Closed loop mixture control using learning data resettable for fuel

Abstract

A closed loop mixture control system for internal combustion engines is responsive to a signal derived from an exhaust gas sensor. The sensor signal is time-integrated in a direction depending on the level of the gas sensor output to derive a first mixture corrective setting of the control system. Second corrective settings or learning data are established for the control system in correspondence with the amount of air supplied to the engine. Each of the latter settings is varied as a function of time in a direction depending on the value of the time-varying first corrective setting relative to a reference so that the second settings are automatically updated to meet varying engine performance such as aging. One of the second corrective settings is selected in response to the detected quantity of the supplied air and multiplied by the first corrective setting to correct the basic mixture control setting of the system toward an optimum value. All of the second corrective settings are reset to appropriate values, for example, '1' at the instant the engine is started if an average value of the second settings is greater than a predetermined value to compensate for different fuel vaporizations.