Introduction
Vehicular Ad-hoc Networks (VANETs) are a cornerstone of future intelligent transportation systems, yet they face significant reliability and latency challenges. Traditional Dedicated Short-Range Communication (DSRC) relies on distributed contention-based access, which often degrades in high-density traffic. This paper proposes a centralized link-level scheduling strategy within a Software Defined Vehicular Network (SDVN) framework to optimize data transmission and guarantee Quality of Service (QoS).
The Limitations of Distributed Contention
Current VANETs primarily utilize the CSMA/CA (Carrier Sense Multiple Access with Collision Avoidance) protocol. While effective for general wireless networking, its "listen-before-talk" nature is ill-suited for the high-speed, safety-critical environment of vehicular communications.
Key issues identified in the study include:
- Uncertain Delays: The random backoff mechanism makes it difficult to guarantee the strict latency required for safety applications.
- Silent Collisions: In broadcast scenarios, vehicles often collide without the sender's knowledge, leading to massive data loss.
- Miss My Packets (MMP): Most vehicular receivers possess only one physical processing unit. If a receiver is busy processing an irrelevant packet when a critical one arrives, the important data is missed entirely.
Centralized Scheduling via SDVN
To address these inefficiencies, the authors leverage the SDVN architecture. This framework introduces a centralized controller—typically located at a Roadside Unit (RSU)—that maintains a global view of the network topology and active connections.
The TSGS Algorithm
The core of the proposed solution is the Transmission Scheduling Greedy Search (TSGS) algorithm. Rather than allowing nodes to compete for channel access, the RSU calculates the optimal "start-sending time" for every connection. The goal is to minimize temporal overlap between transmissions. While finding a global optimum is computationally expensive ($O(M^N)$), the TSGS algorithm offers a low-complexity alternative ($O(MN)$) that is feasible for real-time use in RSUs.
QoS Integration
The scheduler explicitly respects Quality of Service requirements by prioritizing data based on standard Access Categories (AC). This ensures that safety-critical voice and video data receive preferential treatment over background traffic.
Performance Evaluation and Results
The strategy was evaluated using a high-fidelity simulation pipeline. Traffic mobility was handled by SUMO, while network modeling was performed via OMNeT++ and the Veins framework. Real-world map data from the M1 highway near Loughborough University, UK, provided a realistic topology for testing.
The TSGS algorithm demonstrated significant improvements over traditional random access methods:
- Packet Delivery Ratio (PDR): Improved by 15.28% in multi-connection scenarios.
- Collision Avoidance: Enhanced performance by nearly 40%.
- MMP Reduction: Reduced the "Miss My Packets" ratio by approximately 3%.
- Efficiency: Total transmission durations were reduced by up to 33% by avoiding the lengthy backoff cycles inherent in contention-based systems.
Conclusion
By transitioning from chaotic distributed contention to a managed, centralized scheduling approach, SDVNs can significantly enhance the reliability of vehicular communications. The TSGS algorithm provides a practical, low-complexity method for RSUs to manage link-level scheduling, ensuring that critical data arrives both reliably and on time. Future work will focus on scaling this approach to higher vehicle densities and addressing potential control channel overhead.
Summary updated