A Socially Assistive Robot using Automated Planning in a Paediatric Clinical Setting

Abstract

Social robotics has recently focused on developing AI agents that recognise and respond to human emotions. The use of plan-based approaches is promising, especially in domains where collecting data in advance is challenging (e.g., medical domains). We have designed and developed a system that implements a social robot to help children cope with painful and distressing medical procedures in a clinical setting. Our approach uses automated planning as a core component for action selection in order to generate plans that include physical, sensory, and social actions for the robot to use when interacting with humans. A key capability of our system is that the robot’s behaviour adapts based on the affective state of the child patient. The robot must operate in a challenging physical and social environment where appropriate and safe interaction with children, parents/caregivers, and healthcare professionals is crucial. In this paper, we present our system, examine some of the key challenges of the scenario, and describe how they are addressed by our system. We provide an overview of some of the lab-based trials that we have conducted. The system is currently undergoing usability studies at two hospitals, with the intention of moving onto clinical trials.