Advanced Vehicle Collision Avoidance using GNSS, Digital Maps and V2V Communication

The thesis is about an advanced collision avoidance system (CAS) which specifically makes use of
GNSS, digital maps and V2V communication additionally to “conventional” sensors such as radar
and camera. Current prototypes of CAS – which only use those “conventional” sensors for the
detection of surrounding vehicles – are restricted due to the limited detection area of the sensors
and their operational capabilities under various environmental conditions. By means of a sensor
data fusion, the advantages of all the above mentioned sensors are combined in order to increase
the reliability of the CAS. With knowledge of the course of the road ahead and with information of
the surrounding vehicles, a predictive system can be developed.
The thesis approaches the topic systematically by starting with a review of the state of the art
regarding collision avoidance systems. Based on the analysis of the literature, research needs are
identified and the research questions of the thesis are derived. In addition, requirements are
defined.
The aim of the CAS is to detect surrounding vehicles that are on collision course with the ego
vehicle and to initiate a collision avoidance action (braking or steering) automatically in case that
the driver does not react on time. The collision avoidance system is designed by braking it down
into different modules:
- Data fusion
- Motion prediction
- De-escalation and intervention decision
- Trajectory following control
The development modules take into account the research questions and defined requirements. The
algorithms are implemented in a simulation environment and in a test vehicle. Well-defined
scenarios are conducted to answer the research questions and to analyze the test results.