Steven Beauchemin

photo of Dr Beauchemin

Associate Professor

Office: Middlesex College 28C
Tel:519-661-2111 ext. 82073



Dr. Steven Beauchemin is an Associate Professor of Computer Science at the University of Western Ontario, London, Canada. Dr. Beauchemin’s research interests revolve around sensing techniques for the automotive industry and in Advanced Driving Assistance Systems. Dr. Beauchemin leads the RoadLAB research and development program at The University of Western Ontario.

Research Interests

Computer Vision, Signal Processing, Applied Mathematics, Automotive Technologies, Driver Behaviour

Selected Publications

  • SHABBANI S., BEAUCHEMIN, S.S., and BAUER M.A., Analysis of Driver Gaze and Attention to Traffic Signs, submitted to Journal of Advanced Transportation, 2021.
  • SHIRPOUR M., KHAIRDOOST N., BEAUCHEMIN S.S., and BAUER M.A., Traffic-Related Object Detection and Recognition: A Survey and an Approach Based on The Attentional Visual Field of Drivers, submitted to IEEE Transactions on Intelligent Vehicles, 2020.
  • SHABANI S., BEAUCHEMIN S.S., and BAUER M.A., Analysis of Gaze and Driving Sequences to Determine Driver (In)Attention to Traffic Signs, submitted to
  • PEPIN A., BEAUCHEMIN S.S., LEGER S., and BEAUDOIN N., Algorithms to Obtain Low and High-Degree Splines from Discrete Fourier Transforms, submitted to ACM Transactions on Mathematical Software, 2020.
  • RAHMAN J., BEAUCHEMIN S.S., and BAUER M.A., Predicting Driver Behaviour at Intersections Based on Traffic Light Recognition, accepted in IET Intelligent Transport Systems, 2020.
  • KHAIRDOOST N., SHIRPOUR M., BEAUCHEMIN S.S., and BAUER M.A., Real-Time Driver Maneuver Prediction Using LSTM, IEEE Transactions on Intelligent Vehicles, Vol. 5, No. 4, pp. 714-724, 2020.
  • PEPIN A., BEAUCHEMIN S.S., LEGER S., and BEAUDOIN N., A New Method for High-Degree Spline Interpolation: Proof of Continuity for Piecewise Polynomials}, in Canadian Mathematical Bulletin, 1-15, doi:10.4153/S0008439519000742, Cambridge University Press, 2020.
  • ZARDOSHT M., BEAUCHEMIN S.S., and BAUER M.A., Identifying Driver Behaviour in Pre-Turning Maneuvers Using In-Vehicle CANbus Signals, Journal of Advanced Transportation}, Vol. 2018, Article ID 5020648, 10 pages, 2018.


Courses taught in 2020/21
  • CS1026 – Computing Fundamentals I

  • CS3388 – Computer Graphics

  • CS9645 – Introduction to Computer Vision Techniques


  • EXCELLENCE IN TEACHING AWARD 2015-16, UWO Student Council

  • NoAE INNOVATION AWARD 2010}, Network of Automotive Excellence





  • EXCELLENCE IN TEACHING AWARD 1994-95, UWO Student Council.

Research Projects

The RoadLAB Initiative: Predicting Driver Intent

Driving is an essential part of daily life for many and the most prevalent form of mobility in modern societies. With a continued rise in car usage globally, increasing traffic densities in mega cities and highways pose new and challenging threats to safety.


Incorporating the Driver

The advent of driving assistance systems to aid motorists in the driving process has contributed, in part, to improved safety on the road. Many of these systems, such as adaptive cruise control, collision warning, blind spot monitoring and park assist rely on the detection of relevant features in the immediate environment of the vehicle, such as other vehicles, pedestrians, lanes, traffic signs, and other potential obstacles. However, until recently, little research has focused on the monitoring of events and factors that directly concern the driver of the vehicle, despite the fact that 95 per cent of all accidents are caused by human error.

Beauchemin’s team at the University of Western Ontario are taking a novel approach to the design of the next generation of intelligent, Advanced Driving Assistance Systems (i-ADAS), by incorporating the driver as an inherent behavioural agent, with the aim of understanding and predicting his or her driving actions. In collaboration with Professor Michael Bauer, the RoadLab project aims to develop a deeper understanding of the cognitive (cephalo-ocular) task of driving, identifying risk-related factors and integrating these findings into predictive models of driver intent.

The long-term goals of the RoadLab programme include the identification of the cognitive factors involved in driving that impact traffic safety, as well as the definition of sound principles for the design of 'crash-less' automated vehicular safety technologies and the development of i-ADAS; putting driver behaviour prediction and correction at the heart of safety improvement

Eye Movements

In order to determine whether driver intent and driving-related actions can be predicted from qualitative and quantitative analyses of driver behaviour, Beauchemin and his team are establishing the correspondence between cephalo-ocular behaviour and visual stimuli. It has been repeatedly demonstrated that eye movements reflect processes aimed at locating the information required to generate actions in relation to the environment and on this basis, Beauchemin conjectures that eye movements are as predictive of driving actions as they are of physical movement.

Previously developed binocular systems, for estimating the absolute 3D coordinates of where one is looking in the 3D world, have had limited success. The RoadLab team’s creation therefore presents a novel method of determining which objects within a visual scene in front of a vehicle elicit visual responses from drivers. Their method combines a binocular eye gaze tracker with a binocular scene stereo system through an innovative cross calibration procedure, allowing the system to remain precise for significantly larger volumes and distances. The system operates in real time (30 Hz) and is installed in an operational, experimental vehicle. This experimental vehicle was the first of its kind, capable of computing the absolute 3D point of gaze of its driver sufficiently precisely to conduct scientific experiments addressing ocular behaviour in relation to visual stimuli; an important step toward understanding visual attention behaviour and possibly predicting imminent manoeuvres.

Improving Vehicular Safety

Keeping the driver as an active participant in the feedback mechanisms means that contextually-motivated informational support can be provided to offer immediate applications for enhancing safety. The extended possibilities of the RoadLab programme could have a major impact on reducing injuries caused by accidents and their social and economic implications the world over.