Best Multiple Non-linear Model Factors for knock Engine (SI) by using ANFIS

Knock Prediction in vehicles is an ideal problem for non-linear regression to deal with, which use many of the factors of information to predict another factor. Training data were collected through a test engine for the Malaysian Proton company and in various states of speed. Selected six influential factors on the knocking (Throttle Position Sensor (TPS), Temperature (TEMP), Revolution Per Minute (RPM), (TORQUE), Ignition Timing (IGN), Acceleration Position (AC_POS)), has been taking data for this study and then applied to a single cylinder, output factor (output variable) to be prediction factor is a knock. We compare the performance of resultant ANFIS and Linear regression to obtain results shows effectiveness ANFIS, as well as three factors were selected from six non-linear factors to get the best model by using Adaptive Neuro-Fuzzy Inference System (ANFIS). Experiments demonstrate that although soft computing methods are somewhat of tolerant of inaccurate inputs, cleaned data results in more robust models for practical problems.