Aerial Robots Persistent Monitoring and Target Detection: Deployment and Assessment in the Field
Manuel Boldrer, Vít Krátký, Martin Saska · 2025
Read Aerial Robots Persistent Monitoring and Target Detection: Deployment and Assessment in the Field by Manuel Boldrer, Vít Krátký, Martin Saska (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-2504-18832.txt.
Introduction
Manuel Boldrer, Vít Krátký, and Martin Saska field-test drone swarms that watch an area without pause in a 2025 paper. Persistent monitoring means robots keep sweeping a space and catch targets as they appear. Their shared algorithm blends three tools: a synchronization model run in reverse, three-dimensional Lissajous flight curves, and predictive control.
The team flew the method on real aerial robots, up to eleven at once. They define two failure grades. Short glitches, like tracking slips and message delays, form the first. Long outages, like attacks, dead links, or flat batteries, form the second. The system kept monitoring and detecting through both. The authors say the field trials prove the method's strength, resilience, and scale. They present it as ready for real deployments where coverage cannot lapse.
Catalog record
| CREATOR | Manuel Boldrer, Vít Krátký, Martin Saska |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 48 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-18832
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Aerial Robots Persistent Monitoring and Target Detection: Deployment and Assessment in the Field Manuel Boldrer, Vít Krátký, Martin Saska Abstract In this article, we present a distributed algorithm for multi-robot persistent monitoring and target detection. In particular, we propose a novel solution that effectively integrates the Time-inverted Kuramoto model, three-dimensional Lissajous curves, and Model Predictive Control. We focus on the implementation of this algorithm on aerial robots, addressing the practical challenges involved in deploying our approach under real-world conditions. Our method ensures an effective and robust solution that maintains operational efficiency even in the presence of what we define as type I and type II failures. Type I failures refer to short-time disruptions, such as tracking errors and communication delays, while type II failures account for long-time disruptions, including malicious attacks, severe communication failures, and battery depletion. Our approach guarantees persistent monitoring and target detection despite these challenges. Furthermore, we validate our method with extensive field experiments involving up to eleven aerial robots, demonstrating the effectiveness, resilience, and scalability of our solution. Video —https://mrs.fel.cvut.cz/persistent-monitoring-auro2025 Code —https://github.com/ctu-mrs/distributed-area-monitoring keywords Multi-Robot Systems, Distributed Control, Kuramoto model, Persistent Monitoring, Target Detection. 1 introduction Swarm of aerial robots can simultaneously gather data from different locations, providing comprehensive and real-time insights into air quality Bolla and others (2018) , wildlife movements Shah et al. (2020) , natural disasters Bailon-Ruiz et al. (2022) , and crop health Carbone and others (2022) . This coordinated effort can significantly enhance the speed and accuracy of data collection, leading to more informed decision-making and timely responses. In this article, we present a novel algorithm for aerial multi-robot persistent monitoring and detection of targets or events inside a given rectangular mission space. This work is unique in focusing on properties required for deployment in real-world scenarios. In particular, we deployed up to 11 Unmanned Aerial Vehicles (UAVs) in the field to assess the ideas of persistent monitoring and target detection algorithm that we presented in Boldrer et al. (2022a) . 1.1 Related works and our contribution In the realm of distributed multi-robot coverage control, we can identify three distinct families. 1. static coverage : the robots converge to a static equilibrium state that maximizes the covered area Cortes et al. (2004) ; Boldrer et al. (2019) . 2. dynamic coverage/persistent monitoring : the robots coordinate their motion to eventually span all the mission space and, in the case of persistent monitoring, they do it periodically Smith et al. (2011) ; Boldrer et al. (2021) ; Pasqualetti et al. (2012) ; Pinto et al. (2020) ; Franco and others (2015) . 