An instrumental variable approach to non-linear model-based adaptive control of engine speed

An instrumental variable approach to non-linear model-based adaptive control of engine speed regulation was investigated and implemented on a spark ignition internal combustion engine. A four-step version of instrumental variable parameter estimation is used to identify a bias-free and noise tolerant model of the engine dynamics from the by-pass air valve voltage input to the output engine speed. The parametric model representing the engine dynamics is a truncated Volterra series structure. Model-based adaptive control is accomplished through a partitioned non-linear model inversion of the engine model. The desired throttle response and disturbance rejection dynamics are introduced via a two-degree-of-freedom control structure. Performance of the non-linear model-based adaptive control algorithm has been verified experimentally.