Smart Predictions of current and future use of Electric Charging Stations

This project is an innovative one, combining two aspects. On a macro level, the simulation of user behaviour. On a micro level, the prediction of the occupancy rate of an electric charging station. The prediction part will use standard machine learning models and our new transparent models. The transparent models will make it possible to accurately select the best parameters for prediction. These parameters will be reused to fine-tune the simulation, so that simulated driver behaviour reflects reality as closely as possible.

The prediction tool will use Machine Learning techniques and is driver-oriented; it will be used by the driver but will provide the operator/manager with important information about the future use of its network. The simulator will provide the electric charging station manager with a map of the location of charging stations at any time, enabling him to anticipate the behavioural effects of the reorganisation of the network before deploying new charging stations or removing existing ones (under-used, optimisation, maintenance, etc.).