Beyond physics: Tackling the limitations of camera-based perception

As autonomous vehicles and highly automated driver assistance systems take to the roads, one thing has become clear: automated systems need to better understand people. Industry accepted models, based on the physics of the road users to forecast their future position, don’t go far enough to protect and understand pedestrians, cyclists, and other vulnerable road users (VRUs) when inferring actions such as “crossing in front of the vehicle”. This approach doesn’t do a thorough enough job to understand the psychology behind human action or inaction; human behaviour is much more  complex, leading to false positives or failure to predict intention altogether. In this session, Dominic Noy will present the limitations of the industry’s preferred physics-based approach to human intent prediction and how probabilistic machine learning and behavioural psychology can help out.


Dr. Dominic Noy (Principal Behavioural Data Scientist)combines statistics, cognitive science, and machine learning to predict the behaviour of vulnerable road users. He is particularly keen about how to boost the interpretability of the underlying predictive models by using Bayesian methods. He holds a PhD in Experimental Psychology (Psychophysics, Cognitive Modelling, Human-Computer Interaction), a Master’s in Statistics, and a Master’s in Psychology (Psychology of Action), and has been working at Humanising Autonomy since 2018..

 Tuesday 1st Feb | 1pm | Online | Join Here: