Michael Wilbur, Chinmaya Samal, Jose Paolo Talusan, Keiichi Yasumoto, Abhishek Dubey
2020 IEEE 23nd International Symposium on Real-Time Distributed Computing (ISORC) 2020
Centralized route planning services face latency and availability concerns in disaster scenarios, while federated learning enables distributed model training without centralizing sensitive data. This work is innovative because it combines decentralized routing with federated learning to create privacy-preserving, resilient route planning that relies on edge computing. The approach maintains data locality while enabling collaborative learning across the distributed network.
This paper presents a decentralized time-dependent routing approach using federated learning on private fog networks where RSUs collaboratively learn shared prediction models. The system enables route planning without relying on centralized cloud services by leveraging federated learning to collaboratively train models that predict travel times and select optimal routes. All training occurs locally on RSUs with only model weights shared to the central server, avoiding raw data transmission.
The federated learning approach achieved comparable latency and memory efficiency to centralized methods while reducing dependence on cloud infrastructure and protecting data privacy. The system successfully demonstrated time-dependent routing in a simulated Nashville metropolitan area with multiple RSUs. The framework proved effective at learning shared prediction models that improved query response time while maintaining acceptable accuracy loss due to distributed computation.
@inproceedings{wilbur2020decentralized,
author = {Wilbur, Michael and Samal, Chinmaya and Talusan, Jose Paolo and Yasumoto, Keiichi and Dubey, Abhishek},
booktitle = {2020 IEEE 23nd International Symposium on Real-Time Distributed Computing (ISORC)},
title = {Time-dependent Decentralized Routing using Federated Learning},
year = {2020},
organization = {IEEE},
abstract = {Recent advancements in cloud computing have driven rapid development in data-intensive smart city applications by providing near real time processing and storage scalability. This has resulted in efficient centralized route planning services such as Google Maps, upon which millions of users rely. Route planning algorithms have progressed in line with the cloud environments in which they run. Current state of the art solutions assume a shared memory model, hence deployment is limited to multiprocessing environments in data centers. By centralizing these services, latency has become the limiting parameter in the technologies of the future, such as autonomous cars. Additionally, these services require access to outside networks, raising availability concerns in disaster scenarios. Therefore, this paper provides a decentralized route planning approach for private fog networks. We leverage recent advances in federated learning to collaboratively learn shared prediction models online and investigate our approach with a simulated case study from a mid-size U.S. city.},
contribution = {lead},
tag = {decentralization,transit},
keywords = {federated learning, decentralized routing, time-dependent networks, fog computing, privacy-preserving, collaborative learning}
}