< Back to previous page

Project

Embedded Learning and Optimization for Autonomous Vehicle Control

The aim of this PhD position is to develop advanced optimization and machine learning methods that are able to address challenging safety and comfort problems in autonomous driving. Methods such as imitation learning can introduce human likeness in the control behavior. However, they show limitations due to their lack of fundamental and rigorous results on explainability, safety and stability. Therefore, this PhD will work towards a novel methodological and embedded implementation development, which combines both learning human-like behaviour and optimal non-linear control (such as NMPC) for safety objectives to increase the comfort, but also the performance of the Autonomous Driving algorithm. One focus of the thesis, would be on the real-time capabilities of the controller, both in terms of having a deployable code, and having a numerically fast and reliable optimization solver.

Date:2 Jun 2021 →  Today
Keywords:Optimization, Autonomous Driving, Real-Time, Imitation learning, Control, Embedded, ADAS
Disciplines:Numerical computation, Automation and control systems, Embedded systems, Calculus of variations and optimal control, optimisation
Project type:PhD project