A Comparative Review of Artificial Intelligence for the Control of Physical Systems: Case Studies in Power-Grid Dispatch and Autonomous Driving

Guohao Yin

doi:10.47297/taposatWSP2633-456928.20260708

PDF
Export
Share
Collection
Album

A Comparative Review of Artificial Intelligence for the Control of Physical Systems: Case Studies in Power-Grid Dispatch and Autonomous Driving

Volume 7, Issue 8

Views: 0 Downloads: 1

A Comparative Review of Artificial Intelligence for the Control of Physical Systems: Case Studies in Power-Grid Dispatch and Autonomous Driving
A Comparative Review of Artificial Intelligence for the Control of Physical Systems: Case Studies in Power-Grid Dispatch and Autonomous Driving
Hefei No.8 High School, Hefei, Anhui, 230071, China
Vol. 7, Issue 8, Pages: 142-147(2026)
Published：30 August 2026
DOI：10.47297/taposatWSP2633-456928.20260708
Accepted：

Scan QR Code

Yin Guohao.A Comparative Review of Artificial Intelligence for the Control of Physical Systems: Case Studies in Power-Grid Dispatch and Autonomous Driving[J].Theory and Practice of Science and Technology,2026,07(08):142-147.
DOI：

Yin Guohao.A Comparative Review of Artificial Intelligence for the Control of Physical Systems: Case Studies in Power-Grid Dispatch and Autonomous Driving[J].Theory and Practice of Science and Technology,2026,07(08):142-147. DOI： 10.47297/taposatWSP2633-456928.20260708.

摘要

Abstract

Power grids and autonomous vehicles are safety-critical physical systems governed by complex nonlinear dynamics and hard operational constraints. While learning-based artificial intelligence—primarily supervised deep learning and reinforcement learning (RL)—is increasingly proposed to complement classical control

it remains unclear how these methods systematically compare across distinct physical domains. This qualitative comparative review addresses that gap by evaluating learning-based methods against model-based non-learning baselines (such as Model Predictive Control) across power-grid dispatch and autonomous driving.Our main finding is that across both domains

the most viable design pattern for safe real-world deployment is a learned policy operating within a deterministic safety envelope. While learning-based techniques consistently demonstrate significant performance gains in prediction accuracy and data-efficient decision-making

their unconstrained deployment remains unsafe; safe operation is invariably mediated by verification layers

rule-based fallbacks

or human oversight. We further show that execution timescales critically dictate maturity: minute-scale grid dispatch allows thicker verification and supervisory layers than sub-second vehicular control. Finally

we outline the limits of these findings—including purposive case selection

publication bias toward positive results

and non-comparable metrics across evidence tiers—establishing this hybrid safety-envelope architecture as a compelling recurrent design pattern rather than a proven universal optimum.

关键词

Keywords

references

Ibrahim, M. S., Dong, W., & Yang, Q. (2020). Machine learning driven smart electric power systems: Current trends and new perspectives. Applied Energy, 272, 115237.

Bojarski, M., Del Testa, D., Dworakowski, D., Firner, B., Flepp, B., Goyal, P., Jackel, L. D., Monfort, M., Muller, U., Zhang, J., Zhang, X., Zhao, J., & Zieba, K. (2016). End to end learning for self-driving cars. arXiv.

Feng, S., Sun, H., Yan, X., Zhu, H., Zou, Z., Shen, S., & Liu, H. X. (2023). Dense reinforcement learning for safety validation of autonomous vehicles. Nature, 615(7953), 620–627.

Yin, L., Gao, Q., Zhao, L., Zhang, B., Wang, T., Li, S., & Liu, H. (2020). A review of machine learning for new generation smart dispatch in power systems. Engineering Applications of Artificial Intelligence, 88, 103372.

Han, S., Xiao, F., & Cheng, W. (2023). Overview of the application of deep reinforcement learning in autonomous driving systems. Journal of Xihua University (Natural Science Edition), 42(4), 25–31.

Sutton, R. S., & Barto, A. G. (2018). Reinforcement learning: An introduction (2nd ed.). MIT Press.

Rawlings, J. B., Mayne, D. Q., & Diehl, M. M. (2017). Model predictive control: Theory, computation, and design (2nd ed.). Nob Hill Publishing.

Feng, D., Haase-Schütz, C., Rosenbaum, L., Hertlein, H., Gläser, C., Timm, F., Wiesbeck, W., & Dietmayer, K. (2021). Deep multi-modal object detection and semantic segmentation for autonomous driving: Datasets, methods, and challenges. IEEE Transactions on Intelligent Transportation Systems, 22(3), 1341–1360.

Karniadakis, G. E., Kevrekidis, I. G., Lu, L., Perdikaris, P., Wang, S., & Yang, L. (2021). Physics-informed machine learning. Nature Reviews Physics, 3(6), 422–440.

Mayne, D. Q. (2014). Model predictive control: Recent developments and future promise. Automatica, 50(12), 2967–2986.

Marot, A., Donnot, B., Romero, C., Donon, B., Lerousseau, M., Veyrin-Forrer, L., & Guyon, I. (2020). Learning to run a power network challenge for training topology controllers. Electric Power Systems Research, 189, 106635.

Kendall, A., Hawke, J., Janz, D., Mazur, P., Reda, D., Allen, J.-M., Lam, K.-D., Bewley, A., & Shah, A. (2019). Learning to drive in a day. In 2019 International Conference on Robotics and Automation (ICRA) (pp. 8248–8254). IEEE.

Gamble, C., & Gao, J. (2018, August 17). Safety-first AI for autonomous data center cooling and industrial control. DeepMind. Retrieved July 18, 2026, from https://deepmind.google/blog/article/safety-first-ai-autonomous-data-centre-cooling-and-industrial-control/

Sha, L. (2001). Using simplicity to control complexity. IEEE Software, 18(4), 20–28.

Dauner, F., Hallgarten, M., Tian, A., Geiger, A., & Kashyap, A. (2023). Parting with misconceptions about learning-based vehicle motion planning. In Conference on Robot Learning (CoRL) / arXiv: 2306.07962 .