Accelerating transport electrification by machine learning

The technological blossom in artificial intelligence (AI) makes possible numerous advancements in various engineering disciplines. For this project, we intend to use the AI expertise to examine the role that AI technologies can play in accelerating transport electrification, and subsequently contributing to climate action. In view of the strong vehicle industry in Gothenburg, Sweden, Chalmers and Scania AB collectively provide an outstanding research environment for this important topic. More specifically, this project will consolidate data-driven electrochemical battery model, deep reinforcement learning based automated electric vehicle (AEV) speed control and AEV personalized routing recommendation model, which are three biggest challenges in the transition process towards electrified transport systems. The project will be co-supported by AI Innovation of Sweden, and Drive Sweden, with a hope to implement the expected findings in not only the vehicle industry but also transport management sectors. This research is interdisciplinary in nature, and requires expertise from computer science, artificial intelligence, electrochemical engineering, electrical engineering, and transport engineering.  In this project, a new problem is defined, i.e., personalized destination prediction and route recommendation for AEV users. A new notion will also be defined, i.e., the estimated energy consumption of arrival (EECA), inspired by the function of the estimated time of arrival (ETA) in navigation applications. The aims of this proposal are: 1) To extend battery lifetime by controlling operations within its physical limits with an optimal safety bound,
thereby improving the energy utilization efficiency of the battery; 2) To save energy consumption by building a personalized AEV recommendation model utilizing emerging techniques in artificial intelligence, thereby recommending routes with lower energy cost to users; 3) To improve the efficiency of AEV and the user experience by consolidating the electrochemical battery model (using the data-driven method), AEV speed control (using deep reinforcement learning) and AEV recommendation model in real-time, thereby contributing to safer, faster and greener transport.

The project is jointly sponsored by EU and Chalmers foundation, and the total budget is 400,000 Euros.