Comparative study on different balancing conditions of an air filled tyre using statistical features and classification via regression algorithm

Tyre condition monitoring systems (TCMS) are the safety systems used in a vehicle for measuring the condition of tyre like tyre pressure, temperature, balancing etc. In this era of increasing vehicle accidents, these systems are having paramount importance in terms of safety. The current technology TCMS uses direct sensors like pressure sensors or wheel speed sensors etc, which are highly expensive. This paper puts forward an innovative indirect TCMS system using condition monitoring techniques and machine learning. For different five balancing conditions vertical vibrations from fixed wheel hub were extracted from a moving air filled tyre with the help of an accelerometer. The tyres filled with air are considered with different pressure values to represent puncture, normal, idle and high pressure conditions. In feature extraction process, the statistical features were extracted from the acquired signals and the prominent features were selected using J48 algorithm. Selected features were classified with the help of classification via regression algorithm and reasonable high accuracy was obtained. This paper attempts to study the effect of unbalance of the wheel on the classification accuracy of an indirect TPMS system. The results are compared and presented