Title: Prediction-based hierarchical control framework for autonomous vehicles

Authors: Varun Jain; Thomas Weiskircher

Addresses: Department of Mechanical and Process Engineering, University of Kaiserslautern, D-67663 Kaiserslautern, Germany ' Applied Dynamics and Control Research Group, International Center for Automotive Research, Clemson University, Greenville-29607, SC, USA

Abstract: The predictive nature and the constraint handling capability of Model Predictive Controllers (MPC) makes it an appropriate choice for the conceptualisation of autonomous and collision avoidance systems. Such systems aim to make the road driving potentially safer and more comfortable in the future. This research work motivates development of a hierarchical structure based on a MPC and vehicle dynamics control for path planning and collision avoidance scenarios up to the limits of vehicle handling. The proposed idea not only helps to overcome the main challenge concerned with real-time implementation of MPC, but also adds modularity to the structure, whereby the tasks of path planning and vehicle handling can be tackled independently. The control structure can easily be extended for collision avoidance and driver assistance functions with the human driver in the loop. Simulation results with ideal and high fidelity vehicle models indicate the effectiveness of MPC and show the effect of different parameters on the overall performance.

Keywords: MPC; model predictive control; autonomous vehicles; vehicle control; vehicle dynamics; global chassis control; collision avoidance; hierarchical control; constraint handling; vehicle safety; path planning; modularity; driver assistance; simulation; vehicle modelling.

DOI: 10.1504/IJVAS.2014.067867

International Journal of Vehicle Autonomous Systems, 2014 Vol.12 No.4, pp.307 - 333

Received: 26 Mar 2014
Accepted: 21 Sep 2014
Published online: 07 Mar 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article