Analysis of radar sensor based object recognition methods for automotive use

OData support
Supervisor:
Horváth Bálint Péter
Department of Broadband Infocommunications and Electromagnetic Theory

Advanced Driver Assistance Systems for vehicular transportation are becoming a reality as the industry recently began to adapt to new customer requirements. The purpose of these systems is to create a more comfortable and safer driving experience. Well-known ADAS features are for example ABS (Antilock Braking System), ESP (Electronic Stability Program), and ACC (Adaptive Cruise Control). Since 2012 the first two have become a basic requirement in every car in the European Union.

The most important part of the operation is fusing the sensor information together provided by different sensors mounted all over the vehicle. The most important sensors are the ones, that map the area around the vehicle helping to detect surrounding objects and to provide location information of the vehicle.

The sensor data are processed by various algorithms running in parallel to the particular ADAS functions. Based on the results of the algorithms the Electronic Control Unit controls the actuators which in turn implements vehicular maneuvers.

In Chapter 1 a brief description of ADAS functions is presented. Chapter 2 and 3 introduces the reader to automotive sensors, and their applications, especially the radar systems. At the end of Chapter 3 a cost effective sensor configuration is discussed in detail. Chapter 4 introduces the reader to the fundamental algorithms used for object detection, and a Kalman filter based method is presented. Chapter 5 discusses the usage of the Kalman filter through multiple physical problems.

Chapter 6 discusses the realization of a camera-radar sensor fusion system. Moreover this Chapter also presents the basics of the image processing algorithms, and the environment for system verification, which is capable of simulating the sensor. Following the conclusion future development plans are discussed.

Downloads

Please sign in to download the files of this thesis.