Optimization and Mathematical Modelling for Path Planning of Cooperative Urban Autonomous Vehicles

Automated vehicles have the potential to transform the way in which we travel and in which transport is provided. The technology is now sufficiently advanced that their deployment is rapidly approaching, with notable developments such as Tesla’s Autopilot (so-called level 2/3 automation) and the highly-automated (level 4) such as Volvo’s Drive Me. A great deal of attention has been paid on how they will operate and interacts on motorways, but much less effort has been devoted to urban areas, which are potentially much more complex. Urban areas do not always have a notion of well-defined lanes (lane-discipline), making it difficult to deploy current autonomous vehicle technology. In addition, traditional vehicles, parked vehicles, pedestrians and bicycles dynamically influence what would be the optimal trajectory/path for such an autonomous vehicle, both on roads and through intersections. This makes a typical car-following, lane-tracking autonomous control algorithm of less value in the urban context. The goal of this project is to develop an autonomous vehicle controller for groups of vehicles that can take advantage of the higher flow rates available on roads where lanes are not well defined. It will do so by calculating optimal trajectories for a group of autonomous vehicles in interaction with traditional vehicles, parked vehicles, pedestrians and bicycles. This will allow for deployment of autonomous vehicles in areas with irregular lane placement, large numbers of pedestrians, bicycles and parked cars, while taking advantage of platooning of autonomous vehicles both laterally and longitudinally.

The research for this PhD project will focus on the following technical objectives:
1. Develop and optimise a model, for traditional drivers that predicts the path the driver will take in a non-lane-discipline urban environment.
2. Develop a multi-trajectory optimization method for swarms of autonomous vehicles that route themselves around the predicted paths of traditional vehicles as well as other static obstacles.

Recent research in path prediction for shared control automated vehicles and advanced driver assistance systems (ADAS) can be leveraged, as can a wide variety of research in path planning for flying robots (UAVs).

The student will use mathematical algorithms, computer simulations, and control systems to achieve the technical objectives, combining transport micro-simulation methods with optimization techniques such as Mixed-integer Linear Programming (MILP).


How to Apply: You must submit an online PhD application by the deadline stated above. Details of how to apply can be found here: http://www.its.leeds.ac.uk/courses/phd/apply/. You must clearly state the name of your chosen project in the project details. You do not need to include a research proposal, but you do need to upload a ‘statement of motivation’. This should be 1-2 pages and should include information about why you feel you are well suited to the topic. This may refer to your academic background and any other relevant experience and could include an indication of how you would choose to interpret the project. Any enquiries about the application process can be sent to Deborah Goddard ([Email Address Removed])

Informal enquiries about the project can be sent to Professor Richard Romano ([Email Address Removed]).

Entry requirements: Applicants should have a first or upper second class degree (or equivalent) in an appropriate background such as engineering, mathematics or operational research. No prior knowledge of control systems or optimization is required, though obviously it would be an advantage, as would a relevant Masters qualification. A key quality is an enthusiastic and enquiring mind, and the desire to learn new theoretical tools and apply them to real-world problems. Further information about entry requirements can be found here: http://www.its.leeds.ac.uk/courses/phd/apply/