Each year, the number of Electrical Vehicles (EVs) on the roads grows significantly. EVs are expected to reach 2% of the total car population by 2020 and 50% by 2030 which generates important challenges for the grid stability. At the same time, countries are massively investing in renewable energy sources which are volatile in essence. Since it is difficult to store large amounts of energy, EVs offer a major opportunity to use this energy when it is available, thereby contributing to stabilise the whole electrical network. In this project, VUB and Energis aim to develop an efficient and viable way to charge an entire fleet of EVs in a coordinated and controlled way, and to test this approach in lab and real environments. Unlike synthetic ones, real environments are characterised by a high degree of unpredictability. Therefore, the robustness of the EV charging scheduling controller is a main focus, which is why we aim to apply Reinforcement Learning to this problem, a successful technique from Artificial Intelligence well known to achieve robustness.