BMUDF: Hybrid Bio-inspired Model for fault-aware UAV routing using Destination-aware Fan shaped clustering

Abstract

Routing data between unmanned aerial vehicles (UAVs) involves identifying node locations, analyzing residual energy levels, evaluating temporal throughput and packet delivery performance, and identifying other network and node parameters. It assists in forming quality of service (QoS) aware routes. Existing routing models require large data samples to find the optimized path or are highly complex, increasing their computational requirements. Low-complexity models showcase low-performance routing QoS when deployed on large-scale networks. To solve these limitations, a hybrid bioinspired model is proposed for fault-aware UAV routing that uses destination awareness with a fan-shaped clustering process (BMUDF). The model initially collects data from different UAV nodes and their node-level and network-level constraints. These parameters are processed through the particle swarm optimization (PSO) model, which enforces fan-shaped clustering (FSC) for effective routing operations. Our scheme uses the PSO model to identify the initial routing paths cascaded with a genetic algorithm (GA) based destination-aware routing model. These routing paths are evaluated through the QoS matrices like maximum temporal throughput, packet delivery ratio (PDR), delay, and energy consumption. A grey wolf optimization (GWO) model further scrutinizes this routing performance, integrating fault tolerance and route optimization during continuous operations. The GWO model evaluates a trust-based fitness function, which helps to identify faulty nodes, and reconfigures the network with non-faulty nodes to improve its QoS performance under node failures and faults. Integrating the bioinspired models into the proposed system maximizes performance under different network scenarios. This performance compares and validated with an analysis of variance (ANOVA) test with various state-of-the-art models. It is observed that the proposed model showcased an 8.3% lower routing delay, 5.9% lower energy consumption, 1.5% higher packet delivery ratio, and 9.1% higher throughput, which makes it useful for a wide variety of real-time UAV routing scenarios.