Attention Based Bi-LSTM Bi-GRU model for Prediction of Electric Vehicle Charging Demand

The extensive progress and penetration of electric vehicles has led to the development of smart transportation systems, which moves the world towards a clean and sustainable future. The past few years have depicted the incremental volumes of electric vehicles (EVs) that have penetrated the market, leading to an increase in charging demand. As a result, it is necessary to predict the charging demand of EVs to minimize the burden on the electric grid setup and decrease the charging cost to consumers. This research study modelled a novel attention-based Bidirectional Long-Short Term Memory (Bi-LSTM) and Bidirectional Gated Recurrent Unit (Bi-GRU) to predict EV charging demand. The considered data for validating the proposed attention-based Bi-LSTM Bi-GRU (aBi-LSTM Bi-GRU) are time-series data, and the presence of Bi-LSTM and Bi-GRU overcomes the existence of vanishing and exploding gradients, as in the case of classic recurrent deep learning neural network models. The new modelled aBi-LSTM Bi-GRU was tested for its efficacy on an Electric Vehicle Charging dataset. Simulation experiments proved the superiority of the developed deep learning-based predictive model for the considered dataset over the other baseline models employed for comparison.