Engine knock detection using multi-feature classification by means of non-linear mapping

The paper describes an advanced approach solving the knock detection task. It is based on a twolevel feature extraction solution followed by a neural detection step trained in a constructive supervised way. The knock detection is not only based on the popular cycle-by-cycle classification but a tendency index for the knock condition is determined. The experimental environment consisted of several measurement sessions in a special test cell. In order to achieve a most realistic knock database the test engine was mounted on an in-production car.