LoRaWAN Localization using Data Rate Aware Fingerprinting Network

Abstract: Localization is an essential in any Internet-of-Things (IoT) implementation to give meaning to its data. Low Power Wide Area Network (LPWAN) technology such as LoRaWAN enables GPS-free localization through its long-long range communication and low power consumption. Various ranging and fingerprinting techniques have been studied so far. However, they suffer high positioning errors due to the Received Signal Strength (RSSI) variation. In this study, a Data Rate Aware Fingerprinting Network for LoRaWAN localization is developed to mitigate RSSI variation using the Spreading Factor SF information from sensor nodes that preprocess the RSSI out of range data to improve the performance. A publicly available dataset that offers signal strength and SF of the sensor node is used and is split to train, validate, and test the neural network model. A reduced version of the dataset is also utilized wherein messages with less than three receiving base station are discarded. The fingerprinting model achieved a mean positioning error of 293.60 meters on the publicly available dataset and 225.57 meters in the reduced version.

Dependencies

Ensure that the following libraries are installed before running the python. Tensorflow numpy pandas scikit-learn haversine

Training the DRAFNET

To train the DRAFNET using the original antwerp lorawan dataset, run the following script: python MLP_withPCA+SF --trial-name "trial name" --pca=[number of principal component 1-72] --epoch=[trainingepoch] --sf=[1-> sf input on , 2-> sf input off] --oor=[[0]-200dBm [1]-128dBm [2]SF dependent]

to train the DRAFNET using the reduced antwerp lorawan dataset, run python MLP_withPCA+SF_dataset_modified --trial-name "trial name" --pca=[number of principal component 1-72] --epoch=[trainingepoch] --sf=[1-> sf input on , 2-> sf input off] --oor=[[0]-200dBm [1]-128dBm [2]SF dependent]

To run all the trials, run the script python run_all_trials.py For generating the evaluation graph run

EVALUATION

Note, make sure that the .json and H5 file of the trial is available before running the evaluation

python model_evaluator_original.py --trial-name, -- model-name --pca, --sf,--oor for original dataset or python model_evaluator_modified.py --trial-name, -- model-name --pca, --sf,--oor for reduced dataset

