<p>Unmanned Aerial Vehicles (UAVs) are dramatically rising in terms of border surveillance, but with a low battery capacity, significant communication overhead, and prone to node compromise, it is hard to apply it to large scale deployments. This article presents a dual layer secure and efficient drone-based surveillance system that incorporates multi-cluster communication, Fog Computing (FC), and trust-based authentication. A Dual Phase Adaptive Fusion Intelligence Optimizer (DPAFIO) is proposed to find out Cluster Heads (CHs) based on residual energy, trust score and positional efficiency, and to evenly distribute the work load and stabilize the cluster formation. Each cluster is provided with a load-adaptive group of cooperative CHs, serving to distribute workloads, provide redundancy, and failover in a short time, and member drones transmit to the nearest CH to minimize range of transmission and energy loss. A Fog layer, between the cloud and the drone network carries out the in-network aggregation and early-decision-making that decreases the latency and the transmission over long distances. Dynamic trust score-based validation mechanism is used in order to identify the compromised nodes before passing on the data. The results of the simulation indicate that the proposed framework is efficient and stable. The evaluation is based on 30 independent simulation runs across 50 rounds. The average energy consumed is 0.8133 <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\pm \)</EquationSource> </InlineEquation> 0.0263 J/round. This comprises data relay, authentication, control signalling and fog processing. The end-to-end latency is 10.50 ms. The packet reduction rate is 94.99 <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\pm \)</EquationSource> </InlineEquation> 0.014%. The attack detection accuracy is 99.97 <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(\pm \)</EquationSource> </InlineEquation> 0.01%. This is analyzed based on five adversarial models, which are Sybil, collusion, wormhole, replay and packet-drop. The Packet Delivery Ratio (PDR) is 87.48% <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(\pm \)</EquationSource> </InlineEquation> 0.005%. The failover recovery time is 0.00009 <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(\pm \)</EquationSource> </InlineEquation> 0.000020 s. Convergence fitness of DPAFIO is 0.97533 <InlineEquation ID="IEq6"> <EquationSource Format="TEX">\(\pm \)</EquationSource> </InlineEquation> 0.00055. All results are reported as mean <InlineEquation ID="IEq7"> <EquationSource Format="TEX">\(\pm \)</EquationSource> </InlineEquation> 95% confidence interval. T-distribution is calculated on 30 trials to produce the confidence intervals.</p>

A secure and energy-efficient fog-integrated UAV surveillance framework using IntelliGuard and DPAFIO

期刊论文
出版模式：

开放获取
发表日期：

2026年6月24日

查看全文

Dhanvanth Kumar Gude,
Vamshi Krishna Raavi,
Mohit Lalit,
Anurag Jain,
Bhupesh Kumar Dewangan

摘要

Unmanned Aerial Vehicles (UAVs) are dramatically rising in terms of border surveillance, but with a low battery capacity, significant communication overhead, and prone to node compromise, it is hard to apply it to large scale deployments. This article presents a dual layer secure and efficient drone-based surveillance system that incorporates multi-cluster communication, Fog Computing (FC), and trust-based authentication. A Dual Phase Adaptive Fusion Intelligence Optimizer (DPAFIO) is proposed to find out Cluster Heads (CHs) based on residual energy, trust score and positional efficiency, and to evenly distribute the work load and stabilize the cluster formation. Each cluster is provided with a load-adaptive group of cooperative CHs, serving to distribute workloads, provide redundancy, and failover in a short time, and member drones transmit to the nearest CH to minimize range of transmission and energy loss. A Fog layer, between the cloud and the drone network carries out the in-network aggregation and early-decision-making that decreases the latency and the transmission over long distances. Dynamic trust score-based validation mechanism is used in order to identify the compromised nodes before passing on the data. The results of the simulation indicate that the proposed framework is efficient and stable. The evaluation is based on 30 independent simulation runs across 50 rounds. The average energy consumed is 0.8133 \(\pm \) 0.0263 J/round. This comprises data relay, authentication, control signalling and fog processing. The end-to-end latency is 10.50 ms. The packet reduction rate is 94.99 \(\pm \) 0.014%. The attack detection accuracy is 99.97 \(\pm \) 0.01%. This is analyzed based on five adversarial models, which are Sybil, collusion, wormhole, replay and packet-drop. The Packet Delivery Ratio (PDR) is 87.48% \(\pm \) 0.005%. The failover recovery time is 0.00009 \(\pm \) 0.000020 s. Convergence fitness of DPAFIO is 0.97533 \(\pm \) 0.00055. All results are reported as mean \(\pm \) 95% confidence interval. T-distribution is calculated on 30 trials to produce the confidence intervals.