans and robots in a unified view, as well as their joint abilities when collaborating. Additionally, it requires being able to measure the assistance required by operators in order to ensure their safety and well-being. Therefore, this thesis advocates for the use of physical ability metrics, particularly their polytope representations, to address these questions. This thesis proposes a structured view on common physical ability polytopes for humans and robots along with an overview of their applicable evaluation methods. The thesis then explores the use of polytopes as a source of real-time information about human’s and robot’s changing physical abilities and their potential to enhance different aspects of human-robot collaboration.

Finally, this thesis presents an open-source package called ‘pycapacity’, an efficient and user-friendly framework for calculating physical ability metrics of humans and robots, opening doors to their use in the wider community.