On Supervisor Synthesis via Active Automata Learning
Doktorsavhandling, 2021

Our society's reliance on computer-controlled systems is rapidly growing. Such systems are found in various devices, ranging from simple light switches to safety-critical systems like autonomous vehicles. In the context of safety-critical systems, safety and correctness are of utmost importance. Faults and errors could have catastrophic consequences. Thus, there is a need for rigorous methodologies that help provide guarantees of safety and correctness. Supervisor synthesis, the concept of being able to mathematically synthesize a supervisor that ensures that the closed-loop system behaves in accordance with known requirements, can indeed help.

This thesis introduces supervisor learning, an approach to help automate the learning of supervisors in the absence of plant models. Traditionally, supervisor synthesis makes use of plant models and specification models to obtain a supervisor. Industrial adoption of this method is limited due to, among other things, the difficulty in obtaining usable plant models. Manually creating these plant models is an error-prone and time-consuming process. Thus, supervisor learning intends to improve the industrial adoption of supervisory control by automating the process of generating supervisors in the absence of plant models.

The idea here is to learn a supervisor for the system under learning (SUL) by active interaction and experimentation. To this end, we present two algorithms, SupL*, and MSL, that directly learn supervisors when provided with a simulator of the SUL and its corresponding specifications. SupL* is a language-based learner that learns one supervisor for the entire system. MSL, on the other hand, learns a modular supervisor, that is, several smaller supervisors, one for each specification. Additionally, a third algorithm, MPL, is introduced for learning a modular plant model.

The approach is realized in the tool MIDES and has been used to learn supervisors in a virtual manufacturing setting for the Machine Buffer Machine example, as well as learning a model of the Lateral State Manager, a sub-component of a self-driving car. These case studies show the feasibility and applicability of the proposed approach, in addition to helping identify future directions for research.

Supervisory control theory

Finite-state machines

Active learning

Model learning

Automata learning

Discrete-event systems

Online
Opponent: Professor Kai Cai, Osaka City University, Japan.

Författare

Ashfaq Hussain Farooqui

Chalmers, Elektroteknik, System- och reglerteknik

On Active Learning for Supervisor Synthesis

IEEE Transactions on Automation Science and Engineering,;Vol. In Press(2022)

Artikel i vetenskaplig tidskrift

Active Learning of Modular Plant Models

IFAC-PapersOnLine,;Vol. 53(2020)p. 296-302

Paper i proceeding

Modular Supervisory Synthesis for Unknown Plant Models Using Active Learning

IFAC-PapersOnLine,;Vol. 53(2020)p. 324-330

Paper i proceeding

Automatically Learning Formal Models from Autonomous Driving Software

Electronics (Switzerland),;Vol. 11(2022)

Artikel i vetenskaplig tidskrift

Ensuring the correctness of automated systems is crucial. The supervisory control theory proposes techniques to help build control solutions that provide certain correctness guarantees. These techniques rely on a model describing the behavior of the system. Unfortunately, such models are hard to create, thus limiting the industrial adoption of SCT. This thesis aims to improve the situation by providing an approach to automatically learn a model that captures the system's behavior.

To this end, we propose two approaches to integrate active learning and the supervisory control theory. Active learning is a promising technique to learn models by interacting with the system to be learned. Using active learning helps avoid the manual step of creating models, thus allowing the use of supervisory control techniques in the absence of models.

The presented approaches are implemented in a tool MIDES. Two case studies have been undertaken to understand the industrial challenges of the proposed approaches. In the first, the applicability in a manufacturing scenario is studied. In the second, a model of a software component in a self-driving car was learned. Both studies highlight the benefits of the proposed methods while also pointing out their limitations.

Automatiskt bedömning av autonoma fordons korrekthet (Auto-CAV)

VINNOVA (2017-05519), 2018-03-01 -- 2021-12-31.

Systematisk testning av cyberfysiska system (SyTeC)

Vetenskapsrådet (VR) (2016-06204), 2017-01-01 -- 2022-12-31.

Styrkeområden

Produktion

Ämneskategorier

Elektroteknik och elektronik

Reglerteknik

ISBN

978-91-7905-510-3

Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 4977

Utgivare

Chalmers

Online

Online

Opponent: Professor Kai Cai, Osaka City University, Japan.

Mer information

Senast uppdaterat

2023-11-13