Stochastic Control For Idle Speed Stability

Self-Tuning Regulators, based on both Minimum Variance Control theory and Recursive Extended Least Squares method, are applied to fuel injection/spark ignited automotive engines in order to improve idle speed stability. Simplified mathematical models, with consideration for stochastic combustion variation, are used to describe idle speed dynamics. Model parameters and control gains are calculated in every combustion cycle by using a 16-bit microcomputer. Fuel injection rate and alternator load manipulation are independently examined as control forces.

It is founded that (1) these techniques for cotrolling fuel injection rate and alternator load provide over 10% and 30% reduction of engine speed fluctuation, respectively, in comparison with the conventional control systems and (2) this system, in which the control gains are tuned to the appropriate levels, can operate stably in sudden changes of air flow rate and external load.