Enhancing Safety using AR Headsets with Motion Prediction Visualization

Pedestrian collisions are a serious safety issue when considering hazards to elderly persons and to distracted people due to AR gaming. The inability to anticipate an oncoming danger could put these pedestrians at a safety risk endangering their lives. However if we use an AR device to visualise future motion, then this could firstly help elderly wearing them to forsee any oncoming danger and secondly alert a distracted AR gamer if this information is included in the game. The RGBD sensors of an AR device continuously scans the environment to place virtual content. Our prototype idea focuses on applying motion detection and prediction algorithms to this scan data, and communicating it as information to the users. Then other traffic participants, who might potentially get in conflict with an AR headset user can be predicted in advance and informed. This can not only promote the use of AR headsets as a safety aid especially for elderly and disabled but also increase safety in outdoor spaces when using AR. In this paper we share our first results for our prototype based on the above idea where future pedestrian motions are detected and visualised with Unity on the Hololens 2.