PhD Position in Computer Vision and Machine Learning for Autonomous Vehicles

Autonomous Vehicles at the Intersection of Computer Vision, Machine Learning and Human Factors

About the project

Institute for Transport Studies (ITS) is the leading transportation research centre in the UK and one of the Top 10 global leaders in the field. The centre currently has 60+ research staff and PhD students in multi-disciplinary areas of Human Factors, Computer Vision, AI and Machine Learning, Behavior Modeling, Psychology, and Transportation Safety.

Highly automated vehicles (AVs) also known as Level 4 vehicles can operate in self-driving mode in most circumstances and the driver is allowed to involve in any sort of non-driving activities, even sleeping. However, the automated driving mode is only expected to perform in limited spatial areas (geofenced) and under specific legislations; therefore, a human driver still has the option of getting ready to resume the driving control, whenever required.

This project aims at promoting public trust in automated driving based on scientific investigations and findings on the role of Computer Vision Technology and Human Factors in a smooth technology adaptation and transition. Using machine learning, computer vision, and human factors, the research should investigate whether the driver is ready to safely and swiftly take over the control of the vehicle? What are the major contributing human factors in taking over scenarios and to what extent? Under what circumstances the sensors and autonomous mode can perform safer than a human driver and vice-versa?

The research may involve in-cabin situation awareness such as driver behaviour monitoring (eye, head pose, and body pose modelling), as well as road hazard perception and traffic condition understanding (vehicles, pedestrians and cyclists’ detection & tracking).

While in this research we put more weight in Computer Vision part of the research, the PhD candidate is encouraged to perform joint research within our great interdisciplinary team of psychologists and human factor experts in ITS.

The PhD student is expected to contribute towards one of the following research pathways with a clear and feasible research plan and proposal for a 3.5-year PhD journey.

Possible Research Plans/Directions:

• Development of appropriate methodologies, models, tools, and frameworks to increase automation levels of driving and driver comfort using real-world visual datasets and human factors.
• Using Leeds University Driving Simulator facilities/datasets for eye tracking, driver awareness monitoring, situation awareness monitoring and driver-pedestrian interactions.
• Decision making and risk assessment for switching between human and the automated driving mode, based on real-world visual datasets and sensory information in different driving scenarios.
• Driver behaviour monitoring using visual and psychological measures based on eye and head pose tracking, body pose estimation, and activity recognition.
• In-depth understanding of sensor technologies including Vision, RGB-D, Lidar, Radar, and multi-sensor data fusion research for traffic monitoring and hazard perception.

Entry requirements:

• A Masters Degree in a relevant subject e.g. Computer Science, Engineering, AI, Machine Learning.
• Solid background in applied Computer Vision and Machine Learning
• Demonstrated competency in Math and Statistics
• Demonstrated Programming proficiency (Python and/or C)
• Minimum IELTS Score of 6.5 (no band less than 6.0)

Desired competencies:

• Ability to teamwork in a multidisciplinary research group
• Willingness and plans for joint research in human factors, safety, public trust, and social impact of the research.
• Prior publications in the field

How to Apply:

Enquiries may be addressed to Dr Mahdi Rezaei m.rezaei@leeds.ac.uk

Please provide with your CV, Letter of Motivation, and your PhD proposal (no more than 3 pages).

Further details on application procedures:  https://www.leeds.ac.uk/info/130206/applying/91/applying_for_research_degrees

Country Eligibility:

UK/EU/International