Machine vision can provide vehicle motion and environmental information without relying on expensive, data-intensive sensors. The visual-odometry work targets robust camera-based positioning with uncertainty-aware drift correction. Road-condition monitoring extends the sensing question to the surface on which machines operate.
Position estimates can become productivity indicators—cycle time, harsh braking and slow corners—as well as inputs to future autonomous systems.
Big ideas, a little mystery. We respect our students’ privacy while sharing their research. Until they choose to introduce themselves, we use illustrated avatars and purely fictional names, and keep their real names, faces and contact details private.