Torque vectoring allows EVs with multiple motors to deliver precise power to individual wheels, enabling a level of control that combustion engine vehicles cannot match. But calculating and optimising that power while a vehicle is in motion is difficult and requires detailed knowledge of the driving conditions ahead as well as powerful onboard computers, making torque vectoring impractical for everyday consumer EVs.
Published in Vehicle System Dynamics, the Surrey study describes an onboard torque vectoring system that combines a predictive control model with fuzzy logic which adaptively prioritises either vehicle dynamics or energy efficiency, depending on road conditions. The team created a stability-control system that anticipates the curvature of the road ahead, allowing the car to pre-emptively brake when it approaches a bend too fast. According to the researchers, their model also simulated driving using the so-called 'pulse and glide' method to reduce energy consumption, which is beneficial in EVs and cost-effective enough to be widely rolled out. The work was part of the EU’s STEVE project, which has seen Surrey develop several novel approaches to EV torque vectoring.
"This has been an exciting project that has allowed us to make some major advances in powertrain control for electric vehicles,” said Professor Aldo Sorniotti, head of the Centre for Automotive Engineering at the University of Surrey.
“We believe that our work will allow new advanced torque vectoring techniques to become useable in ordinary electric vehicles, delivering research that will directly assist drivers in the very near future."
Riccardo Groppo, CEO of project partner Ideas & Motion, added: "It has been a pleasure working with the University of Surrey on the STEVE project. In particular, the technical collaboration was fundamental to making progress on the inverter and control algorithm for the Light Electric Vehicle we developed. I think we have accomplished excellent results, setting the basis for further collaboration.”
