Engine Speed Control using Online ANN for Vehicle with EMDAP-CVT

Abstract

Controlling engine speed corresponding to load variations and road condition has always been a challenge to automotive engineers. However, with the introduction of Electro-Mechanical Dual Acting Pulley Continuously Variable Transmission (EMDAP-CVT),maintaining constant engine speed based on either its optimum control line or maximum engine power characteristic could be made possible. This paper describes the simulation work in this area carried out by the Drivetrain Research Group at the Automotive Development Centre, Universiti Teknologi Malaysia, Skudai Johor. The developed drive train model is highly non-linear; it could not be controlled satisfactorily by common linear control strategy such as PID controller. To overcome the problem, the use of Artificial Neural Network (ANN) is employed to indirectly control the engine speed by adjusting pulley CVT ratio. Computer simulations showed that applying artificial neural network (ANN) into drive train model could select a proper transmission ratio where the engine could be maintained to run at certain desired engine speed.