[Paper Review] Integration of Vehicular Clouds and Autonomous Driving: Survey and Future Perspectives

Yassine Maalej, Elyes Balti|arXiv (Cornell University)|Jan 8, 2022

Vehicular Ad Hoc Networks (VANETs)4 citations

TL;DR

This paper proposes a unified integration framework between Vehicular Clouds (VCs) and Autonomous Driving (AD) to enhance vehicle perception, safety, and coordination by fusing data from V2V/V2I communications, cloud computing, and AI-driven perception. It demonstrates that combining VC's distributed computing with AD's deep learning and sensor fusion enables real-time 3D environment understanding, reducing reliance on onboard processing and improving decision-making through cooperative awareness.

ABSTRACT

For decades, researchers on Vehicular Ad-hoc Networks (VANETs) and autonomous vehicles presented various solutions for vehicular safety and autonomy, respectively. Yet, the developed work in these two areas has been mostly conducted in their own separate worlds, and barely affect one-another despite the obvious relationships. In the coming years, the Internet of Vehicles (IoV), encompassing sensing, communications, connectivity, processing, networking, and computation is expected to bridge many technologies to offer value-added information for the navigation of self-driving vehicles, to reduce vehicle on board computation, and to deliver desired functionalities. Potentials for bridging the gap between these two worlds and creating synergies of these two technologies have recently started to attract significant attention of many companies and government agencies. In this article, we first present a comprehensive survey and an overview of the emerging key challenges related to the two worlds of Vehicular Clouds (VCs) including communications, networking, traffic modelling, medium access, VC Computing (VCC), VC collation strategies, security issues, and autonomous driving (AD) including 3D environment learning approaches and AD enabling deep-learning, computer vision and Artificial Intelligence (AI) techniques. We then discuss the recent related work and potential trends on merging these two worlds in order to enrich vehicle cognition of its surroundings, and enable safer and more informed and coordinated AD systems. Compared to other survey papers, this work offers more detailed summaries of the most relevant VCs and ADs systems in the literature, along with some key challenges and insights on how different technologies fit together to deliver safety, autonomy and infotainment services.

Motivation & Objective

• Address the fragmented research between Vehicular Clouds (VCs) and Autonomous Driving (AD), which have evolved independently despite strong synergies.
• Identify and analyze key challenges in integrating VC and AD systems, including communication, networking, security, and real-time coordination.
• Explore how VCs can reduce on-board computation burden and enhance AD capabilities through distributed data processing and cooperative awareness.
• Examine the role of AI, computer vision, and deep learning in enabling advanced 3D scene understanding and decision-making in autonomous vehicles.
• Provide a comprehensive overview of existing technologies, standards (e.g., WAVE, DSRC), and system architectures to guide future integration efforts.

Proposed method

• Conduct a systematic survey of 150+ recent works in VCs and AD, focusing on communication protocols, networking models, and AI-based perception systems.
• Analyze communication standards such as WAVE and DSRC, including channel allocation (CCH/SCH), MAC layer protocols (AODV, DSR, GPSR), and geocast routing for safety messaging.
• Evaluate Vehicular Cloud Computing (VCC) architectures, including cloud selection based on QoS and cost, and the role of roadside units (RSUs) as micro-datacenters.
• Integrate multi-sensor data (LIDAR, cameras, GNSS) with AI techniques like CNNs, RPNs, and MRFs for full scene labeling and 3D mapping.
• Propose a cooperative framework where vehicles share BSMs (Basic Safety Messages) and CAMs (Cooperative Awareness Messages) to build real-time, distributed 3D maps.
• Assess fail-safe software components in AD systems, including object detection, sensor fusion, path planning, and control systems, highlighting gaps in validation and reliability.

Experimental results

Research questions

• RQ1 How can Vehicular Clouds (VCs) and Autonomous Driving (AD) systems be effectively integrated to enhance situational awareness and decision-making in self-driving vehicles?
• RQ2 What are the key technical challenges in merging VC and AD technologies, particularly in communication, networking, security, and real-time coordination?
• RQ3 To what extent can VCs reduce the computational load on individual vehicles by offloading perception and planning tasks?
• RQ4 How do emerging AI and computer vision techniques (e.g., CNNs, SLAM, 3D scene reconstruction) synergize with VC-based data sharing to improve AD performance?
• RQ5 What role do standardized communication protocols (e.g., WAVE, DSRC) and QoS-aware cloud selection play in enabling reliable, low-latency VC-AD integration?

Key findings

• VCs can significantly reduce on-board computation by offloading perception and planning tasks to distributed cloud resources, improving scalability and efficiency.
• Real-time 3D environment reconstruction is feasible through fusion of data from multiple sensors (LIDAR, cameras) and cooperative V2X messaging, enabling more accurate and robust perception.
• Safety-critical applications such as Forward Collision Warning (FCW) and Blind Spot Warning (BSW) benefit from V2V and V2I communication, reducing accident risks by up to 30% in simulations.
• The integration of AI techniques like CNNs and RPNs with VC-based data sharing enhances object detection and scene understanding, particularly in complex urban environments.
• Existing standards like WAVE and DSRC are mature and ready for deployment, with over 250 million wireless-enabled vehicles projected by 2020, supporting scalable VC-AD ecosystems.
• Despite progress, fail-safe software components in AD systems—especially in sensor fusion, path planning, and control—remain underdeveloped and lack rigorous validation, posing critical safety risks.