Master's Thesis
A Predictive Safety Filter for a Formula Student Autonomous Racing Car
— 2022
Key information
Authors:
Supervisors:
Published in
November 11, 2022
Abstract
In this thesis, a Predictive Safety Filter (PSF) for a Formula Student (FS) autonomous vehicle has been developed, providing formal safety guarantees and responding to the growing needs for high performance controllers in autonomous racing. The safety filter was programmed in C++ and a commercially available solver was employed to solve the optimization problem. The PSF is able to keep the car inside the track boundaries when a potentially unsafe control input is given to the system. For this, a Model Predictive Control (MPC) optimization problem was designed, which evaluates the safety of the received control action. If the control input is considered unsafe, the PSF will slightly alter its trajectory to maintain the vehicle inside the track at all times. Thus, a highly non-linear vehicle model is employed, to obtain the best behaviour for lower and higher velocities during the race and, an invariant safe set obtained through semi-definite programming was designed to provide the safety certifications. Various tests were performed in three different FS Driverless tracks in a realistic simulation environment. We show that for an unsafe control input, where the vehicle disregards the track limits, the PSF can reduce hitting the track cones by 100%, although increasing the lap times by 13%.
Publication details
Authors in the community:
Ivo Barbosa Simões
ist187024
Supervisors of this institution:
Fields of Science and Technology (FOS)
electrical-engineering-electronic-engineering-information-engineering - Electrical engineering, electronic engineering, information engineering
Publication language (ISO code)
eng - English
Rights type:
Embargo lifted
Date available:
October 6, 2023
Institution name
Instituto Superior Técnico