Driver information systems help drivers to navigate efficiently through a r
oad network and avoid congestion. However, while a driver could save time b
y complying with a route switch recommendation, they might experience an in
crease in accident risk or cause other drivers to experience an increased r
isk. If this is the case, it may be worth considering the accident risk of
different routes as part of the route recommendation process. The integrati
on of accident predictive models within a traffic network model, designed t
o simulate route guidance, enabled these issues to be investigated.
Dynamic route guidance, based on a minimum journey time criterion, is estim
ated in a congested, urban network to result in accident reductions of a si
milar proportion to distance reductions, but significantly less than journe
y time reductions. The effect of considering accident risk in route recomme
ndations was investigated. Deterring drivers from minor roads or from high
risk junction crossings failed to produce significant accident reductions w
hen compared to route recommendations based on travel time alone.