Abstract
Nowadays present driver assistance systems are mainly focused on supporting the driver by warning him on risky situations (1), (2). Examples of these features are lane departure warning, blind spot detection, speed limit assistant or forward collision warning. Other systems, like adaptive cruise control, are still considered more a comfort function rather than a safety feature. Next steps are and will be done in the direction of not only warning but also acting over the vehicle behaviour (braking, steering…) to fully support the driver. Examples of first available functions are lane keeping assistance and collision mitigation by braking. Furthermore, data fusion techniques combined with by wire technologies and enhanced vehicle dynamics information, can provide a better understanding and estimation of the potential risk in order to move towards full collision avoidance concepts. This paper reviews enhancement possibilities for next generation ADAS using data fusion techniques. Although new radar, lidar and image sensors are also considered, special emphasis is placed in the use of enhanced digital maps and C2X technologies that can provide an extended horizon and field of view on the vehicle’s path that can be used. First, the possible improvements of present ADAS solutions and the new functions that data fusion algorithms can provide for a next generation of driver assistance systems are discussed. In this sense, also specific needs, challenges and requirements for full collision avoidance are analysed. Then, a possible data fusion architecture, that takes into account the usage of several sources of information (radar, image sensors, lidar, digital maps, C2X and ego-data) is introduced and described (see figure 1). Afterwards several CTAG research activities in the field of data fusion for next generation ADAS are explained introducing some of the results achieved with a dynamic driving simulator and a research vehicle. Finally next steps and main concludions are outlined.
KEYWORDS – Data Fusion, Digital Maps, C2X, Collision Avoidance, Driver Assistance Systems