The reliability and safety of axle bearings in passenger and freight rolling stock has always been of concern to the rail industry. Wayside condition monitoring of in-service axle bearings has been based primarily on hot boxes. These instruments use expensive infrared sensors which can only detect problems once they have become critical and are prone to measurement errors. Moreover, due to their high cost hot boxes are usually spaced at relatively great distances from each other. Hence faulty axle bearings can seize at any time in between and result in a derailment. During tests in Long Marston using rolling stock with artificially induced axle bearing faults it has been proven that high-frequency acoustic emission is capable of detecting such faults regardless of their severity. The Ph.D. candidate during this project will develop a signal processing methodology to assess the signals acquired from acoustic emission sensors in order to enable the quantification of the severity of the faults. This would be a major breakthrough based on what has already been achieved. The successful outcome of the project will benefit profoundly Network Rail and other rail infrastructure managers who face significant challenges in ensuring the reliability of axle bearings of various types of rolling stock.