Applying Machine Learning to Reduce Overhead in DTN Vehicular Networks

• DOI: 10.1109/sbrc.2014.12
• Publication Date: 2014-10-22T21:17:11Z
• Authors: Lourdes P. Portug...; Cesar A.C. Marcondes; Hermes Senger; Luciana Arantes
• Keywords: Delay-tolerant networking; Routing algorithm

VANETs benefit from Delay Tolerant Networks (DTNs) routing algorithms when connectivity is intermittent because of the fast movement vehicles. Multi-copy DTN spread message copies to increase delivery probability but increasing network overhead. In this work we apply machine learning reduce overhead by discriminating worst intermediate nodes for transmission copies. The scenario a VANET public buses that follow specific routes and schedules. This repetitive behavior creates an opportunity applying trained classifiers predict occurrence performance-related events. As main contribution, our method decreases without degrading probability.