Antenna and Channel-Aware Fault-Tolerant Communication Architecture for Safety-Critical Vehicular Ad Hoc Networks

Abstract

The intelligent transportation systems that rely on Vehicular Ad Hoc Networks (VANETs) as their communication backbone data safety-critical applications require fault-tolerant, low-latency data communication; the failure of which can result in fatal accidents and disastrous traffic jams [4]. Nevertheless, the frequent connectivity and disconnectivity, and changing topology, high mobility and harsh wireless environments in VANETs are major cause of communication breakdown, whereas the classical protocols are not suitable to support the high reliability demands. In the given paper, a new communication architecture, Antenna and Channel-Aware Fault-Tolerant Communication Architecture (ACFTCA) is introduced which increases the reliability, adaptability and fault resilience of communication between vehicles. The offered architecture is based on the exploitation of two-layer approach where the routing mechanism is based on adaptive antenna diversity and the channel state. In particular, a real-time Channel State Information (CSI) feedback component is employed that monitors the quality of the links, foresees degradation and pre-eminently switches antennas or re-routes the transferred messages to keep communication alive. It uses 2 x 2 MIMO antenna system that is capable of beamforming, in order to attain a better-spatial diversity and additional signal robustness to the mobile-induced fading. Simultaneously, a hybrid fault detection scheme is established that incorporates anomaly detection at signal level and redundancy-based data forwarding to recover instantly in the event of the sudden degradation of a link, or a node failure. In contrast to the prior work, ACFTCA incorporates physical-layer awareness into network layer fault detection and recovery so as to achieve cross-layer tolerance to faults. The system is simulated by co-simulation based on the NS-3 simulator of network behaviour and the MATLAB simulator of physical-layer modelling with realistic mobility tracing and fading of channels. Performance measures, including packet delivery ratio (PDR), end-to-end delay, the link recovery time and communications reliability are examined with different densities of a vehicular network and mobility patterns. Outcome is that, ACFTCA provides up to fifteen percent of improvement in PDR and used to decrease fault recovery duration by sixty percent when compared against baseline fault-tolerant protocols. The architecture is also found to be scalable and with little overhead thus can be used in next generation vehicular environment where it is warranted especially where continuous and robust communication support is needed between autonomous and connected vehicles.