Novel Dataset and Transformer for Driver Activity Recognition via IR-UWB Radar

This research explores driver activity recognition using a novel dataset and input-size-agnostic Vision Transformer, potentially improving in-cabin safety. The use of IR-UWB radar is particularly interesting, given its potential for robust performance in challenging lighting conditions.

• •Presents a new dataset for driver activity recognition using IR-UWB radar.
• •Employs an input-size-agnostic Vision Transformer architecture.
• •Aims to improve driver safety by accurately recognizing activities.