Quantum Long Short-Term Memory-Assisted Optimization for Efficient Vehicle Platooning in Connected and Autonomous Systems

Abstract

Vehicle platooning, especially when dedicated to carrying goods, represents a forward-looking approach to optimizing logistics and freight transportation using autonomous vehicles. In this study, we propose to employ Quantum Long Short Term Memory (QLSTM) models to predict the vehicle dynamics of a leading vehicle of the platoon. This predictive capability allows the following vehicles to adjust their behaviours dynamically. By doing so, we aim to optimize control strategies and maintain string stability within vehicle platoons. This approach leverages the unique computational advantages of quantum computing, particularly in processing complex temporal data, potentially leading to more accurate and efficient dynamic systems in vehicular platoon infrastructure. The simulation results indicate that the QLSTM model is highly efficient by learning more information in fewer epochs compared to traditional Long Short Term Memory (LSTM) models. This efficiency contributes to minimizing control errors, enhancing the precision and reliability of vehicle dynamics in the context of autonomous vehicle platooning. This research not only enhances the predictability of autonomous vehicle platoons but also opens pathways for research into how quantum computing can be integrated into real-time dynamic systems analysis and control.