3. Target detection/pursuit-evasion : the goal is to maximize the probability of detecting targets or events in the mission space Mavrommati et al. (2017) ; Robin and Lacroix (2016) ; Durham et al. (2012) . Figure 1: Experimental verification of the proposed approach in the field with 11 UAVs. In this article, we propose a unified solution both for persistent monitoring (for all the robots) and target detection. In particular, we want to synthesize an algorithm that steers each robot to continuously span all the mission space, providing guarantees for target detection in finite time. In the literature, there are a few works that address this problem. In Borkar et al. (2016) ; Borkar et al. (2020) solutions based on Lissajous curves Richards (1902) are provided to solve the problem. In particular, the authors provide to the robots precomputed trajectories to follow through a tracking algorithm, however no information about the neighbors state is considered during the mission, making the solution not robust to synchronization errors. Recently, Nath and Ghose (2024) exploited the properties of Lissajous knots to solve the problem of dynamic coverage of 3D structures. In Boldrer et al. (2022a) , we combined the Time-inverted Kuramoto dynamics Boldrer et al. (2021) with Lissajous curves properties to obtain a distributed algorithm, which promotes robustness and resiliency during the mission. A coordination strategy based on guiding vector field and consensus is proposed in Yao et al. (2021) . Nevertheless, it requires the desired distance between robots to be specified. Moreover, the time-inverted Kuramoto model offers the possibility to reach also alternative equilibrium configurations autonomously, i.e., \kappa -cluster equilibrium points (see Section 2 ). Another crucial lack in the literature is the absence of deployment and assessment of generic coverage algorithm on aerial robotic platforms in the field. In fact, the mentioned works implement their algorithms in simulated or controlled environments with limited areas of interest. The presence of an ideal localization system is often assumed, and the implementation is often centralized. These environments lack presence of realistic disturbances and communication delays due to their relative simplicity. A limited number of multi-robot coverage algorithms have been validated in field experiments, Apostolidis and others (2022) ; Barrientos and others (2011) ; Shah et al. (2020) ; Datsko et al. (2024) . The mentioned implementations rely on precomputed trajectories, which simplifies coordination, but often reduce adaptability during execution. As a result, robustness to uncertainties, failures, or tracking errors is limited in such approaches. Finally, we cover the case where the robots experience unexpected events, e.g., malicious attacks, communication delays, tracking errors and other failures such as battery depletion or communication loss, which often occur during real-world deployment. Also this research line is quite unexplored Liu et al. (2021) ; Zhou and Kumar (2023) . In Boldrer et al. (2022b) , we provided sufficient conditions for stability and showed the degree of resiliency of the system to attacks and failures, without providing an effective solution to mitigate their effects. In Borkar et al. (2020) , the authors proposed a method to reconfigure the formation that allows to add, remove and replace robots from the swarm. However, it needs all the robots in the network to be aware of the new parameter changes, which may introduce scalability issues. Our contribution is threefold. Firstly, we designed a novel approach for aerial robots persistent monitoring and target detection in the three dimensional case, relying on the theoretical study we presented in Boldrer et al. (2022a) for ground robots. Notice that, with respect to existing methods for persistent monitoring and target detection Borkar et al. (2016) ; Borkar et al. (2020) , our solution does not rely on precomputed trajectories, but it implements a constant feedback action on the basis of the neighbors state. Secondly, we show how the proposed approach can be adopted in case of short-time failures (type I), such as tracking errors, and long-time failures (type II), such as malicious attacks or battery depletion, providing a detailed analysis on the conditions to obtain persistent monitoring and target detection guarantees. Even by considering these additional challenges, we preserve the distributed nature of the approach, and its scalability. Finally, we bridge the gap between theory and practice by implementing the proposed algorithm in the field. In particular, we implement the approach on up to 11 UAVs and analyzed the collected data, thereby validating the proposed method. 2 Problem description and proposed solution In this section, we provide the problem description and the proposed solution that rely on our previous work Boldrer et al. (2022a) . Notice that with respect to Boldrer et al. (2022a) , in this section,