A precrash system based on radar for automotive applications

In present-day cars, safety systems like airbag or seat belt pretensioner reduces accident consequences for passengers. Future generations of cars will support additional safety functions such as precrash, collision warning and broke assistant. For safety functions like precrash the timely deployment of seat belt pretensioner and some other safety systems is important, but is a very difficult problem. In this paper, we describe a Precrash-system for cars, based on 4 short-range radars (24GHz) and the associated signal processing, which provides information about a crash relevant situation. The system generates the information immediately before an imminent (inevitable) accident happens. For crash-relevant situation, it provides information about the location, distance and relative velocity of the cash-relevant object as exact as possible to achieve an optimum effect. Other important requirements for the signal processing task were very low computation time and very low memory consumption. Therefore a special basic approach for signal processing, based on classification, was realized and will be introduced.