Temporal neuro-control of idle engine speed

Neural network architectures are proposed to model and control the spark-ignition (SI) engine idle speed. Static and dynamic multilayer neural networks are used to develop plant models and plant inverse function models. A neural network is trained to learn the system's input-output relationship. Another is trained to model the inverse relationship of the plant and is included in the controller. The developed controller, in series with appropriate filters, is then used to control the throttle angle and the spark advance angle control signals to track a desired speed and pressure of the engine. The paper demonstrates how computationally inexpensive neural networks can effectively learn online both the modeling and control tasks for this nonlinear, dynamical and complex process.