Hybrid vehicles are characterized by a powertrain consisting of at least two energy conversion machinesworking in different energy domains. It also requires at least two different kinds of on-board energy storage. One possible, and by far the most pursued approach is the hybrid electric vehicle (HEV). Its powertrain consists of an internal combustion engine (ICE) and an electric machine (EM). Since two totally different machines working in different energy domains is being employed, one has to determine when to use one device over another.
One technique or operation strategy for doing this is load leveling. The idea in load leveling is to use the electric machine to force the ICE to operate at its peak efficiency at all times during the driving cycle. Artificial neural networks and fuzzy logic control are used to implement a load leveling strategy. The resulting vehicle control unit, a supervisory controller, coordinates the powertrain components. The controller has the ability to adapt to different drivers and driving cycles. This allows a control strategy which includes both fuel-economy and performance modes. The development was done for the Ohio State University FutureTruck Challenge project.