LLM Adaptive PID Control for B5G Truck Platooning Systems

Area of Science:

• Intelligent Transportation Systems
• Wireless Communication Networks
• Artificial Intelligence in Control Systems

Background:

• Truck platooning requires advanced control systems for safety and efficiency.
• 5G and Beyond 5G (B5G) networks offer enhanced connectivity for vehicle-to-vehicle (V2V) communication.
• Traditional PID controllers may struggle with dynamic network conditions like latency and packet loss.

Purpose of the Study:

• To develop and evaluate an adaptive PID controller for truck platooning using AI.
• To investigate the impact of communication parameters (latency, packet loss, range) on controller performance.
• To explore the integration of Large Language Models (LLMs) for real-time system updates.

Main Methods:

• Developed a Deep Learning (DL) model to emulate an adaptive PID controller.
• Simulated communication impairments including latency, packet loss, and limited range.
• Utilized GPT-3.5-turbo (a Large Language Model) for instantaneous performance feedback to the controller.

Main Results:

• The DL-emulated adaptive PID controller demonstrated effectiveness in truck platooning scenarios.
• Controller performance was analyzed under various communication constraints, highlighting robustness.
• LLM integration provided real-time updates, enhancing system responsiveness.

Conclusions:

• AI-enhanced adaptive PID controllers are viable for truck platooning in B5G networks.
• LLMs show promise for real-time control system optimization in advanced communication environments.
• This research provides a foundation for safer and more efficient autonomous vehicle operations.