The convergence of artificial intelligence (AI), edge computing, and next-generation wireless networks is revolutionizing the way autonomous systems operate in mission-critical environments. Drone-based delivery systems have emerged as a vital component of smart healthcare logistics, enabling rapid transport of medical supplies, especially in geographically dispersed or inaccessible regions. However, the increasing reliance on autonomous drones introduces new security challenges, particularly in safeguarding data transmission.

The integration of deep learning techniques, particularly Convolutional Neural Networks (CNNs), within 6G-powered edge cloud infrastructures offers a promising solution to address security vulnerabilities while maintaining real-time operational efficiency. The 6 G network, with its ultra-high-speed and low-latency capabilities, is crucial for enabling real-time communication between drones and the edge cloud infrastructure, thereby enhancing the security and efficiency of drone operations.

In a paper presented at the 2025 IEEE Conference on Communications and Network Security (CNS), researchers introduce the next-generation deep learning-integrated security architecture for drone applications based on 6G-enabled edge cloud networks. The proposed system leverages CNNs for real-time network traffic analysis and anomaly detection, enabling the identification of cyber threats such as spoofing, jamming, and unauthorized access. After covering the related work on this subject, the researchers present the proposed architecture, methodology, and evaluation. 

Next-Generation Deep Learning Framework

The framework comprises three core components: 

  • Drone Security Architecture
    Unmanned aerial vehicles (UAVs) or drones equipped with onboard sensing and communication units operate in an urban environment and communicate through 6G wireless networks. These high-speed and low-latency wireless networks ensure real-time communication between drones and edge cloud infrastructure.
  • Methodology—
    The CNN-enabled system leverages learning models to intelligently classify, prioritize, and allocate drone tasks to the most optimal edge cloud node based on real-time constraints such as wireless bandwidth, security demands, task urgency, and node availability.
  • Security Method
    Integrates all essential components—Edge Cloud Controller, Drone Tasks, and Wireless Tasks—into a cohesive policy-driven security enforcement model. This end-to-end integration between CNN task scheduling, wireless communication, and edge-based security governance ensures a robust and resilient environment for drone operations in urban areas.

Proposed framework

 

The entire architecture supports high mobility and dynamic scalability. The drones are capable of seamless handovers between 6G networks and cloud nodes without disruption. By integrating 6G wireless technology, AI-based scheduling, and layered security protocols, this framework addresses the critical challenges in drone traffic management, cybersecurity, and real-time decision-making in innovative city environments.

Evaluation & Results

To evaluate the proposed CNN-integrated Scheduling Security (CNN-SS) framework, the researchers deployed a real-world experimental testbed using both Python and Android technologies, emulating a 6G-enabled distributed edge cloud infrastructure for drone applications. The testbed comprises multiple edge cloud nodes, a fleet of autonomous drones equipped with Android-based control modules, and simulated 6G wireless connectivity. 

The evaluation metrics include end-to-end latency, task success rate, and security validation ratio. Results are collected and visualized to show the improved security detection rate and reduced latency, confirming the efficiency and robustness of the CNN-SS algorithm in 6G edge cloud drone networks.

In the result analysis, we determined the performance of different methods. The results illustrate the superior scalability and reduced delay of the proposed CNN-SS scheduling algorithm in high-load conditions, making it more suitable for real-time and large-scale edge computing scenarios such as drone package delivery or smart logistics.

End-to-end latency for 1000 drone tasks from source to destination

 

According to the researchers, the results also indicated that CNN-SS is highly effective for drone-based security tasks, underscoring CNN-SS's superiority in 6G environments, where high-speed data flows demand resilient detection. It proves adaptable to resource-constrained edge devices, whereas DSPDM and DCNNS fail to handle attack surges. 

Conclusion and Future Work

The study presents a novel deep learning-based security framework, CNN-SS, for efficient and secure task scheduling in drone applications operating over 6G-enabled edge cloud networks. The proposed CNN-SS architecture is not only scalable for large drone fleets but also adaptable to various mission-critical applications, including medical supply delivery, surveillance, and emergency response. 

Despite these promising outcomes, several aspects remain open for further exploration, and the researchers have plans to:

  • Extend the CNN-SS framework by incorporating federated learning to ensure privacy-preserving model training across distributed drones. 
  • Consider adaptive scheduling mechanisms that utilize real-time threat intelligence and energy consumption models to optimize resource allocation.
  • Evaluate the robustness of the CNN-SS framework under realistic environmental and adversarial conditions. 

These enhancements aim to evolve CNN-SS into a comprehensive and scalable security solution for the rapidly advancing 6G-based drone ecosystem.

Interested in learning more about Cybersecurity and Drone Security? The IEEE Xplore Digital Library offers over 48,000 publications on Cybersecurity and over 2,000 publications on Drone Security.

Interested in acquiring full-text access to this collection for your entire organization? Request a free demo and trial subscription for your organization.