Model-free Learning for Safety-critical Control Systems: A Reference Governor Approach

Liu, Kaiwen; Li, Nan; Kolmanovsky, Ilya; Rizzo, Denise; Girard, Anouck

doi:10.23919/acc45564.2020.9147255

Cited by 9 publications

(9 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, for more complex scenarios including those involving multiple vehicle interactions, an explicit model is typically not available. Therefore, an emerging approach is to train a safety supervisor through learning, as investigated in References 26,[31][32][33] Note that the learning objectives of those aforementioned learning-based approaches to control policy design and a learning-based approach to safety supervisor design are different: The former aims at developing a nominal control policy that achieves optimal performance with regard to (the expected value of) a reward function, while the latter focuses on the safety aspect and typically pursues minimum modification to the nominal control.…”

Section: Introductionmentioning

confidence: 99%

Improving autonomous vehicle in‐traffic safety using learning‐based action governor

Han

Tseng

et al. 2022

Adv Control Appl

Self Cite

View full text Add to dashboard Cite

The Action Governor (AG) is a supervisory scheme augmenting a nominal control system in order to enhance the system's safety and performance. It acts as an action filter, monitoring the action commands generated by the nominal control policy and adjusting the ones that might lead to undesirable system behavior. In this article, we present an approach based on learning to developing an AG for autonomous vehicle (AV) decision policies to improve their safety for operating in mixed‐autonomy traffic (i.e., traffic involving both AVs and human‐operated vehicles (HVs)). To achieve this, we demonstrate that it is possible to train the AG in a traffic simulator that is capable of representing in‐traffic interactions among AVs and HVs. We illustrate the effectiveness of this learning‐based AG approach to improving AV in‐traffic safety through simulation case studies.

show abstract

Section: Introductionmentioning

confidence: 99%

Improving autonomous vehicle in‐traffic safety using learning‐based action governor

Han

Tseng

et al. 2022

Adv Control Appl

Self Cite

View full text Add to dashboard Cite

show abstract

“…Simulation results are reported which illustrate learning and vehicle response during step commands, sine-and-dwell tests and when driving conditions change. This paper is distinguished from [28] by providing more details of LRG algorithms, extending the theoretical analysis of LRG, providing detailed proofs, and illustrating a practical application to the fuel truck rollover avoidance. This paper is also different from [29], where the latter focuses on an integration of a neural network into LRG to handle variability in road conditions and speed up online computations; however, for such an approach based on a neural network, theoretical constraint enforcement guarantees are no longer available.…”

Section: Introductionmentioning

confidence: 99%

Safe Learning Reference Governor: Theory and Application to Fuel Truck Rollover Avoidance

Liu¹,

Li²,

Kolmanovsky³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

This paper proposes a learning reference governor (LRG) approach to enforce state and control constraints in systems for which an accurate model is unavailable; and this approach enables the reference governor to gradually improve command tracking performance through learning while enforcing the constraints during learning and after learning is completed. The learning can be performed either on a black-box type model of the system or directly on the hardware. After introducing the LRG algorithm and outlining its theoretical properties, this paper investigates LRG application to fuel truck rollover avoidance. Through simulations based on a fuel truck model that accounts for liquid fuel sloshing effects, we show that the proposed LRG can effectively protect fuel trucks from rollover accidents under various operating conditions.

show abstract

“…The present paper considers the application of a recently proposed Learning Reference Governor (LRG) [24][25][26][27] to performing ARPOD maneuvers. The LRG is an add-on scheme to a nominal control system and is used to enforce pointwise-in-time state and control constraints.…”

mentioning

confidence: 99%

“…The LRG is an add-on scheme to a nominal control system and is used to enforce pointwise-in-time state and control constraints. Its viability has been previously demonstrated in applications to vehicle rollover avoidance [24][25][26] and misfire avoidance in spark-ignition engines [27]. In our application to spacecraft control, the LRG monitors and modifies the command generated by a higher-level ARPOD planning algorithm when it becomes necessary to avoid constraint violations.…”

mentioning

confidence: 99%

“…The learning can be performed through experimentation with an actual spacecraft before the mission or through simulations on a high-fidelity, black-box model of the spacecraft (e.g., its digital twin). The safety-critical version of the LRG [25] guarantees constraint enforcement both during and after learning, and is adopted in this work. A notable feature of our safety-critical LRG approach is that learning can be terminated at any time and the resulting operation of the system is guaranteed to be safe (of course, conservative if the learning is interrupted too early).…”

mentioning

confidence: 99%

See 1 more Smart Citation

Learning Reference Governor for Constrained Spacecraft Rendezvous and Proximity Maneuvering

Ikeya

Liu

Girard

et al. 2023

Journal of Spacecraft and Rockets

View full text Add to dashboard Cite

Spacecraft automated rendezvous, proximity maneuvering, and docking (ARPOD) play significant roles in many space missions, including on-orbit servicing and active debris removal. Precise modeling and prediction of spacecraft dynamics can be challenging due to the uncertainties and perturbation forces in the spacecraft operating environment and due to the multilayered structure of its nominal control system. Despite this complication, spacecraft maneuvers need to satisfy required constraints (thrust limits, line-of-sight cone constraints, relative velocity of approach constraints, etc.) to ensure safety and achieve ARPOD objectives. This paper considers an application of a learning-based reference governor (LRG) to spacecraft ARPOD operations to enforce constraints without relying on a dynamic model of the spacecraft during the mission. Similar to the conventional reference governor (RG), the LRG is an add-on supervisor to a closed-loop control system, serving as a prefilter on the command generated by the ARPOD planner. The LRG modifies, if it becomes necessary, the reference command to a constraint-admissible value to enforce specified constraints. The LRG is distinguished, however, by the ability to rely on learning instead of an explicit model of the system; and it guarantees constraints’ satisfaction during and after the learning. In this paper, the LRG is applied to the control of combined translational and rotational motions of a chaser spacecraft, and three case studies with different sets of safety constraints and thruster assumptions are used to demonstrate the benefits of the LRG in ARPOD missions.

show abstract

Model-free Learning for Safety-critical Control Systems: A Reference Governor Approach

Cited by 9 publications

References 9 publications

Improving autonomous vehicle in‐traffic safety using learning‐based action governor

Improving autonomous vehicle in‐traffic safety using learning‐based action governor

Safe Learning Reference Governor: Theory and Application to Fuel Truck Rollover Avoidance

Learning Reference Governor for Constrained Spacecraft Rendezvous and Proximity Maneuvering

Contact Info

Product

Resources

About