AI-cloud-based Vehicle Management Strategies for Electrified Vehicles

The commitments to reducing greenhouse gas (GHG) emissions and fossil energy consumption are driving the world to leverage sustainable vehicle and transportation techniques. Electrified vehicles (EVs), such as plug-in hybrid EVs (PHEVs) and battery EVs are the current panacea to renovate the transport sector for a sustainable vision. PHEVs represent a judicious solution to achieve sustainability but without sacrificing the convenience of end-users, as corroborated by market data. The worldwide EV sales have seen an exponential increase since the recent decade, in which PHEVs take an important part, and the sales in 2019 noticeably reach 2.2 million.

The hybrid configuration of two power sources means some trade-offs need to be made so as to fully exploit PHEV’s merits, i.e., the improved energy efficiency and reduced emissions. It is worth mentioning that these trade-offs also exist in other hybrid configurations, such as battery-fuel cell hybrid EVs and those with different battery types. Customised vehicle management strategies are required in order to fully exploit EV merits. This proposal aims to design a two-tier energy control strategy (ECS) leveraging machine learning and Vehicle-to-Cloud (V2C), and an energy-efficient and robust velocity planning scheme via Vehicle-to-Infrastructure (V2I). The proposed ECS aims to optimise the power demand allocation in a real-time and blended manner for all trips to achieve energy efficiency. The velocity planning scheme aims at design energy-efficient driving velocity profile that can accommodate traffic dynamics by leveraging deep learning and V2I. The joint optimisation of ECS and velocity planning will be developed to reduce energy consumption during a trip cooperatively. The proposed management strategies target to collectively achieve trip-wise and real-time energy management in real travel contexts for improving the energy efficiency of EVs. The outcomes are expected to reduce energy consumption and emissions, and improve battery lifetime and user satisfaction through intelligent controls without adding extra expenses.

The project is funded by Chalmers AI Research Center (CHAIR) for two years with a total amount of 1.372 million SEK. To fully address the proposed objectives, we establish an interdisciplinary team including academic experts and industrial partners (CEVT and Volvo Cars) in the areas of automatic control, machine learning, driving behaviour modelling, traffic modelling, and automotive engineering.

Partner organizations

• Volvo Cars (Private, Sweden)
• China-Euro Vehicle Technology (CEVT) AB (Private, Sweden)