A novel predictive link state indicator for ad-hoc networks

Abstract

Mobile Ad-hoc Networks (MANET) and more specifically their vehicular variant (VANET) have to deal with fast changing channel conditions, specifically in urban areas. Routing protocols that have to build end to end paths over such volatile links typically react to link failure. In this paper we present a novel PHY layer based link state indicator which aims to predict such failure. The proposed link state indicator is related to the IEEE 802.11 standard and relies on the OFDM decoding process. PhySimWifi is a detailed and accurate implementation of the OFDM-based IEEE 802.11 standard that provide access to all the steps of a packet reception and incorporates realistic channel models. When using it in the ns-3 simulator, the received packets decoding errors / success gives us the material to compute our predictive estimator. This new link state indicator is entirely based on the PHY level. The efficiency of the proposed indicator is validated by reference to PHY and NET packet reception ratio of the monitored link. The efficiency of the proposed new indicator is validated by comparing it with an Signal to Noise Ratio based predictive link state estimator.