A Caltech Library Service

Enhancing the Performance of a Safe Controller Via Supervised Learning for Truck Lateral Control

Chen, Yuxiao and Hereid, Ayonga and Peng, Huei and Grizzle, Jessy (2019) Enhancing the Performance of a Safe Controller Via Supervised Learning for Truck Lateral Control. Journal of Dynamic Systems, Measurement, and Control, 141 (10). Art. No. 101005. ISSN 0022-0434. doi:10.1115/1.4043487.

[img] PDF - Published Version
See Usage Policy.


Use this Persistent URL to link to this item:


Correct-by-construction techniques, such as control barrier functions (CBFs), can be used to guarantee closed-loop safety by acting as a supervisor of an existing legacy controller. However, supervisory-control intervention typically compromises the performance of the closed-loop system. On the other hand, machine learning has been used to synthesize controllers that inherit good properties from a training dataset, though safety is typically not guaranteed due to the difficulty of analyzing the associated learning structure. In this paper, supervised learning is combined with CBFs to synthesize controllers that enjoy good performance with provable safety. A training set is generated by trajectory optimization that incorporates the CBF constraint for an interesting range of initial conditions of the truck model. A control policy is obtained via supervised learning that maps a feature representing the initial conditions to a parameterized desired trajectory. The learning-based controller is used as the performance controller and a CBF-based supervisory controller guarantees safety. A case study of lane keeping (LK) for articulated trucks shows that the controller trained by supervised learning inherits the good performance of the training set and rarely requires intervention by the CBF supervisor.

Item Type:Article
Related URLs:
URLURL TypeDescription
Chen, Yuxiao0000-0001-5276-7156
Hereid, Ayonga0000-0002-4156-2013
Grizzle, Jessy0000-0001-7586-0142
Additional Information:© 2019 by ASME. Manuscript received August 23, 2018; final manuscript received April 3, 2019; published online June 3, 2019. The work of Yuxiao Chen, A. Hereid, and Huei Peng is supported by NSF Grant CNS-1239037. The work of J. Grizzle is supported by Toyota Research Institute (TRI). Funding Data: National Science Foundation (Grant No. CNS-1239037; Funder ID: 10.13039/501100008982). Toyota Research Institute (TRI) (Funder ID: 10.13039/501100004405).
Funding AgencyGrant Number
Toyota Research InstituteUNSPECIFIED
Subject Keywords:control barrier function, supervised learning, trajectory optimization
Issue or Number:10
Record Number:CaltechAUTHORS:20190926-133052511
Persistent URL:
Official Citation:Chen, Y., Hereid, A., Peng, H., and Grizzle, J. (June 3, 2019). "Enhancing the Performance of a Safe Controller Via Supervised Learning for Truck Lateral Control." ASME. J. Dyn. Sys., Meas., Control. October 2019; 141(10): 101005.
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:98879
Deposited By: Tony Diaz
Deposited On:26 Sep 2019 21:05
Last Modified:16 Nov 2021 17:42

Repository Staff Only: item control page