Abstract

The progression in developing autonomous electric vehicles (AEVs) leads to a demand for innovative solutions that make use of their energy storage capacities. Alongside, the advances in the energy transition towards renewable energy with rising numbers of distributed energy resources (DER) offer the opportunity for prosumers to bidirectionally interact with each other. The interaction could either occur directly through a common grid or indirectly through smart mobile energy storage (e.g. AEVs). Within the above-described environment, a prosumer or an AEV can independently make decisions for optimization based on a certain objective that may dynamically change. This paper introduces a framework within which the energy supply of multiple prosumers individually and an AEV is autonomously optimized. The optimization is achieved by solving several Mixed Integer Linear Programming (MILP) problems and using Machine Learning (ML) algorithms to forecast energy generation and demand. The presented framework is then tested and evaluated by simulation in the Berlin area. The results show that both prosumers and AEVs can benefit from offering locally generated energy. Advantages can either be the reduction of power supply total expenses, time-saving or maximizing the use of renewable energy.