MDH recently presented some of their simusafe research entitled Supervised Learning for Road Junctions Identification using IMU at the International Conference on Advances in Signal Processing and Artificial Intelligence (ASPAI’ 2019)’, http://www.aspai-conference.com/.
abstract: Events and interactions that take place in road junctions (i.e., roundabouts and intersections) are interesting to Naturalistic Driving Studies NDS analyzing safety and behavioral aspects of road users. This work employs machine learning techniques to develop a model that could automatically identify various road junctions encountered during ND using Inertial Measurement Unit IMU sensory data. The research is based on ND data of 16 volunteers part of SimuSafe project collected between May and August 2018. The supervised learning approach utilizes GPS signal to label road structures. The best performing model using Random Forest classifier achieved 90% precision on roundabout identification and recorded superior recall in comparison to using smart camera to spot roundabout signs.