Joint Optimization of Multi-Hop Broadcast Protocol and MAC Protocol in Vehicular Ad Hoc Networks

Area of Science:

• Vehicular Ad Hoc Networks (VANETs)
• Wireless Communication Protocols
• Machine Learning Applications

Background:

• Vehicular ad hoc networks (VANETs) face challenges with low-delay and high-reliability message transmission.
• Existing MAC protocols struggle to minimize data collisions, impacting VANET performance.
• Hidden node problems and varying vehicle densities complicate reliable communication.

Purpose of the Study:

• To propose a novel Q-learning MAC protocol for VANETs to enhance beacon and emergency message delivery.
• To reduce message delay and improve reliability in highway scenarios.
• To mitigate the impact of hidden nodes and optimize MAC layer performance under diverse traffic conditions.

Main Methods:

• A Q-learning MAC protocol is developed, utilizing beacon messages and two-hop neighbor detection.
• Contention window values are trained using Q-learning based on vehicle density to minimize beacon message collisions.
• A multi-hop broadcast protocol adjusts contention windows for emergency messages, incorporating neighbor state information.

Main Results:

• The proposed Q-learning MAC protocol effectively reduces beacon message collision probability.
• The multi-hop broadcast protocol demonstrates reduced forwarding delay and improved reliability for emergency messages.
• Simulation results confirm superior performance compared to existing related protocols across various metrics.

Conclusions:

• The Q-learning MAC protocol offers a significant improvement for VANET communication reliability and delay.
• The adaptive multi-hop broadcast protocol enhances emergency message dissemination in highway environments.
• This research provides an effective solution for critical message delivery in VANETs.