Model-predictive Trajectory Tracking for Autonomous Vehicles

Trajectory tracking for autonomous driving based on model predictive control (MPC). Author: Jan Filip, Czech Technical University in Prague

Control structure:

a) Longitudinal velocity MPC based on LTI model,  

b) Lateral MPC with LPV model based on a single-track vehicle model with linear tire cornering force characteristic,

c) The desired path created using spline curves and approximated piecewise-linearly during controller runtime for tracking error calculation.

d) Velocity trajectory generated using the forward-backward iterative algorithm and a simplified “single isotropic tire” model.

Longitudinal and lateral control of vehicle motion implemented in Simulink. Controller performance verified in simulation in a race track scenario using high-fidelity vehicle model in IPG Carmaker, with simulated OTS RT3002 state estimator. 

Simulator: IPG Carmaker 6.0.4

Location: Nardo Handling Track

Car model: Tesla Model S

QP Solver: qpOASES

Description of the graphical overlay:

* Left-hand side plots indicate:  

1) Velocity tracking  

2) Longitudinal acceleration  

3) Lateral acceleration

* Right-hand side plots indicate:  

1) Crosstrack error  

2) Heading error (orange with side slip compensation)  

3) Steering angle (orange = open-loop, blue = closed-loop)

This post was created with our nice and easy submission form. Create your post!

What do you think?

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Path Planning, self driving car

Simulation of pathfinding for self-driving vehicles (car and semi truck with trailer