A Fault-Tolerant Routing Strategy in Wireless Mobile Ad-Hoc Networks

This paper introduces a novel methodology for routing in wireless ad-hoc networks through the implementation of an Enhanced Neural Gas Algorithm, which is further complemented by Competitive Hebbian Learning (CHL). These networks are susceptible to node failures and face distinct challenges related to interruptions in network connectivity, necessitating the need for fault tolerance to ensure resilience and reliability. We establish an implicit neighborhood ranking system that enables only the most locally optimal nodes to participate in routing decisions. Additionally, we propose an enhanced 2-hop acknowledgment protocol that utilizes beacon timestamp values to precisely evaluate and identify the most efficient routing paths, prioritizing nodes based on the promptness of their responses. We also incorporate a Trust and Reputation Mechanism that employs both Ranked Neighborhood and Location Information, thus enhancing the dependability of the data exchanged between participating nodes. In addition, we optimize the generated routing paths through a consensus learning approach. Our experiments show promising results on various performance metrics against some established routing protocols.