The autonomous emergency brake (AEB) is an active safety system in vehicles, which aims to reduce the severity of the rear-end collisions. An AEB uses distance measurement sensors to detect an emergency situation and performs full brake when an accident becomes unavoidable. In this study, an autonomous emergency braking system is presented based on the combination of information measured by a laser sensor, an ultrasonic sensor and an infrared sensor. This thesis presents a new approach based on the arithmetic mean error and priority. The system calculates safety indicators such as time-to-collision, time headway, and enhanced time-to-collision. This system is implemented on a car model with promising results. Fuzzy decision-making algorithms are designed to apply appropriate accelerations. The test results show that the enhanced time-to-collision is more precise than other indicators, but this indicator is not enough, as it is not defined when the relative speed between the two cars is constant. So, it should be fused with other safty indicators such as time-to-collision.

Keywords: collision detection, autonomous emergency brake, sensor fusion, decision making algorithms.