Networked Control Systems Secured by Quantum Key Distribution

Hai-Jin Ding, Re-Bing Wu, Qian-Chuan ZhaoarXiv 2018 · arXiv

Abstract

Networked Control Systems (NCS) are critical to modern industrial infrastructure but face a fundamental trade-off between cryptographic security and real-time control performance. This paper proposes a novel framework integrating Quantum Key Distribution (QKD) to achieve high-level security without the latency penalties of traditional encryption. By utilizing QKD-generated keys for One-Time Pad (OTP) encryption via simple XOR operations, the system maintains stability and precision. Furthermore, a Kalman-filter approach is introduced to mitigate noise from "raw" quantum keys, enhancing key utilization efficiency.

Introduction: The Security-Performance Trade-off

In an NCS, controllers and plants communicate over digital networks, making them vulnerable to cyber-attacks. Traditional encryption standards like AES or RSA rely on computational complexity, which introduces significant communication delays and "jitter." In high-speed control loops, such as robotic servos, these delays can degrade performance or lead to system instability. Moreover, these standards are theoretically susceptible to future quantum computing attacks. While One-Time Pad (OTP) encryption offers information-theoretic security, distributing the necessary volume of random keys over classical channels is impractical.

Proposed QKD-Secured Architecture

The authors address these challenges by replacing classical key exchange with Quantum Key Distribution (QKD), specifically the BB84 protocol.

Architecture of the QKD-secured NCS, showing the interaction between Alice (controller) and Bob (plant) via classical and quantum channels.

The framework consists of:

  1. Dual-Channel Communication: A controller (Alice) and a plant (Bob) are linked by a classical data channel for control signals and a quantum channel for key generation.
  2. Simplified Encryption: With a continuous stream of truly random QKD keys, computationally intensive algorithms are unnecessary. The system employs the XOR operation, minimizing processing time and packet overhead.
  3. Kalman Filtering for Raw Keys: To bypass the slow post-processing of QKD (error correction), the authors propose using "raw" keys—which may contain bit errors—to encrypt lower-order control data. A Kalman filter at the controller treats these decryption errors as white noise, maintaining precision while maximizing key rates.

Experimental Results

The approach was benchmarked using a mechanical servo system, comparing QKD+XOR against standard DES and AES implementations (AES-128, 192, 256).

Latency and Control Precision

The XOR-encrypted system achieved an average round-trip delay of 0.0645s, nearly half that of AES-256 (0.1307s). This reduction in latency directly translates to superior control accuracy.

Comparison of control performance (position vs. time) for various encryption methods. XOR encryption closely tracks the reference signal, while higher-order AES variants introduce significant oscillation and error.

As shown in the figure above, the Mean Square Error (P) for the XOR system was 0.4044, significantly lower than the 0.6300 observed with AES-256. The increased complexity of AES leads to queueing delays that compromise the stability of the control loop.

Security and Noise Robustness

Despite its simplicity, the XOR algorithm achieves a normalized security level of 0.9002 when paired with high-frequency QKD key updates, rivaling the "perfect" security of AES-256 without the performance cost. Additionally, the Kalman filter successfully suppressed fluctuations from raw keys with up to a 10% error rate, proving the system's robustness to imperfect quantum hardware.

Conclusion and Future Work

This research demonstrates that by moving security from the algorithm to the key via QKD, NCS can achieve "unbreakable" security while improving real-time performance. The inherent noise-handling capabilities of control systems (e.g., Kalman filters) allow for more efficient use of quantum resources.

Future work will focus on scaling this point-to-point model to complex multi-agent industrial networks and addressing potential vulnerabilities in quantum repeaters within larger topologies.

Summary updated

BibTeX
@misc{ding2018networked,
  author = {Hai-Jin Ding and Re-Bing Wu and Qian-Chuan Zhao},
  title = {{Networked Control Systems Secured by Quantum Key Distribution}},
  year = {2018},
  eprint = {1805.09015},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1805.09015},
}
Figure