LAUSR

The ACANTO project is developing robotic assistants to aid the confidence and recovery of older adults. A key requirement of these assistants is aiding with navigation in complex and potentially chaotic environments. Prior work has addressed this for a single user, using a single robotic assistant in an intelligent environment. However, for therapeutic purposes, ACANTO supports social groups and group activities. ACANTO’s robotic assistants must, therefore, be able to plan the motion of groups of older adults walking together. This requires an efficient navigation solution that can handle large numbers of users and that can operate rapidly on embedded computing devices. To increase user confidence, the solution must encourage group cohesion without trying to impose its own rigid structure; it must try to maintain the natural (de facto) group structure despite unpredictable behaviours and environmental conditions. Our on-the-fly group motion planner addresses these challenges by: using intelligent environment information to develop behavioural traces, clustering traces to determine groups, constructing a predictive model of the groups as a whole, and finding an optimal suggested trajectory using statistical model checking. In this work, we describe our proposed approach in detail and validate some of its novel aspects on the ETH Zürich pedestrian motion dataset.