Alerting Misbehaviour for Autonomous Driving Systems
Computer Vision
Investigating methods to predict unsupported scenarios in autonomous driving systems trained using imitation learning. Using unsupervised models like autoencoders to enable online healing of such systems to prevent accidents.