A Multi-Layered Assessment System for Trustworthiness Enhancement and Reliability for Industrial Wireless Sensor Networks

Abstract

The decision-making process in Industrial Wireless Sensor Networks heavily relies on the information provided by smart sensors. Ensuring the trustworthiness of these sensors is essential to prolong the lifetime of the network. Additionally, dependable data transmission by sensor nodes is crucial for effective decision-making. Trust management approaches play a vital role in safeguarding industrial sensor networks from internal threats, enhancing security, dependability, and network resilience. However, existing trust management schemes often focus solely on communication behaviour to calculate trust values, potentially leading to incorrect decisions amidst prevalent malicious attacks. Moreover, these schemes often fail to meet the resource and dependability requirements of IWSNs. To address these limitations, this paper proposes a novel hybrid Trust Management Scheme called the Multi-layered Assessment System for Trustworthiness Enhancement and Reliability (MASTER). The MASTER scheme employs a clustering approach within a hybrid architecture to reduce communication overhead, effectively detecting and mitigating various adversarial attacks such as Sybil, Blackhole, Ballot stuffing, and On–off attacks with minimal overheads. This multifactor trust scheme integrates both communication-based trust and data-based trust during trust estimation, aiming to improve the lifetime of industrial sensor networks. Furthermore, the proposed MASTER scheme utilizes a flexible weighting scheme that assigns more weight to recent interactions during both direct and recommendation (indirect) trust evaluation. This approach ensures robust and precise trust values tailored to the specific network scenario. To efficiently process and glean insights from dispersed data, machine learning algorithms are employed, offering a suitable solution. Experimental results demonstrate the superior performance of the MASTER scheme in several key metrics compared to recent trust models. For instance, when 30% of malicious Sensor Nodes (SNs) exist in a network comprising 500 sensor nodes, the MASTER scheme achieves a malicious behaviour detection rate of 97%, surpassing the rates of other models. Even after the occurrence of malicious SNs exceeding 30%, the False Negative Rate (FNR) in the MASTER scheme remains lower than other models due to adaptive trust functions employed at each level. With 50% malicious SNs in the network, the MASTER scheme achieves a malicious behaviour detection accuracy of 91%, outperforming alternative models. Moreover, the average energy consumption of SNs in the MASTER scheme is significantly lower compared to other schemes, owing to its elimination of unnecessary transactions through clustered topology utilization. Specifically, with 30% and 50% malicious SNs in the network, the MASTER scheme achieves throughput rates of 150 kbps and 108 kbps, respectively, demonstrating its efficiency in challenging network scenarios. Overall, the proposed MASTER scheme offers a comprehensive solution for enhancing security, trustworthiness, and collaboration among sensor nodes in IWSNs, while achieving superior performance in various metrics compared to existing trust models.