Packet Size-Aware Broadcasting in VANETs With Fuzzy Logic and RL-Based Parameter Adaptation

• Main Authors: Celimuge Wu, Xianfu Chen, Yusheng Ji, Fuqiang Liu, Satoshi Ohzahata, Tsutomu Yoshinaga, Toshihiko Kato
• Format: Article
• Language: English
• Published: IEEE 2015-01-01
• Series: IEEE Access
• Subjects: Vehicular ad hoc networks; broadcast protocol; reinforcement learning; fuzzy logic
• Online Access: https://ieeexplore.ieee.org/document/7335572/

Most existing multi-hop broadcast protocols for vehicular ad hoc networks do not consider the problem of how to adapt transmission parameters according to the network environment. Besides the propagation environment that determines the channel bit error rate, packet payload size has a significant effect on the packet loss rate. In this paper, we first discuss the effect of packet size on the packet reception ratio, and then propose a broadcast protocol that is able to specify the best relay node by taking into account the data payload size. The proposed protocol employs a fuzzy logic-based algorithm to jointly consider multiple metrics (link quality, inter-vehicle distance, and vehicle mobility) and uses a redundancy transmission approach to ensure high reliability. Since the fuzzy membership functions are tuned by using reinforcement learning, the protocol can adapt to various network scenarios. We use both real-world experiments and computer simulations to evaluate the proposed protocol.