DATAMUt: Deterministic Algorithms for Time-Delay Attack Detection in Multi-Hop UAV Networks
Keiwan Soltani, Federico Corò, Punyasha Chatterjee, Sajal K. Das · 2025
Read DATAMUt: Deterministic Algorithms for Time-Delay Attack Detection in Multi-Hop UAV Networks by Keiwan Soltani, Federico Corò, Punyasha Chatterjee, Sajal K. Das (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-07670.txt.
Introduction
Keiwan Soltani, Federico Corò, Punyasha Chatterjee, and Sajal K. Das address time-delay attacks in multi-hop drone networks. In such an attack, malicious drones deliberately delay forwarding packets, which is dangerous for time-sensitive missions like search and rescue. Deliberate delay is hard to distinguish from ordinary network delay in dynamic drone networks.
Their system, DATAMUt, represents the network's timing behavior as a weighted time-window graph. Two deterministic polynomial-time algorithms then detect attacks, one for global and one for local network knowledge. Simulations show far lower message overhead than machine learning approaches, and execution times hundreds of times faster. The authors present the method as a lightweight alternative to centralized detection.
Catalog record
| CREATOR | Keiwan Soltani, Federico Corò, Punyasha Chatterjee, Sajal K. Das |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 51 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-07670
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
DATAMUt: Deterministic Algorithms for Time-Delay Attack Detection in Multi-Hop UAV Networks Keiwan Soltani, Federico Corò, Punyasha Chatterjee, Sajal K. Das Abstract Unmanned Aerial Vehicles (UAVs), also known as drones, have gained popularity in various fields such as agriculture, emergency response, and search and rescue operations. UAV networks are susceptible to several security threats, such as wormhole, jamming, spoofing, and false data injection. Time Delay Attack (TDA) is a unique attack in which malicious UAVs intentionally delay packet forwarding, posing significant threats, especially in time-sensitive applications. It is challenging to distinguish malicious delay from benign network delay due to the dynamic nature of UAV networks, intermittent wireless connectivity, or the Store-Carry-Forward (SCF) mechanism during multi-hop communication. Some existing works propose machine learning-based centralized approaches to detect TDA, which are computationally intensive and have large message overheads. This paper proposes a novel approach DATAMUt , where the temporal dynamics of the network are represented by a weighted time-window graph ( TWiG ), and then two deterministic polynomial-time algorithms are presented to detect TDA when UAVs have global and local network knowledge. Simulation studies show that the proposed algorithms have reduced message overhead by a factor of five and twelve in global and local knowledge, respectively, compared to existing approaches. Additionally, our approaches achieve approximately 860 and 1050 times less execution time in global and local knowledge, respectively, outperforming the existing methods. 1 Introduction Unmanned Aerial Vehicles (UAVs), or drones, are remotely operated or autonomous aircraft that are used in various applications such as surveillance, agriculture, and disaster relief [ 6 ] . UAVs are cyber-physical systems (CPS) as they integrate computation (embedded software and control systems) with physical components (UAVs flying in the physical world) [ 19 ] . UAVs in particular and CPS in general require monitoring capabilities to detect and possibly mitigate erroneous and safety-critical behavior at runtime. For some specific applications like data collection, real-time monitoring, search and rescue, surveillance, and payload delivery, a collection of UAVs is used, which collaborates to gather information through multi-hop communication and relay it to a central ground station. These UAVs construct a flexible network, generally called flying ad-hoc networks or FANETs [ 18 ] , to expand network coverage. UAV networks [ 16 ] are vulnerable to various attacks such as flooding [ 1 ] , false data injection [ 21 ] , location spoofing [ 2 ] , jamming [ 14 ] , wormhole [ 3 ] , blackhole [ 9 ] , grayhole [ 5 ] , including Time Delay Attack (TDA) [ 25 ] . TDAs pose a significant threat to UAV networks, particularly in time-critical applications where real-time data is essential for effective decision making. In these attacks, malicious UAVs intentionally delay packet forwarding, disrupting the flow of information. Suppose a UAV A sends a packet to C through B . Normally, B would immediately forward the packet when encountered with C . However, malicious B introduces a delay \epsilon , which hinders timely delivery. Fig. 1 illustrates the scenario. This subtle tactic can have significant consequences. For example, in military reconnaissance, delayed intelligence could hinder strategic planning and jeopardize mission success [ 26 ] . Similarly, in search and rescue operations, delayed communication could hinder rescue efforts, as illustrated in Fig. 2 . In this scenario, a UAV relays an emergency message about an injured person on a mountain to a distant base station (BS) via multiple intermediate UAVs. The original green path represents the quickest route to the BS. However, the presence of a malicious node imposes a delay, which diverts the message along a slower red path. This can result in critical problems, including delayed emergency response, network congestion, and reduced system reliability. Such disruptions jeopardize the effectiveness of CPS, putting those in need of urgent assistance at risk. Even in civilian applications, such as package delivery [ 12 , 20 ] or traffic monitoring, these attacks can cause disruptions and economic losses. \epsilon A B B^{\prime} B^{\prime\prime} C Figure 1 : Normal and delay forwarding behaviors in the UAV network. In addition, TDAs can act as a catalyst for other types of attacks such as omission attacks [ 24 ] , false data injection attacks [ 11 ] , and data manipulation attacks [ 11 ] . TDA detection is challenging in UAV networks due to the inherent characteristics, of the network such as dynamic topology, UAV mobility, and error-prone wireless communication. These can mask short delays, making it difficult to distinguish malicious behavior from normal network latency. Furthermore, the store-carry-forward (SCF) mechanism, where UAVs store and carry data until they encounter the intended recipient, can further obscure malicious delays. An attacker can exploit this mechanism to blend their delayed transmissions with legitimate delays caused by network conditions. Figure 2 : Effect of TDA in UAV network for emergency response Existing research on TDA detection in UAV networks is limited and focuses mainly on centralized approaches [ 25 , 26 ] . These methods often require extensive computational resources and multiple iterations to learn the network’s behavior, making them unsuitable for time-sensitive scenarios. To address these gaps, we propose polynomial-time, deterministic algorithms, both centralized and distributed, under the name DATAMUt ( D eterministic A lgorithms for T ime-Delay A ttack Detection in M ulti-Hop U AV Ne t works), for detecting TDAs in UAV networks. These algorithms have lower overhead and require only a single run to identify malicious UAVs, offering a significant improvement over existing approaches. Our contributions are as follows: • We introduce an algorithm to construct a weighted graph, named T ime- Wi ndow G raph (TWiG) , which effectively captures the temporal dynamics of UAV encounters in a multi-hop UAV network. • Based on TWiG , we present an algorithm that leverages global knowledge of the network to identify malicious UAVs. This algorithm will serve as a benchmark for evaluating the performance of subsequent algorithms designed for more realistic scenarios. • We also propose a distributed algorithm for a more realistic scenario that operates with local knowledge, where each UAV only has information about its 2-hop neighborhood. This algorithm effectively detects TDAs, even with limited information. • Compare with the existing HOTD approach [ 26 ] , we demonstrate that our approaches reduce message overhead by a factor of five and twelve in global and local knowldege, respectively. Additionally, our approaches achieve approximately 860 and 1050 times less execution time in global and local knowldge, respectively. The rest of the paper is organized as follows: Section 2 reviews related works. Section 3 describes the system model. Section 4 formulates the problem. Sections 5 and 6 present our proposed algorithms to detect malicious nodes using global and local knowledge scenarios, respectively. Section 7 evaluates the performance of our proposed methods. Finally, Section 8 concludes the paper. 2 Related Works In this section, we describe the existing works on TDA in UAV networks. Although TDA is a well-known attack in wired networks and static wireless networks, such as CPS based smart power grids [ 4 , 8 ] , industrial control systems [ 23 ] , IoT networks [ 28 ] , cellular networks [ 7 ] , time-sensitive networks (TSN) [