Project PI : Foong Shaohui
Task PI : Soh Gim Song, Tan U-Xuan, Yuen Chau
Project Period : 20 January 2020 to 27 July 2022
UAVs are increasingly used in search-and-locate operations by emergency service personnel. However, their operation is restricted to outdoor environments as they depend on GNSS for localisation and their limited on-board autonomy capabilities prevent them from doing unsupervised and safe navigation in confined and cluttered environments indoors. In addition, walls in buildings attenuate RF signals, degrading wireless communications between UAVs and operators. As a result, it is not feasible for emergency service personnel to deploy drones indoors for search-and-locate operations.
To enable indoor deployment of drones, Project BOREALIS will develop technologies to enable manned-unmanned teaming between a human and three UAVs (1) in close proximity, (2) in GNSS-denied and cluttered indoor environments, and (3) in widely varying lighting conditions from sunlight to low-light. With clutter, the momentary loss of visual line-of-sight and wireless communications between the human and the UAVs can cause the UAVs to lose track of the human. To effectively follow a human continuously under such scenarios, the UAVs in the team have to stay close to the human while avoiding obstacles. The UAVs also need to find collision-free paths to accomplish the task collaboratively without impeding one another.