Development of a detection road users system in vehicle A-pillar blind spots

102nd International Scientific and Technical Conference "Intelligent Systems of Driver Assistance: Development, Research, Certification"

运输工程;计算机科学

Beresnev, P.O.^1 ; Tumasov, A.V.^1 ; Zeziulin, D.V.^1 ; Porubov, D.M.^1 ; Orlov, L.N.^1

Nizhny Novgorod State Technical University N.a. R.E. Alekseev, Minin Str., 24, Nizhniy Novgorod

603950, Russia^1

关键词: Blind spots;    Existing systems;    Onboard computers;    Pedestrian recognition;    Road users;    Side view mirrors;    Structural elements;    Warning signals;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/386/1/012008/pdf
DOI  :  10.1088/1757-899X/386/1/012008

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

In connection with the increase in the number of cars in the world, there is an increase in the number of road accidents. The structural elements of the car do not allow the driver to assess fully the road situation. Such zones are called blind spots. Mostly the collision of cars happens in blind spots due to the lack of visibility. This is especially true for commercial vehicles. This article presents a system that allows you to expand the visibility of the car, due to the use of displays in the windshield pillars. A review of the existing systems was carried out. However, the use of only displays is not safe, for this, a pedestrian recognition system integrated in the windshield pillars based on neural networks can be used. The system allows detection, estimation of the distance to the object and giving out a warning signal about a possible collision. It consists of an on-board computer module, a video camera and an display. This system allows you to reduce the number of accidents near a pedestrian crossing or with a sudden pedestrian exit on the roadway. Interior elements have been designed and installed, allowing a good fitment of the displays into the windshield pillars. Exterior elements have been designed and installed in such way that allowed to keep the visibility of the car. The external side-view mirrors were replaced with video cameras. Experimental studies of the system were carried out. The pedestrian recognition system has demonstrated its performance by the results of the tests, however, it is necessary to debug the coefficients taking into account the approach time to the object.