# AdaptiveFog: A Modelling and Optimization Framework for Fog Computing in Intelligent Transportation Systems

arXiv: https://arxiv.org/abs/2106.15477 · HTML: https://mlsanity.com/paper/2106.15477/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Jun 2021.

## Built on by (most cited first, 4 of 4)

- [Imitation Learning-based Implicit Semantic-aware Communication Networks: Multi-layer Representation and Collaborative Reasoning](https://mlsanity.com/paper/2210.16118/) (2022)
- [Time-sensitive Learning for Heterogeneous Federated Edge Intelligence](https://mlsanity.com/paper/2301.10977/) (2023)
- [Distributed Traffic Synthesis and Classification in Edge Networks: A Federated Self-supervised Learning Approach](https://mlsanity.com/paper/2302.00207/) (2023)
- [Adversarial Learning for Implicit Semantic-Aware Communications](https://mlsanity.com/paper/2301.11589/) (2023)

## Abstract (arXiv)

Fog computing has been advocated as an enabling technology for computationally intensive services in smart connected vehicles. Most existing works focus on analyzing the queueing and workload processing latencies associated with fog computing, ignoring the fact that wireless access latency can sometimes dominate the overall latency. This motivates the work in this paper, where we report on a five-month measurement study of the wireless access latency between connected vehicles and a fog/cloud computing system supported by commercially available LTE networks. We propose AdaptiveFog, a novel framework for autonomous and dynamic switching between different LTE networks that implement a fog/cloud infrastructure. AdaptiveFog's main objective is to maximize the service confidence level, defined as the probability that the latency of a given service type is below some threshold. To quantify the performance gap between different LTE networks, we introduce a novel statistical distance metric, called weighted Kantorovich-Rubinstein (K-R) distance. Two scenarios based on finite- and infinite-horizon optimization of short-term and long-term confidence are investigated. For each scenario, a simple threshold policy based on weighted K-R distance is proposed and proved to maximize the latency confidence for smart vehicles. Extensive analysis and simulations are performed based on our latency measurements. Our results show that AdaptiveFog achieves around 30% to 50% improvement in the confidence levels of fog and cloud latencies, respectively.