In complex missions such as disaster rescue and field inspection, UAVs often need to perform collaborative operations across different regions. However, when multiple UAV groups enter communication-restricted areas and are unable to connect with trusted ground authorities, how to complete identity authentication, group fusion, and subsequent security management has become a key challenge in UAV swarm applications.

This issue of the AMOVLAB 10,000 RMB Scholarship Paper Sharing focuses on the research work participated in by Manting Gan from Anhui University, China. The paper has been accepted by IEEE Transactions on Mobile Computing, and has received the First Prize of the AMOVLAB Campus Scholarship for its outstanding research achievements, with a prize of 10,000 RMB.

The paper proposes a distributed management scheme for UAV groups in offline environments, enabling UAV groups to achieve autonomous fusion and certificate updates even without online support from centralized authorities.

The paper link is provided at the end of the article for direct access. If this article is helpful to you, please feel free to like, share, and recommend it.

01

Research Background

UAVs have been widely applied in tasks such as remote sensing, surveying and mapping, road inspection, and emergency rescue. However, limited by endurance, operational range, and onboard computing capability, a single UAV is often unable to independently cover complex and large-scale missions.

After multiple UAVs form a group, they can improve system efficiency, robustness, and reliability through distributed collaboration, task sharing, and shared sensing and computing resources.

As the mission area continues to expand, UAV groups managed by different mobile stations also need to dynamically merge. After completing operations in one area, one group can share resources with another UAV group, combine operational coverage, and continue advancing toward farther regions.

The challenge is that such missions often take place in communication-restricted environments such as mountainous areas and disaster zones, where UAV groups may already be beyond the coverage of trusted ground authorities.

Figure 1. UAV Group Fusion Scenario

Existing Technical Challenges

Existing UAV group management schemes can generally be divided into three categories: centralized approaches, blockchain-based approaches, and multi-center approaches.

Although each approach has its own advantages, they usually encounter four challenges in scenarios involving offline trusted authorities and continuously changing UAV groups:

Strong dependence on online services:
Group fusion, identity authentication, or certificate updates require continuous participation from trusted authorities. Once the UAV group moves beyond communication coverage, the process may be interrupted.

Insufficient flexibility for fusion:
Some solutions require new members to reapply for certificates from the central authority, while others rely on blockchain-based identity records or cluster-head nodes, making it difficult to support autonomous and continuous fusion among multiple groups.

Difficulty in continuous certificate updates:
When UAV group members change, existing certificates may no longer fully match the identity relationships of the new group. If updates still rely on centralized computation and distribution, they cannot be completed in offline environments.

Unfriendly resource overhead:
Blockchain maintenance, complex cryptographic operations, or dedicated tamper-resistant hardware increase onboard computing, storage, and deployment costs.

This research focuses on a clear objective: when the Trusted Authority (TA) is unavailable, multiple UAV groups can still complete secure fusion, and the merged members can autonomously update their certificates.

To achieve this goal, the authors use Mobile Proactive Secret Sharing (MPSS) as the core technology and construct a distributed, lightweight UAV group management scheme.

02

Research Highlights

This research applies Mobile Proactive Secret Sharing (MPSS) technology for the first time to UAV group security management in scenarios where the Trusted Authority (TA) is offline.

By combining threshold signature technology, the research constructs a distributed autonomous management architecture, reducing real-time dependence on trusted authorities and ground stations.

Decentralized Autonomous Group Fusion Technology

During the initialization and registration stage, the Trusted Authority (TA) and Mobile Station (MS) still participate.

After entering the mission area, group authentication, key updates, and group fusion are completed collaboratively by UAV nodes.

MPSS can be understood as follows:

The system divides critical secrets into multiple shares, with different UAVs holding different secret shares. During group fusion, UAV nodes exchange and update their respective “secret shares” to generate the required key relationships for the new group without reconstructing the complete secret.

Before fusion, threshold signatures are used to verify member identities in batches. Authentication is established only when a predefined number of legitimate members participate together, thereby reducing the risk of unauthorized nodes joining the network.

The entire architecture does not use a fixed cluster head, avoiding reliance on a single node for system reliability.

Figure 2. UAV Group Fusion System Model

Lightweight Certificate Self-Update Mechanism

After group fusion, UAVs use newly generated sub-keys to update their identity certificates without waiting for the central authority to recalculate or distribute certificates.

The update process mainly consists of lightweight operations such as modular multiplication and hash computation, making it suitable for devices with limited onboard computing capability.

Meanwhile, the proposed scheme utilizes the trust transfer relationship of the certificate chain, extending the existing trust relationship:

“TA authenticates the Mobile Station — Mobile Station authenticates UAVs”

to the newly formed UAV group.

This ensures that the certificate source remains verifiable and the identity relationship remains traceable.

High-Fault-Tolerance and High-Security Cryptographic Architecture

The proposed scheme adopts a (t, n) threshold signature mechanism.

Among n members, a collective signature can be completed as long as at least t legitimate members participate.

Therefore, up to n-t nodes can fail or become disconnected without affecting authentication, group fusion, or certificate updates.

The paper also provides a formal security analysis based on the Random Oracle Model and discusses common attacks such as impersonation attacks, tampering attacks, and man-in-the-middle attacks.

The design goal is not only to achieve “successful fusion”, but also to achieve verifiably secure fusion.

The dynamic update of key shares also means that the leakage of historical shares does not directly result in the compromise of subsequent session keys.

Continuous Multi-Group Fusion Adaptability

After fusion, the newly formed group generates a unified group public key.

