Traditional vehicle platooning systems rely
on predefined heuristics and rule-based approaches,
which lack adaptability to dynamic traffic conditions
and unexpected changes in vehicle behavior. This paper
utilizes reinforcement learning (RL) for vehicle platooning and illustrates with simulation. The simulation,
developed in Pygame, models four autonomous vehicles,
with RL optimizing their speed to maintain an efficient
and collision-free platoon formation. The RL agent
employs a Q-learning framework to dynamically adjust
vehicle speeds based on real-time distance measurements
and a reward function that promotes smooth traffic
flow. By leveraging inter-vehicle communication, the
system ensures secure and coordinated decision-making
among autonomous vehicles. The experimental results
demonstrate the system’s capability to maintain spacing,
enable vehicle merging into a platoon, and adjust speeds
when a leading vehicle exits, while highlighting the
significance of robust network connectivity and security
for reliable autonomous platooning. By ensuring both
coordination and resilience against cyber-physical risks
such as spoofing or jamming, the proposed framework
enhances safety awareness in autonomous driving and
critical infrastructure transportation.