Predictive Driving Guidance of Full Electric Vehicles Using Particle Swarm Optimization

This paper focuses on the development of computational algorithms for determining online energy-based driving guidance for an electric vehicle (EV) endowed with regenerative breaking system capabilities. A predictive decision support system is designed to optimally distribute energy flow between the instantaneous power demand requested by the driver for the powertrain engine and the different auxiliaries relating to comfort performance, such as the heating system. The proposed methodology uses an online particle swarm optimization (PSO) algorithm to search for a global optimum relative to specific objective functions, which take into account battery autonomy, driving comfort indexes, and travel time. Our methodology has been validated for a heavy motorized quadricycle vehicle using hardware-in-loop (HIL) simulations, for which the energy management system has been implemented in a digital signal processing (DSP) board communicating through a controller area network (CAN) protocol.