Iterative learning air-fuel ratio control with adaptation in spark ignition engines

The paper addresses the problem of air-fuel ratio (AFR) stabilization applying learning techniques. The proposed strategy consists in iterative redesign of control coefficients based on the control obtained in a previous step in conjunction with an adaptive scheme. The interesting feature of the proposed solution is that the adaptive control attempts to match not the AFR dynamics but rather the error dynamics originated by the substitution of the feedforward control in the real system. Results of application are reported and discussed.