-
[show abstract]
[hide abstract]
ABSTRACT: Automotive pedestrian protection systems will be introduced in the EU in short term to reduce the number of accidents and injury fatalities. As with any safety issue, a comprehensive approach comprising both active and passive safety elements should be followed and it is also valid for pedestrian protection. Passive safety short term solutions can be contact sensor systems that trigger raisable engine hoods. This paper discusses an innovative approach for pedestrian detection and localization, by presenting a system based on two short range radars and an array of passive infrared thermopile sensors, aided with probabilistic techniques for detection improvement. The two short range radars are integrated in the front bumper of the test vehicle. They are able to observe and track multiple targets in the region of interest. However, one difficulty is to distinguish between pedestrians and other objects. Therefore, a second sensor system is required to classify pedestrians reliably. This system consists of spatial distributed thermopile sensors which measure the object presence within their respective field-of-view independently. These measurements are then validated and fused using a mathematical framework. Thermopiles are excellent to detect the thermal radiation emitted by every human. However, a robust signal-interpretation algorithm is mandatory. In this work a statistical approach combining Dempster-Shafer theory with occupancy-grid method is used to achieve reliable pedestrian detection. Thermopile and radar sensors use independent signature-generation phenomena to develop information about the identity of objects within the field of view. They derive object signatures from different physical processes and generally do not cause a false alarm on the same artifacts. The integration of the sensor readings from the radar and thermopile system is conducted using unifying sensor-level fusion architecture.
Information Fusion, 2005 8th International Conference on; 08/2005
-
[show abstract]
[hide abstract]
ABSTRACT: The growing interest toward active safety systems leads to a rapid development of different sensors and systems. One key aspect for all systems is the knowledge about the car environment. This knowledge is essential for safety applications such as PreCrash. A PreCrash system generates information about location, moving direction and relative velocity of critical objects in the car environment immediately before an imminent and inevitable accident happens. Different types of actuators can benefit from this information for in-time deployment and the choice of the right deployment energy. In this paper a PreCrash system is proposed, which uses two short range radars and a laser scanner to obtain the required environmental data. The short range radars and the laser scanner will be mounted in the front of a test vehicle. The sensors are able to observe multiple targets within a field of view in front of the car. To develop a highly reliable environment sensing system, competitive fusion approaches using different types of sensors are appropriate. The challenge in realizing such a fusion system is the optimal utilization of the detection performance of each sensor type and the developing of a fast architecture for using it in the automotive environment. As two examples for such a fusion strategy a commonly utilized decision level fusion approach is discussed and a newly developed grid fusion approach is introduced. It is based on a virtual grid in front of the car in which the sensor readings are mapped into. The architecture of the grid is directly adapted to the requirements of the application and the sensor specifications. The data extraction out of the grid is based on simple and well known strategies. Decision strategies to increase the certainty of object identity are also proposed in this paper.
Information Fusion, 2005 8th International Conference on; 08/2005
-
[show abstract]
[hide abstract]
ABSTRACT: Whenever addressing pedestrian related injury focal points in automotive accidentology, a comprehensive approach comprising both active and passive safety elements should be followed. Passive safety short term solutions can be contact sensor systems that trigger raisable engine hoods and an active safety element could be the brake assist. However, an important enabler for a future pedestrian protection system is a suitable, low-cost, environment-friendly sensing technology for pedestrian detection, supported by a fast and reliable algorithm for object localization. This paper discusses such an innovative approach for pedestrian detection and localization, by presenting a system based on an array of passive infrared thermopile sensors, aided with probabilistic techniques for detection improvement. The distributed thermopile sensors (sensor-array) detect the object presence within their respective field-of-view independently from each other. These measurements are then validated and fused using a mathematical framework. The focus of this paper is on the signal interpretation of the thermopile sensors. Since passive thermopile sensors are prone to background influences and can detect only the relative temperature changes, a robust signal-interpretation algorithm is essential. In this respect, a statistical approach combining Dempster-Shafer-theory with occupancy-grid method is used to achieve reliable pedestrian detection. The performance of the proposed approach is discussed by presenting some experimental results.
Intelligent Vehicles Symposium, 2005. Proceedings. IEEE; 07/2005
