Time-optimal Race Car Driving using an Online Exact Hessian based Nonlinear MPC Algorithm

Paper in proceedings, 2017

This work presents an embedded nonlinear model predictive control (NMPC) strategy for autonomous vehicles under a minimum time objective. The time-optimal control problem is stated in a path-parametric formulation such that existing reliable numerical methods for real-time nonlinear MPC can be used. Building on previous work on timeoptimal driving, we present an approach based on a sequential quadratic programming type algorithm with online propagation of second order derivatives. As an illustration of our method, we provide closed-loop simulation results based on a vehicle model identified for small-scale electric race cars.