EU Logo

6FP Logo

EUCAR Logo

The SAFESPOT Project is co-funded by the European Commission Information Society and Media and supported by EUCAR.

SP2 - INFRASENS

SP2

The challenge faced by INFRASENS was to extend the role of the infrastructure in the area of preventive road safety by integrating information gathered by roadside sensors with information from the vehicles themselves, considered in this context as ‘mobile sensors’.

The response has been to create an Infrastructure-based Platform which is complementary to the Vehicle-based Platform, SAFEPROBE, so that together they permit a cooperative approach to the reduction of accidents. It was intended that the Platform runs in roadside units installed at critical parts of the road network ("black spots").

An important objective was to define an architecture which would be flexible enough to be adaptable to different types of safety applications, and hence different data needs, and also suitable for different road environments.

Central to the platform is the Data Fusion Module which undertakes all the necessary data processing to be able to produce high quality and reliable data. This is 'written' on the Local Dynamic Map, i.e. the dynamic data base, in the roadside unit, so that it is available for the SAFESPOT Applications.

The Data Fusion is made up of two main processes: Object Refinement and Situation Refinement. While the aim of the former is to increase the level of detail and accuracy regarding a specific detected object (e.g. an obstacle or a vehicle), the latter consists of a set of modules whose output helps to provide a more precise picture of the driving context: e.g. weather conditions (e.g. visibility levels), traffic events, and vehicle trajectories.

The implementation of these modules involved developing new algorithms able to increase the accuracy of interpretation of data received from roadside and vehicle-based sensors. This also included the development of techniques for fusing non-homogeneous data from a variety of sources (i.e. fixed and mobile sources, conventional and non conventional measurements).

A fundamental feature of the Data Fusion Module was the specification of a common interface which permitted the input of data from various sources, the most important being:

  • Roadside sensing systems which provide critical information on safety-related events or conditions involving the driving environment;
  • SAFESPOT-equipped vehicles which communicate data from onboard sensing systems via the VANET (ad hoc vehicle network).

It is also possible to receive data from external sources through the same interface, such as Traffic Control Systems or Traffic Management Centres.

The roadside sensing systems which - for experimental purposes - have been integrated within the INFRASENS platform, include an innovative network of micro-sensors as well as more conventional systems. In combination with the other data sources, their role is to extend the drivers’ 'vision'.

All of these systems include pre-processing modules to convert the raw data into useful measures for the Data Fusion Modules.

THE SENSING SYSTEMS

A roadside-mounted laserscanner is used to detect the position of road users at dangerous intersections:

Roadside mounted laserscanner

A CCTV camera system is used to trace vehicle movements and register their speed:

CCTV camera system

A CCTV camera system is used to detect the presence of fog or rain and estimate the visibility range:

CCTV camera system

A passive RFID system (with readers on the roadside) is used to detect passing vehicles and their driving direction (e.g. ghost drivers):

Passive RFID system

A CCTV camera system is used for ice detection by means of image filtering:

CCTV camera system

An Infrared camera is used for animal and pedestrian detection:

Infrared camera

A Wireless Sensor Network offers a low-cost, low-maintenance solution to detect vehicle passage and direction. The traffic-related information can then be used for generating safety warnings at 'black spots' on the road network.

Wireless Sensor Network