This article presents a multisensor system developed in order to avoid road crashes. The system deals with data issued from sensors placed on an experimental vehicle driving in a real environment. The aim of this project is to design a hardware and software system which can help to avoid accidents. The integration of the multisensor system aims at fusing different information sources in order to take safer decisions. A lot of heterogeneous and asynchronous data must be taken into consideration for the study and achievement of an on-board driving aid. This information given by different sensors passes through a Controller Area Network. The sensors limitations are taken into account and the system takes advantage of each sensor as well as possible. The sensor set allows the reconstruction of a local road map as accurately as possible. (A)
Samenvatting