Performance Evaluation of Forecasted Data and Trained Data in Long-Short Term Memory Prediction of Device To Device Communication in Lora Based Network

Summary

This work presents performance evaluation of forecasted data and trained data in long-short term memory (LSTM)
prediction of device to device (D2D) communication in long range (LoRa) based network. The received signal strength
indicator (RSSI), the path loss and the distance between the nodes were logged. The RSSI was subdivided into train
and test data. The trained data was used to train an LSTM model and forecast was made using same model. It was
observed that the forecasted data followed the same trend with the test data with root mean square error (RMSE) of
4.848 and 5.153 for each node when moving towards each other and RMSE of 4.68 and 4.17 when moving apart. It is
recommended that other environmental factor that result to varying propagation properties like shadowing, interface,
reflection and diffraction should be considered. The major contribution of this paper to knowledge is that forecasting
of the RSSI which is not done by any study reviewed has been presented.

Index Terms

Browse additional articles authored by one or more contributors to this paper.