The European Commission has been actively supporting a number of safety related projects, as part of its initiative to pursue the goal of considerably reducing road accidents by about 50% until 2010.
PReVENT-ProFusion and other on-going Advanced Driver Assistance Systems (ADAS) projects have clearly identified some basic functional, performance and cost limits of current sensors and ADAS systems. These limits can be seen as breakthrough opportunities at hardware and software levels, helping in the fulfilment of applications objectives and in the effective deployment of advanced driver assistance systems. Some hardware aspects concern sensing technologies (for which a recent boost has been observed regarding vision devices), communication interfaces (data buses) and housing. Software aspects concern sensor data pre-processing and application interfaces (access to raw and pre-processed data for instance). PReVENT addresses mainly the software aspects.
Three main functional requirements, relevant to ADAS systems, will be addressed in ADOSE:
Reliable obstacle detection and classification, with reduced environmental influences and false positive rate at high recognition rates:
- Detection of pedestrians in urban and extra urban scenarios for warning night-vision applications, collision mitigation and avoidance of pedestrians at night.
- Detection of all road users including pedestrians in urban scenarios in the near field for pre-crash applications, collision mitigation and avoidance systems as well as emergency braking.
- Fast, precise and reliable measurements of both range and angle of closing objects with high resolution and wide field of view under all weather conditions (in Vulnerable Road Users & Collision Mitigation function).
- Visibility measurements (fog, mist) which are very important for the assistance task to the preventive safety functions.
Performance, robustness and reliability of ADAS can be improved through sensor enhancements at hardware and/or software levels, sensor data fusion and communication systems, and by introducing new sensor technologies.
Among the technological needs addressed to sensor enhancement there are:
- Sensors with wider field of view and higher spatial resolution (e.g. overcoming of present ranging camera with large area custom pixel and low resolution, CMOS imager with more than 1 Mpixels)
- Sensors with higher sensitivity and high dynamic range (e.g. CMOS imager with enhanced response in the near infrared band-NIR, ranging camera with higher fill factor)
- Sensors with higher temporal resolution and low-processing power (e.g. sensor for pre-crash application)
- Multifunctional sensors as "application independent" sensors
- New (friction, visibility, etc.) and improved (radar, vision, etc.) sensors