Service Orchestration in Autonomous Vehicle Networks: Leveraging Vehicular Fog and Edge Computing

The presence of connected and autonomous vehicles is transforming the transportation landscape, improving safety, efficiency, accessibility, and sustainability. However, the disparity between the computational capabilities of autonomous vehicles and their embedded computations requires offloading computation tasks. To address this issue, RoadSide Units (RSUs) have been introduced, bringing computing units closer to the vehicles. By performing real-time computation tasks at the network edge, RSUs enable faster decision-making and reduce reliance on remote cloud servers. Another solution that enhances the capabilities of smart transportation systems is the collaboration of neighboring vehicles within the framework of vehicular fog computing. However, existing research in task placement for vehicular networks often overlooks the quality of service (QoS) in terms of vehicle-service connectivity, focusing primarily on load balancing and latency reduction. To tackle this problem, our work proposes a task orchestration solution that considers vehicle mobility factors to strike a balance between minimizing latency and improving QoS. The objective is to minimize disconnection and avoid waiting times for service relocation. We evaluated the effectiveness of our proposed solution through a comparative evaluation, which demonstrated that our approach offers the best tradeoff between average service latency and QoS improvement in terms of vehicle connectivity. These results indicate that our solution effectively enhances both performance and reliability.