to plot all the performance graph, run python graph_test.py

--trial-name ""MLP -- model-name "MLP"--pca, --sf,--oor

Name		Name	Last commit message	Last commit date
Latest commit History 65 Commits
KNN_modified		KNN_modified
KNN_original		KNN_original
MLP_modified		MLP_modified
MLP_original		MLP_original
__pycache__		__pycache__
extratress_modified		extratress_modified
extratress_original		extratress_original
figures		figures
models		models
Baseline-Copy1.ipynb		Baseline-Copy1.ipynb
Baseline.ipynb		Baseline.ipynb
KNN.py		KNN.py
KNN_experiment.py		KNN_experiment.py
KNN_mod.py		KNN_mod.py
KNN_model_performance_mean_modified.png		KNN_model_performance_mean_modified.png
KNN_model_performance_mean_original.png		KNN_model_performance_mean_original.png
LICENSE		LICENSE
MLP+PCA=10+SF.png		MLP+PCA=10+SF.png
MLP+PCA=10+SF_SFD.png		MLP+PCA=10+SF_SFD.png
MLP+PCA=10.png		MLP+PCA=10.png
MLP+PCA=10_SFD.png		MLP+PCA=10_SFD.png
MLP+PCA=3+SF.png		MLP+PCA=3+SF.png
MLP+PCA=3+SF_SFD.png		MLP+PCA=3+SF_SFD.png
MLP+PCA=3.png		MLP+PCA=3.png
MLP+PCA=3_SFD.png		MLP+PCA=3_SFD.png
MLP+PCA=40+SF.png		MLP+PCA=40+SF.png
MLP+PCA=40+SF_SFD.png		MLP+PCA=40+SF_SFD.png
MLP+PCA=40.png		MLP+PCA=40.png
MLP+PCA=40_SFD.png		MLP+PCA=40_SFD.png
MLP+PCA=44+SF.png		MLP+PCA=44+SF.png
MLP+PCA=44+SF_SFD.png		MLP+PCA=44+SF_SFD.png
MLP+PCA=44.png		MLP+PCA=44.png
MLP+PCA=44_SFD.png		MLP+PCA=44_SFD.png
MLP+PCA=5+SF.png		MLP+PCA=5+SF.png
MLP+PCA=5+SF_SFD.png		MLP+PCA=5+SF_SFD.png
MLP+PCA=5.png		MLP+PCA=5.png
MLP+PCA=5_SFD.png		MLP+PCA=5_SFD.png
MLP+PCA=7+SF.png		MLP+PCA=7+SF.png
MLP+PCA=7+SF_SFD.png		MLP+PCA=7+SF_SFD.png
MLP+PCA=7.png		MLP+PCA=7.png
MLP+PCA=7_SFD.png		MLP+PCA=7_SFD.png
MLP+SF.png		MLP+SF.png
MLP+SF_SFD.png		MLP+SF_SFD.png
MLP-3 Layers.ipynb		MLP-3 Layers.ipynb
MLP.ipynb		MLP.ipynb
MLP.png		MLP.png
MLP_SFD.png		MLP_SFD.png
MLP_model_performance_mean_modified.png		MLP_model_performance_mean_modified.png
MLP_model_performance_mean_original.png		MLP_model_performance_mean_original.png
MLP_withPCA+SF.py		MLP_withPCA+SF.py
MLP_withPCA+SF_dataset_modified.py		MLP_withPCA+SF_dataset_modified.py
MLP_withPCA=3.ipynb		MLP_withPCA=3.ipynb
MLP_withPCA=4.ipynb		MLP_withPCA=4.ipynb
MLP_withPCA=5.ipynb		MLP_withPCA=5.ipynb
MLP_withPCA=6.ipynb		MLP_withPCA=6.ipynb
MLP_withPCA=7.ipynb		MLP_withPCA=7.ipynb
MLP_withPCA=8.ipynb		MLP_withPCA=8.ipynb
MLP_withPCA=9.ipynb		MLP_withPCA=9.ipynb
OOR1+SF_model_performance_mean_original.png		OOR1+SF_model_performance_mean_original.png
OOR1+SF_performance_mean_modified.png		OOR1+SF_performance_mean_modified.png
OOR1+SF_performance_mean_original.png		OOR1+SF_performance_mean_original.png
OOR1_model_performance_mean_original.png		OOR1_model_performance_mean_original.png
OOR1_performance_mean_modified.png		OOR1_performance_mean_modified.png
OOR1_performance_mean_original.png		OOR1_performance_mean_original.png
OOR2+SF_model_performance_mean_original.png		OOR2+SF_model_performance_mean_original.png
OOR2+SF_performance_mean_modified.png		OOR2+SF_performance_mean_modified.png
OOR2_model_performance_mean_original.png		OOR2_model_performance_mean_original.png
OOR2_performance_mean_modified.png		OOR2_performance_mean_modified.png
OOR2_performance_mean_original.png		OOR2_performance_mean_original.png
README.md		README.md
SF_normalize_experiment.py		SF_normalize_experiment.py
best_MLP.py		best_MLP.py
dataset_creation.py		dataset_creation.py
dataset_plotter.py		dataset_plotter.py
errorplotter.py		errorplotter.py
ext_model_performance_mean_modified.png		ext_model_performance_mean_modified.png
ext_model_performance_mean_original.png		ext_model_performance_mean_original.png
ext_model_performance_mean_reduced.png		ext_model_performance_mean_reduced.png
extratress.py		extratress.py
extratress_mod.py		extratress_mod.py
graph_test.py		graph_test.py
haversine_script.py		haversine_script.py
lorawan_antwerp_2019_dataset.csv		lorawan_antwerp_2019_dataset.csv
lorawan_antwerp_2019_dataset_withSF.csv		lorawan_antwerp_2019_dataset_withSF.csv
model_evaluator_modified.py		model_evaluator_modified.py
model_evaluator_original.py		model_evaluator_original.py
model_performance_mean.png		model_performance_mean.png
model_test.py		model_test.py
model_test_pca.py		model_test_pca.py
run_all_evaluation.py		run_all_evaluation.py
run_all_trials.py		run_all_trials.py

Navigation Menu

Repository files navigation

LoRaWAN Localization using Data Rate Aware Fingerprinting Network

Dependencies

Training the DRAFNET

EVALUATION

About

Releases

Packages

Contributors

Languages

Folders and files

Latest commit

History

Repository files navigation

LoRaWAN Localization using Data Rate Aware Fingerprinting Network

Dependencies

Training the DRAFNET

EVALUATION

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages