Networking in High Mobility Vehicular and Aerial Environments

You are reading version 1 of this direction, frozen 2026-10-04. The latest version is version 2.

Version 1 of 2519 papers579 conceptsfrozen 2026-10-04

In traditional networking, we often assume that the infrastructure is static and the users move slowly. However, in the emerging Internet of Vehicles (IoV) and Unmanned Aerial Vehicle (UAV) networks, this assumption collapses. When nodes move at high speeds, the network topology changes on a millisecond scale, rendering standard routing tables obsolete before they can even be propagated.

The core challenge in this field is volatility. As noted by Weidong Xiang and colleagues in their work on Wireless Access in Vehicular Environments, high velocities result in extremely short "dwelling times"—the window during which a vehicle is within range of a fixed access point. This fleeting connectivity creates a "dark area" problem where traditional handshakes and data transfers fail.

The Failure of Static Protocols

In a static network, a path from point A to point B is a sequence of stable links. In a vehicular ad-hoc network (VANET), that path is a flickering graph. Yang Yang 0001 and others have highlighted that high-speed mobility causes neighbor information to become "stale" almost instantly, leading to the "broadcast storm" problem where nodes desperately flood the network to find a route that no longer exists.

Even when links are active, they are unreliable. Mohsen Sardari and Faramarz Hendessi point out in Infocast that the mismatch between vehicle velocity and routing update frequency prevents reliable bulk data dissemination. To solve this, researchers are moving away from trying to maintain a "perfect" map of the network, instead embracing probabilistic and predictive models.

Predictive Intelligence and Social Routing

If we cannot know exactly where a node is, we must predict where it will be. One approach treats the road network as a Markov Decision Process (MDP). Okyoung Choi and Song Chong demonstrated that by leveraging traffic statistics and predictable schedules (like those of public buses), they could create delay-optimal routing policies that outperform traditional protocols.

Others look at the "social" behavior of vehicles. Xiaoqiang Xiao and Weixun Ning argue that vehicles aren't random moving points; they have social attributes like "centricity" and "location preference." By routing messages toward vehicles that are socially "central" to a traffic pattern, the network can maintain higher delivery rates despite physical link instability.

Offloading to the Moving Edge

As autonomous driving increases the demand for local computation, the network must do more than just pass messages; it must share work. Vehicular Edge Computing (VEC) treats cars and roadside units (RSUs) as a distributed cloud.

However, offloading a task to a car that might drive out of range in three seconds is risky. Jie Xu 0001 and Lixing Chen address this using a Multi-Armed Bandit approach, where the system learns which vehicles are reliable "compute nodes" over time, even with delayed feedback. Similarly, Qi Qi 0001 and Zhanyu Ma use Deep Reinforcement Learning to navigate the trade-offs between vehicle mobility and the data dependencies of complex computational tasks.

The Generative and Software-Defined Future

The most recent shift in the field is toward more holistic, "knowledge-driven" architectures. Geng Sun 0001, Dusit Niyato, and Hongyang Du have proposed using Generative AI (GAI) to synthesize network topologies and estimate spectrum maps from sparse, noisy data. Instead of reacting to changes, GAI allows the network to autonomously formulate optimization problems that account for future uncertainty.

To manage this complexity, researchers like Dajun Zhang and F. Richard Yu advocate for Software-Defined Vehicular Networks (SDVNs). By moving the "brain" of the network to a centralized controller (often hosted in the cloud or on a satellite, as suggested by Minghui LiWang), the system can maintain a global view of the traffic and coordinate resources more effectively than any individual vehicle could on its own.

Ultimately, the goal of this research is to turn the inherent chaos of high-speed mobility into a structured, albeit dynamic, resource—transforming every passing car and drone into a reliable component of a global sensing and computing platform.

Papers in this version

Showing 60 of 519 papers.

History & citationVersion 2 · 2026-10-04

Current · Permanent link

  1. Version 2 (latest) · 2026-10-04 · 826 papers
  2. Version 1 · 2026-10-04 · 519 papers
Cite this direction

Latest version (a living page):

@misc{lacuna_dynamic_network_topology_dynamic_topolog_083d2a80,
  title = {dynamic network topology (dynamic topology management)},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/dynamic-network-topology-dynamic-topology-management/rdr_083d2a801c4549feada816b5899d4256},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 1, frozen text):

@misc{lacuna_networking_in_high_mobility_vehicular_an_083d2a80_v1,
  title = {Networking in High Mobility Vehicular and Aerial Environments},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/networking-in-high-mobility-vehicular-and-aerial-environments/rdr_083d2a801c4549feada816b5899d4256/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}