When connecting with a third or fourth UAV group in subsequent operations, the same authentication and share-update procedures can continue to be used without returning to the Trusted Authority for re-registration.

For large-scale missions executed in multiple stages, this capability of “continuous operation while performing group fusion” is closer to real-world operational requirements than one-time network formation.

03

Experimental Testing

The proposed scheme was implemented and tested on both a real UAV platform and a PC simulation environment.

Its feasibility was evaluated from three perspectives:

• Computational overhead
• Communication overhead
• Security properties

Experimental Environment and Parameter Configuration

Hardware Environment

The UAV platform uses the AMOVLAB P600 research UAV, equipped with an NVIDIA Carmel 1.4 GHz ARM processor and running Ubuntu 18.04. The simulated Trusted Authority platform uses an Intel Core i7-11700 processor, 16 GB memory, and Windows 10. The introduction of a real onboard UAV platform makes the evaluation closer to practical engineering environments with limited resources.

Figure 3. AMOVLAB Prometheus 600 UAV

Cryptographic Parameters

The experiment uses the BLS12381 elliptic curve, SHA-384 hash function, and AES symmetric encryption. The corresponding cryptographic operations are implemented based on the Miracl Core cryptographic library.

Computational Overhead Experimental Results

As the size of the UAV group increases, the computational overhead of each stage shows different growth trends:

• The computational overhead of intra-group key negotiation, cross-group key negotiation, and certificate updates increases approximately linearly, remaining relatively stable overall.
• The group fusion stage requires multiple nodes to exchange and reorganize key shares. Therefore, its computational overhead increases quadratically with the number of members and becomes the main source of overall computational cost.
• The proposed scheme does not require the Trusted Authority or ground control units to remain online to perform fusion computation. Instead, the computation is completed in a distributed manner by UAV nodes, achieving offline autonomy and reducing dependence on centralized infrastructure. Compared with the related schemes selected in the paper, this scheme demonstrates advantages in computational efficiency and security feature coverage.

Figure 4. Computational Overhead of Each Stage in the Proposed Scheme

Communication Overhead Experimental Results

The communication overhead during intra-group and cross-group key negotiation remains relatively stable. During the certificate update stage, no additional data exchange is required, resulting in zero communication overhead. Group fusion requires encrypted key shares to be transmitted between nodes, causing communication overhead to increase quadratically with the group size.

This is also a cost that distributed solutions need to consider: Removing centralized nodes does not mean eliminating costs; instead, the cost is transferred to internal interactions among group members.  The paper points out that high-speed communication technologies such as 5G/B5G can reduce transmission latency. However, for larger-scale UAV groups, how to further reduce share interaction remains an important direction for future optimization.

Figure 4. Communication Overhead of Each Stage in the Proposed Scheme

Security Performance Validation

• Formal Security Proof：A rigorous security proof is completed based on the Random Oracle Model and the Elliptic Curve Discrete Logarithm Problem (ECDLP). The proposed scheme can effectively resist signature forgery attacks. Within polynomial time, no attacker can break the keys or forge legitimate identities.
• Comparison of Security Properties：Compared with mainstream solutions, the proposed scheme simultaneously satisfies the following security properties: Mutual authentication, No dependence on online centralized authorities, Fairness, High fault tolerance, Multi-group adaptability, Autonomous fusion, Certificate self-update. Furthermore, the scheme does not require TPD hardware support, providing the most comprehensive security functions and the strongest adaptability to different application scenarios.
• Attack Resistance Capability: The proposed scheme can effectively defend against: UAV impersonation attacks, Message tampering attacks, Man-in-the-middle attacks

The hash mechanism ensures message integrity, while the threshold signature mechanism guarantees identity legitimacy, providing comprehensive protection against UAV group communication security risks.  The core significance of this research is that, under the constraint of a Trusted Authority being offline, it still establishes a complete operational loop integrating identity authentication, group fusion, and certificate updates. Experimental results demonstrate competitive computational efficiency. Meanwhile, the results also clearly show that as the UAV group size increases, the computational and communication costs during the fusion stage require further optimization.

04

Application Scenarios

The proposed scheme focuses on decentralized, offline, and dynamic scenarios. It is suitable for long-distance, large-scale UAV group collaborative operations without ground communication support. The following three scenarios are the most representative:

Disaster Emergency Rescue

Disasters such as earthquakes, floods, and landslides may damage ground communication stations and control centers.

Rescue UAVs from different regions and different deployment batches can perform autonomous authentication and fusion at the operation site, continuously carrying out: Search and rescue, Supply delivery, Environmental monitoring missions.

Long-Distance Field Inspection

In beyond-visual-line-of-sight missions such as: Power transmission line inspection, Oil and gas pipeline inspection, Border security monitoring, Forest fire prevention. UAV groups frequently move in and out of ground communication coverage. The proposed scheme supports multiple inspection groups merging during operations and autonomously completing identity and certificate updates, reducing mission interruptions caused by reconnecting to centralized systems.

Collaborative Operations in Complex Airspace

In dynamic missions such as urban mapping and large-scale meteorological monitoring, different batches of UAVs need to reorganize according to mission progress. The continuous fusion capability enables the system to adapt to topology changes while maintaining trusted identities and secure communications as the collaboration scope expands.

05

Resource Information

Paper Title:
Distributed and Autonomous Group Management Supporting Group Fusion for UAVs

Paper Link:
https://ieeexplore.ieee.org/document/11397563

Published Journal:
IEEE Transactions on Mobile Computing

DOI:
10.1109/TMC.2026.3665690