Integration of navigation data into driver assistance system

OData support
Supervisor:
Bányász Gábor
Department of Automation and Applied Informatics

More than 31000 fatal road accidents happened in Europe in 2010. This is a huge number, so car manufactures are placing big emphasis on reducing this number. Besides passive safety systems there are increasing number of active safety elements in cars. While passive safety systems reduce the aftermath of accidents, active safety systems try to avoid them, hereby to minimize human casualties and financial losses. For example, the camera based active safety systems are capable of keeping lanes, can do emergency braking and are able to display traffic signs. The problem with traffic sign displaying functions is that not every sign is detectable for technological and infrastructural reasons. This may cause invalid speed limit displayed to the driver. The system performance can be increased by the fusion with navigation. This can be realized with a protocol called ADASIS, which is a method to send the electronic horizon to the other electronic control units via CAN bus of the vehicles. The electronic horizon is a model containing the attributes of the car’s movement and the map attributes around the car.

The aim of my thesis is to develop a software module, which is capable of reconstructing the electronic horizon. While writing the thesis I deepened my knowledge of the CAN communication and the ADASISv2 protocol. I had to make an existing recontructor better and more optimized. To ensure testability, I also had to interface it with a visualization, which is a map containing the position of vehicles and other text data. The visualization is realized in Qt.

The development of the reconstructor was in C++ programming language. After planning and remaking, in the beginning I did the tests with simulation files. Then I used recorded navigation data to test. Finally, I recorded a video and CAN data stream in a car and I used it to verify the correct operation of my software module.

Downloads

Please sign in to download the files of this thesis.