State-of-charge estimation for electric scooters by using learning mechanisms

Because of its nonlinear discharge characteristics, the residual electric energy of a battery remains an open problem. As a result, the reliability of electric scooters or electric vehicles is lacking. To alleviate this problem and enhance the capabilities of present electric scooters or vehicles, we propose a state-of-charge learning system that can provide more accurate information about the state-of-chArge or residual capacity when a battery discharges under dynamic conditions. The proposed system is implemented by learning controllers, fuzzy neural networks, and cerebellar-model-articulation-controller networks, which can estimate and predict nonlinear characteristics of the energy consumption of a battery. With this learning system, not only could it give an estimate of how much residual battery power is available, but it also could provide users with more useful information such as an estimated traveling distance at a given speed and the maximum allowable speed to guarantee safe arrival at the destination.