FelixOtt/gnss_federated_learning

Federated Few-Shot Learning for Distributed Interference Classification in GNSS Data

Latest update: 2024-10-30

Project size: 11 MB

Introduction

Federated learning (FL) enables multiple devices to collaboratively train a global model while maintaining data on local servers. Each device trains the model on its local server and shares only the model updates (i.e., gradient weights) during the aggregation step. A significant challenge in FL is managing the feature distribution of novel, unbalanced data across devices. In this paper, we propose an FL approach using few-shot learning and aggregation of the model weights on a global server. We introduce a dynamic early stopping method to balance out-of-distribution classes based on representation learning, specifically utilizing the maximum mean discrepancy of feature embeddings between local and global models. An exemplary application of FL is orchestrating machine learning models along highways for interference classification based on snapshots from global navigation satellite system (GNSS) receivers. Extensive experiments on four GNSS datasets from two real-world highways and controlled environments demonstrate that our FL method surpasses state-of-the-art techniques in adapting to both novel interference classes and multipath scenarios.

We publish our source code for distributed interference classificaton as well as for the adaptation to new interference classes and new environmental scenarios.

This project can be used for research purposes to addresse the following problems:

• How can DL models be designed to adapt to novel interference sources and varying propagation environments?

• What techniques can be developed to identify and learn the unique signal patterns and features of each interference source?

• How can machine learning models effectively track and classify interference sources across different locations and environments?

• What methods can integrate machine learning techniques into a practical pipeline for efficient, accurate, and robust GNSS interference detection and localization according to their difficulty of interference class?

Installation

To install the required dependencies, you can use pip and run the following command:

conda env create -f environment.yaml

This will create the FederatedLearning environment. Then, install the package by using the follwing command:

pip install .

Usage

First, to train the few-shot model, go to Few_shot_learning and change config file as per your dataset and run the follwing command:

python trainer.py 

This will complete the initial few-shot learning.

For federated learning, navigate to the respective folder based on your training preference. You can either use parallel computing or sequential computing. For parallel computing, go to the Flower_federated folder. In the config.py file, update the global model path, which is located in the Few_shot_learning folder under outputs. Also, make embedding_diff True. To train using parallel GPUs, navigate to the Scripts folder and execute the appropriate command.

bash run_env_fed.sh --strategy=" "--port='' --run_name=''--max_epochs=' '--min_epochs= ' ' --mode='earlystopping'

If you prefer sequential training, head to the Pytorch_federated folder and adjust the global model path accordingly. In training files of clients, change emb_diff[:new_class_index] = 0.

If you want to run it then use:

bash run_seq_gpu.sh

For the federated learning strategy you have the following options:

choices=[
  "FaultTolerantFedAvg", #Only for parallel
  "FedAdagrad",          #Only for parallel
  "FedAdam",   
  "FedAvg", 
  "FedXgbNnAvg",         #Only for parallel
  "FedAvgAndroid",       #Only for parallel
  "FedAvgM",
  "FedOpt",
  "FedProx",
  "FedYogi",             #Only for parallel
  "QFedAvg",
  "FedMedian",
  "FedTrimmedAvg",
]

For testing, change the file path in test.py and run:

python test.py

python collect_results.py

python plot.py

In addition, the Test folder contains files to calculate the t-SNE plot (tsne.py). In the Notebook folder, we have different files for visualization and calculation.

Dataset

The datasets con be downloaded from the following websites:

For our spectrum highway dataset 1 recorded along a highway bridge, see the following website: https://gitlab.cc-asp.fraunhofer.de/darcy_gnss/FIOT_highway

For our spectrum highway dataset 2 recorded along a highway bridge, see the following website: https://gitlab.cc-asp.fraunhofer.de/darcy_gnss/fiot_highway2

For our dataset recorded in a controlled environment with a high-frequency antenna, see the following website: https://gitlab.cc-asp.fraunhofer.de/darcy_gnss/FIOT_LC_laboratory

For our dataset recorded in a controlled environment with a low-frequency antenna, see the following website: https://gitlab.cc-asp.fraunhofer.de/darcy_gnss/controlled_low_frequency

References

For more information of the dataset and results, see our publication. If you use our dataset for your research, please consider citing:

@inproceedings{gaikwad_heublein,
  author = {Nishant S. Gaikwad and Lucas Heublein and Nisha L. Raichur and Tobias Feigl and Christopher Mutschler and Felix Ott},
  title = {{Federated Learning with MMD-based Early Stopping for Adaptive GNSS Interference Classification}},
  booktitle = {\href{https://arxiv.org/abs/2410.15681}{arXiv preprint arXiv:2410.15681}},
  month = oct,
  year = {2024}
}

For more information on the datasets, see:

@inproceedings{heublein_raichur_ion,
  author = {Lucas Heublein and Nisha L. Raichur and Tobias Feigl and Tobias Brieger and Fin Heuer and Lennart Asbach and Alexander Rügamer and Felix Ott},
  title = {{Evaluation of (Un-)Supervised Machine Learning Methods for GNSS Interference Classification with Real-World Data Discrepancies}},
  booktitle = {\href{https://www.ion.org/publications/abstract.cfm?articleID=19887}{Proc. of the Intl. Technical Meeting of the Satellite Division of the Institute of Navigation (ION GNSS+)}},
  pages = {1260--1293},
  month = sep,
  year = 2024,
  address = {Baltimore, MD},
  doi = {10.33012/2024.19887}
}

@inproceedings{heublein_feigl_dataset,
  author = {Lucas Heublein and Tobias Feigl and Thorsten Nowak and Alexander Rügamer and Christopher Mutschler and Felix Ott},
  title = {{Evaluating ML Robustness in GNSS Interference Classification, Characterization \& Localization}},
  booktitle = {\href{https://arxiv.org/abs/2409.15114}{arXiv preprint arXiv:2409.15114}},
  month = sep,
  year = {2024}
}

Acknowledgment

This work has been carried out within the DARCII project, funding code 50NA2401, supported by the German Federal Ministry for Economic Affairs and Climate Action (BMWK), managed by the German Space Agency at DLR and assisted by the Bundesnetzagentur (BNetzA) and the Federal Agency for Cartography and Geodesy (BKG). Additionally, this work was supported by the Bavarian Ministry of Economic Affairs, Regional Development and Energy through the Center for Analytics – Data – Applications (ADA-Center) within the framework of „BAYERN DIGITAL II“ (20-3410-2-9-8).

License

This work is licensed under a CC BY-NC-SA 4.0: Creative Commons Attribution-Noncommercial-ShareAlike, see https://creativecommons.org/licenses/by-nc-sa/4.0/

Kontakt

If you have any questions or tips to improve the datasets, contact us:

Felix Ott: felix.ott@iis.fraunhofer.de

Lucas Heublein: lucas.heublein@iis.fraunhofer.de

Nordostpark 84, 90411 Nürnberg, Germany, GoogleMaps

Fraunhofer Institute for Integrated Circuits IIS, Nürnberg, Germany