Improvement of Spark Ignition Knock Detector Performance by Learning Control

The improvement of the system response and the stability of the Knock Control System (KCS) by using Learning Control is discussed.

The learning control keeps the feedback results in the computer memory and changes the feedback gain according to the data in the memory.

An application method of the learning control to KCS is studied.

Improvement of the learning control by adding analogy and extinct functions are explained.

The new KCS with learning control is realized by a 16 bit micro-computer, and the test results of the new KCS on 1.8l vehicle are also described.