Inverse Model Predictive Control (IMPC) Based Modeling and Prediction of Human-Driven Vehicles in Mixed Traffic
Modeling and predicting human-driven vehicle behaviors are critically important for the planning and control of autonomous vehicles in mixed traffic which includes both autonomous and human-driven vehicles. Despite the tremendous efforts on this problem, the traditional general driver model-based approaches are subject to prediction accuracy issues and the state-of-the-art data-driven heuristic approaches are subject to scalability issues. To this end, this paper proposes a novel inverse model predictive control (IMPC) based approach to address these issues. The approach can learn the internal control process of human drivers through the automatic learning of their cost functions in a novel IMPC setup, which could result in improvements in both the accuracy and scalability. The approach was implemented and validated with realistic human driver studies. The experiments illustrated that the proposed approach could achieve a better accuracy and a better scalability for unseen scenarios compared to existing approaches.
- Record URL:
-
Availability:
- Find a library where document is available. Order URL: http://worldcat.org/issn/23798858
-
Supplemental Notes:
- Copyright © 2021, IEEE.
-
Authors:
- Guo, Longxiang
- Jia, Yunyi
- Publication Date: 2021-9
Language
- English
Media Info
- Media Type: Web
- Features: References;
- Pagination: pp 501-512
-
Serial:
- IEEE Transactions on Intelligent Vehicles
- Volume: 6
- Issue Number: 3
- Publisher: Institute of Electrical and Electronics Engineers (IEEE)
- ISSN: 2379-8858
- Serial URL: http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=7274857
Subject/Index Terms
- TRT Terms: Autonomous vehicles; Drivers; Human beings; Motor vehicles; Predictive models; Process control; Vehicle mix
- Subject Areas: Highways; Operations and Traffic Management; Vehicles and Equipment;
Filing Info
- Accession Number: 01785365
- Record Type: Publication
- Files: TRIS
- Created Date: Oct 22 2021 5:16PM