Matching and decision for vehicle tracking in road situation

We present the development of a multi-objects matching algorithm with ambiguity removal entering into the design of a dynamic perception system for intelligent vehicles. The originality of this system lies in the use of theories such as fuzzy mathematics and belief theory which allow the handling of inaccurate us well as uncertain information Moreover, these theories allow both numeric and symbolic data fusion. We started from the hypothesis that we have some sensors providing redundant information in time. We develop in the article the problem of matching between the prediction (known objects) and the perception result (perceived objects). This makes it possible to update a dynamic environment map for a vehicle. The belief theory will enable us to quantify association belief on each perceived and known object. Some conflicts can appear in the case of object appearance or disappearance or in the case of a bad perception or a confused situation. These conflicts are removed or solved using an assignment algorithm, giving a solution called the "best" and so ensuring multi-objects tracking.