An integrated trust-based secure routing with intrusion detection for mobile Ad Hoc network using adaptive snow geese optimization algorithm
V. Nivedita, Chin‐Shiuh Shieh, Mong‐Fong Horng
Abstract
Wireless Sensor Networks (WSN) and Mobile Ad Hoc Networks (MANET) are pivotal technologies widely used across various applications. However, the rise in wired and wireless technologies has also increased the frequency of attacks, compromising security, increasing packet loss, and reducing routing efficiency. Detecting denial-of-service (DoS) attacks remains a critical challenge, with issues in accuracy, scalability, and handling diverse attack methods. Existing methodologies face numerous challenges concerning the performance constraints of the detection system, the scalability and stability of the system, and the capacity to utilize extensive data effectively. To address these challenges, this research work proposes a cluster-based routing protocol integrated with a Stacked Convolutional Sequential Autoregressive Encoding Network (SCSAEN). The approach begins with density-based Adaptive Soft clustering (DAS) to maintain cluster stability during node mobility. The cluster head is selected using the Elk Herd Optimization (EHO) algorithm, which ensures resilience in dynamic MANET environments. The ASGO- TSPCPTrustNet algorithm performs in two ways: (i) Initially, the TrustSync Packet Control Protocol (TSPCP) computes the multi-attribute trust value to enhance network security; (ii) subsequently, the optimal route is determined utilizing the Adaptive Snow Geese Optimization Algorithm (ASGO). Additionally, SCSAEN-based intrusion detection is implemented to identify various attacks, including zero-day and DoS attacks. The performance of the proposed method is assessed using various metrics, including energy consumption, network lifetime, packet delivery ratio (PDR), attack detection rate, and computational time. The proposed method achieves high attack detection rates of 98 %, exhibits a high throughput of 99 Mbps, and consumes less energy at 0.6 mJ than the existing methods such as the Movement-Aware Routing Protocol based on Hybrid Optimization (MARP-HO), Fuzzy Chaotic Adaptive Particle Swarm Optimization (F-CAPSO), Epsilon Swarm Optimized Cluster Gradient using a deep belief classifier for multiple attack intrusion detection (ESOCG), improved heap optimization for intrusion detection models (IHO-MA), and the Multi-head Self-Attention based Gated Graph Convolutional Network (MSA-GCNN). The experimental findings demonstrate that the proposed method outperforms existing methods such as MARP-HO, F-CAPSO, ESOCG, IHO-MA, and MSA-GCNN in terms of energy efficiency, network lifetime, packet delivery ratio, and computational time.