Cascaded Model Predictive Control for Shared Autonomous Electric Vehicles Systems with V2G Capabilities

Abstract

Shared autonomous electric vehicles (SAEVs) are being introduced in pilot programs and they are expected to be commercially available by the next decade. In this work, we propose a methodology for the joint optimisation of vehicle charging, vehicle-to-grid (V2G) services and fleet rebalancing in mobility systems using SAEVs. The proposed model is implemented as a cascaded model predictive control (MPC) optimisation framework with two different timescales. The first MPC scheme, called energy layer, abstracts the fleet of SAEVs as an aggregate storage system for the sake of model scalability, and it optimises fleet charging and V2G services to minimise electricity cost over a long timescale (hours). The second MPC scheme, called transport layer, optimises short-term vehicle routing and relocation decisions to minimise customers’ waiting times while taking into account the charging constraints derived from the energy layer. A case study using transport and electricity price data for the city of Tokyo is used to validate the model. Results demonstrate that our approach is computationally scalable and it can be applied to large-scale scenarios. In addition, it allows to significantly reduce charging costs with limited impact on passengers’ waiting times