mSAIL: Milligram-Scale Multi-Modal Sensor & Analytics Monitoring Platform for Monarch Butterfly Migration Tracking
The voluneteer light and temeperature data for 2018, 2019, 2020 are stored on ./Light_Data and ./Temp_Data. Volunteer data coordinates are intentionally made inaccurate in the resolution of 10km and we will gradually populate more data as we get permission from volunteers.
-
In order to generate the training data, you need to simply run the Generate_trainset_light.m and Generate_trainset_tmp.m.
-
In order to generate the testing data, you need to simply run the Generate_testset_light.m and Generate_testset.m.
Neural network models are located in ./Model_codes. It includes both Python and Jupyter notebook versions. You can simply run the Mobicom_Temp.ipynb and Mobicom_light.ipynb to generate the heatmap results. The pretrained neural network models are located on ./Trained_models/ .
Heatmap_Mobicom.m and Visualization_Sampling.m contain the codes to generate the localization results from the heatmaps generated from the NN models. They also include the codes to generate the following pictures.
The platform as a whole as well as individual chips are available at cost to academic researchers. If you have any questions about the chips, please contact the following emails: hunseok@umich.edu , inhee.lee@pitt.edu , and blaauw@umich.edu .