Development of frequency modulated continuous wave (FMCW) radar based shape identification system

In the recent years, there has been growing interest on radar technology in the automotive industry. Autonomous Driving Assistance System (ADAS) is becoming inevitable in automotive industry because of higher safety demand in road traffic and Frequency Modulated Continupus Wave (FMCW) radar sensors are commonly used in ADAS. There is a strong relation between ADAS technology and identfying shape of target because the more accurate prediction of obstacles, the more robustness in driving assistance systems. This thesis investigates the use of millimeter wave band FMCW radar for shape identification. In the past years, machine learning has earned great interest in various scientific areas and object detection problem is not one exception. Unlike other studies in the literature, proposed identification method uses machine learning based classifier and spectrum characteristics extracted from radar signature for determining object's shape. Developed system is tested with objects placed behind black cover and approximately % 98 identification accuracy is achieved.