In this paper a novel hybrid prediction model is proposed which enables deployment of unmanned aerial vehicle (UAV) as flying base stations (UAV-BS). A UAV can assist the base station when it is overloaded. Hence, based on the forecast of the network traffic both long term as well as short term features of the network traffic are taken care. In the proposed hybrid model, cyclical encoder is used to encode the periodic features, and Long Short Term Memory (LSTM) and Temporal Convolution Network (TCN) for long term and short term feature extraction. In our hybrid model the network traffic for the specific weekday and particular hour each day are considered. Simulation results shows the error rate of our model is 0.030 (RMSLE). Compared with the existing LSTM, TCN and ARIMA models our hybrid prediction model outperform the base line models.
Deployment of UAV Base Station using Hybrid Prediction Model
15.03.2024
487398 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Public Bicycle System Station Deployment and Scale Prediction
Trans Tech Publications | 2013
|Deployment of UAV Base Station in Cellular Networks: Models and Algorithms
Springer Verlag | 2022
|