Gain-scheduled model predictive controller for vehicle-following trajectory generation for autonomous vehicles

Abstract

As the development of autonomous vehicles accelerates, the need to enhance the comfort characteristics for those vehicles has become important. In the present article, an enhanced vehicle-following motion planner algorithm is presented. The aim of the algorithm is to smoothen the repetitive braking and acceleration behaviour during vehicle following in traffic jam situations. The algorithm uses the information gathered from Lidar sensor, cameras and vehicle-embedded sensors in real time to construct the range vs. range-rate diagram, and it computes the desired velocity trajectory for the speed controller. The algorithm is based on the Gain-Scheduled Model Predictive Controller (MPC), where at least one MPC controller is designed to handle one of the three vehicle-following operating conditions: speed control, headway control and emergency brake control. The algorithm allows the designer to manipulate two vehicle following variables: standstill distance between lead vehicle and ego vehicle, and the headway time gap. The algorithm is experimentally validated on a full-size passenger vehicle.

KEYWORDS:

• Autonomous vehicle
• vehicle following
• range vs. range-rate
• adaptive cruise control
• gain-scheduled MPC

Acknowledgements

The authors would like to acknowledge the Canada Research Chairs programme, Canada Foundation for Innovation and Ontario Research Fund for funding the research. Also, the authors would like to acknowledge the Waterloo Region Emergency Services Training and Research Center (WRESTRC) for facilitating the field testing of the proposed research.

Disclosure statement

No potential conflict of interest was reported by the author(s).