On Partial Adoption of Vehicle-to-Vehicle Communication: When Should Cars Warn Each Other of Hazards?

Abstract: The emerging technology of Vehicle-to-Vehicle (V2V) communication over vehicular ad hoc networks promises to improve road safety by allowing vehicles to autonomously warn each other of road hazards. However, research on other transportation information systems has shown that informing only a subset of drivers of road conditions may have a perverse effect of increasing congestion. In the context of a simple (yet novel) model of V2V hazard information sharing, we ask whether partial adoption of this technology can similarly lead to undesirable outcomes. In our model, drivers individually choose how recklessly to behave as a function of information received from other V2V-enabled cars, and the resulting aggregate behavior influences the likelihood of accidents (and thus the information propagated by the vehicular network). We fully characterize the game-theoretic equilibria of this model. Our model indicates that for a wide range of our parameter space, V2V information sharing surprisingly increases the equilibrium frequency of accidents relative to no V2V information sharing, and that it may increase equilibrium social cost as well.