Charles Hartsell, Nagabhushan Mahadevan, Shreyas Ramakrishna, Abhishek Dubey, Theodore Bapty, Taylor T. Johnson, Xenofon D. Koutsoukos, Janos Sztipanovits, Gabor Karsai
Proceedings of the Workshop on Design Automation for CPS and IoT, DESTION@CPSIoTWeek 2019, Montreal, QC, Canada 2019
Learning-enabled CPS face unique challenges in demonstrating safety and correctness properties during design phases before deployment. This work is innovative because it integrates model-based design with learning component development, providing systematic approaches for architectural modeling, LEC training, and safety verification. The methodology enables developers to systematically address challenges in mixing formal assurance with empirical machine learning.
This paper presents a model-based design methodology for assurance-based learning-enabled cyber-physical systems that supports architectural modeling, LEC training, and safety assurance. The approach uses Domain Specific Modeling Languages to specify system architectures and integrates them with learning-enabled component development. The methodology includes support for multiple development workflows including supervised and reinforcement learning approaches.
The methodology successfully supported end-to-end development of learning-enabled systems from architectural specification through training and deployment. The framework demonstrated integration of multiple assurance techniques including formal verification, static analysis, and runtime monitoring. The approach proved effective at documenting design artifacts that support safety assurance arguments for complex systems.
@inproceedings{Hartsell2019,
author = {Hartsell, Charles and Mahadevan, Nagabhushan and Ramakrishna, Shreyas and Dubey, Abhishek and Bapty, Theodore and Johnson, Taylor T. and Koutsoukos, Xenofon D. and Sztipanovits, Janos and Karsai, Gabor},
booktitle = {Proceedings of the Workshop on Design Automation for {CPS} and IoT, DESTION@CPSIoTWeek 2019, Montreal, QC, Canada},
title = {Model-based design for {CPS} with learning-enabled components},
year = {2019},
month = {apr},
pages = {1--9},
abstract = {Recent advances in machine learning led to the appearance of Learning-Enabled Components (LECs) in Cyber-Physical Systems. LECs are being evaluated and used for various, complex functions including perception and control. However, very little tool support is available for design automation in such systems. This paper introduces an integrated toolchain that supports the architectural modeling of CPS with LECs, but also has extensive support for the engineering and integration of LECs, including support for training data collection, LEC training, LEC evaluation and verification, and system software deployment. Additionally, the toolsuite supports the modeling and analysis of safety cases - a critical part of the engineering process for mission and safety critical systems.},
bibsource = {dblp computer science bibliography, https://dblp.org},
biburl = {https://dblp.org/rec/bib/conf/cpsweek/HartsellMRDBJKS19},
category = {workshop},
contribution = {colab},
doi = {10.1145/3313151.3313166},
file = {:Hartsell2019-Model-based_design_for_CPS_with_learning-enabled_components.pdf:PDF},
keywords = {model-based design, learning-enabled components, safety assurance, domain-specific modeling, architectural design, CPS development},
project = {cps-autonomy},
tag = {ai4cps},
timestamp = {Wed, 20 Nov 2019 00:00:00 +0100},
url = {https://doi.org/10.1145/3313151.3313166},
month_numeric = {4}
}