Geolocation assisted routing protocols for vehicular networks

摘要

The class of flooding-based DTN routing protocols that leverage (transitive) encounter probabilities have been shown to perform well in selected simulations and scenarios, however they are especially sensitive to heterogeneous mobility models in which some nodes' mobility pattern is on a significantly different timescale than others. In particular, military and disaster response scenarios can exhibit abrupt topology changes. We analytically show that the worst-case inputs to these existing DTN routing algorithms can drastically reduce their performance. In light of such scenarios, we develop new protocols that inherit the benefits of existing schemes, while leveraging geographic assistance to enable faster recovery from abrupt topology changes.