Galileo / EGNOS Enhanced Driver Assistance
Background & Objectives
In the 1990s, the number of accidents with injured persons stagnated at about 380 000 cases per year. The improvement of passive safety systems in vehicles thereby just compensated the increasing traffic volume. However, the decreasing number of traffic casualties since the year 2000 can be mainly traced to the increasing market penetration of active assistant systems like ESP. But the accident rate on inner-city roads decreased only by 1.6% p.a., whereas the number of accidents on country roads declined by around 4% p.a.
This might be explained by the fact that ESP and other current assistant systems like speed-control radar, adaptive cruise control or brake assistant systems develop their maximum effect at higher speed levels and are laid out to focus on the driving tube. The number
of available assistant systems for inner-city driving is, however, not yet noteworthy.
A similar development can be observed for the number of traffic fatalities. Country-road users profit continuously from improved passive safety systems, whereas the positive trend has continued in inner-city areas since 1998 at only a diminished level because accidents in those areas often involve vulnerable road users like bicyclists or pedestrians; 94% of all traffic casualties within cities involve pedestrians. Therefore the need for action to reduce accidents in inner-city areas is obvious.
There are several algorithms for data fusion of information from different sources and even heterogeneous sensor networks. Data fusion works so long as the underlying assumptions are given (e.g. zero-mean, white noise). As soon as the assumptions are not valid, data fusion may yield worse results than using just the ‘correct’ sensor data. This is expected in urban environment data due to:
- satellite navigation suffering from bad visibility and multi-path effects;
- maps may be outdated;
- object and lane detecting sensors may be miscalibrated.
The main challenge then is to differentiate valid from invalid data to achieve the best result for the environmental model to be used for safety applications:
- Ego-Positioning: Satellite position and vehicle odometer can be fused to a vehicle state. These sensors should enable vehicle autonomous integrity monitoring.
- object detection/map: If stationary objects (e.g. poles) are available from map data, these landmarks should be verified by appropriate sensors. If this works for a period and then no further matches are found, the map seems to be out-dated.
- object detection/calibration: The innovation between landmarks from map data and landmarks from sensor data should be zero-mean, white noise. If this is not the case over a period, the sensor seems to be miscalibrated. Estimating the error may enable online calibration.
GENEVA focuses on high-precision road telematics applications. The project concentrates on applications with a social and public dimension, and with a high level of innovation i.e. highly Advanced Driver Assistance Systems (ADAS).
The general objective of the GENEVA project is to develop an innovative application within the context of advanced driver assistance for high precision, reliable and certifiable use. The application will contribute to the adoption of EGNOS/EDAS and the introduction of Galileo in one of the most important market segments i.e. the European automotive industry.
The application ‘Urban Assist’ is targeted to road telematics applications. Urban Assist will bring benefits, e.g. increased road safety to all European citizens, and has a broad social and public dimension. Sophisticated technology with a high level of innovation is needed for the implementation of the application.
Work performed & results
The output of GENEVA will be an application with clear market implementation focus using GNSS as a primary positioning technology.
Urban Assist aims to significantly reduce the number of urban road accidents and consequently the number of injured and killed people in road traffic. Besides this main target, collision avoidance has positive influence on other aspects important for the community. It helps to reduce traffic jams, resource consumption, air pollution and CO2emissions.
Improved safety for all road users shall be enabled through suitable, inexpensive sensors with short time-to-market and safety-relevant driver assistance systems for collision avoidance and collision mitigation.
The key topics therefore are:
- Driver support in standard situations to improve safety before or when entering a critical situation;
- Development of a scalable, inexpensive high-accuracy positioning system for urban environment with built-in integrity monitoring;
- Enriched existing map technologies for attributes needed for safety-relevant driver-assistance systems in urban environment and their standardisation;
- Development of algorithms to verify attributed maps with sensors for environmental perception like camera/image processing and laser/radar;
- Development of algorithms to verify calibration of sensors for environment perception.