We analyzed traffic scenarios for determining accident dy namics and identifying when dangerous situations might occur.
The accident patterns that we addressed were similar to those
used by a safety system for identifying possible hazards in
traffic (e.g., as in [4] and [16]). Requirements on the exchange
of safety data, such as delivery latency, information content,
and size, were derived from these analyses. The spatial and
temporal relevance of data were also investigated, e.g., the set
of vehicles that have data of interest for a subject vehicle or
when certain information needs to be provided to a vehicle
to maximize the possibility of avoiding a collision.