Securing Electrical Drive Systems Against Man-in-the-Middle Attacks Using S-Box Optimized AES Encryption

The Protection of electric drive systems is significant in modern industrial applications to ensure reliable operations. These systems consist of multiple interconnected control units that communicate wirelessly, making them vulnerable to cyber threats. This paper focuses on ensuring secure end-to-end communication within a drive system that includes a 2.2 kW induction motor drive controlled by the F28379D microcontroller and connected through the ESP8266 Wi-Fi microcontroller. To address the security challenges of such interconnected systems, this study proposes a solution utilizing an optimized Advanced Encryption Standard (AES) algorithm to safeguard data against man-in-the-middle (MitM) attacks. The microcontroller is connected to the cloud through the ThingSpeak platform for control program deployment and drive parameter monitoring. The cloud environment handles both private and public data sources, which introduces vulnerabilities to cyber-attacks. This vulnerability is mitigated by the encryption and decryption processes. The security of the AES encryption is enhanced through diffusion and confusion mechanism, using the optimized S-box, which adds confusion during the substitution phase. The optimized S-box is generated using the Differential Evolution (DE) algorithm for encrypting data within the communication network to mitigate data visibility. The optimization performance is evaluated using nonlinearity, avalanche effect, and bit independence metrics, while the AES algorithm’s performance is analyzed in terms of encryption throughput and latency. Results indicate that the nonlinearity of the DE-optimized S-box is slightly lower compared to the genetic algorithm (GA)- optimized S-box, with similar trends observed for the avalanche effect and bit independence criterion. However, the DE-optimized S-box demonstrates superior performance in encryption throughput and latency for various data sizes compared to the GA-optimized S-box. This demonstrates that the proposed optimization algorithm can handle data encryption in less time, which is important for real-time applications in industrial drive systems.