Radius of Robust Feasibility for Ground Coverage in Aerial Sensor Networks

Vanshika Datta, C. Nahak · 2025

Read Radius of Robust Feasibility for Ground Coverage in Aerial Sensor Networks by Vanshika Datta, C. Nahak (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-2510-20213.txt.

Introduction

Vanshika Datta, C. Nahak, and J. C. Yao build a robust plan for ground cover by drone-based sensors whose spots may shift. Each drone sensor casts a sector shape on the ground. Its size is set by height, view field, and aim. Cover drops fast when a sensor drifts from its planned spot. Their plan bakes drift room right into the task.

A score called the radius of robust feasibility gauges how much drift a plan can take. The hard worst-case task has no smooth form and is NP-hard. So a shared greedy step turns each sensor by local facts and aims at high-value zones. In sim tests, the plan keeps strong cover on rough land and in mixed states.

Catalog record

CREATORVanshika Datta, C. Nahak
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE61 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2510-20213

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Radius of Robust Feasibility for Ground Coverage in Aerial Sensor Networks Vanshika Datta, C. Nahak Abstract Sensors are vital for environmental monitoring, yet their effectiveness diminishes under spatial uncertainty. We propose a robust optimization framework for maximizing the coverage of aerial directional sensors under spatial uncertainty. Each sensor projects a truncated sector on the ground, parameterized by its altitude, field of view, and orientation. To address sensor displacement uncertainty, we introduce the radius of robust feasibility (RRF) as a measure of tolerance against positional perturbations. We formulate an exact expression for the RRF of aerial sensor networks and embed it into the coverage maximization model as a robustness constraint. Our approach guarantees that the optimized configuration remains feasible under bounded uncertainty. A distributed greedy algorithm based on Voronoi partitioning is used for orientation adjustment, ensuring scalable and adaptive deployment toward high-impact regions. Experimental results validate the effectiveness of our model in preserving robust coverage across complex terrain and varying uncertainty conditions. MSC 49K99; 65K10; 90C29; 90C46; 90C34 1 Introduction Sensors possess the inherent ability to perceive, record, and react to their surrounding environment. This fundamental capability underpins their widespread application in diverse domains such as airport surface surveillance, border monitoring, disaster response, smart agriculture, infrastructure inspection, and urban traffic management [ 1 , 2 ] . As sensors evolved from passive data collectors to active decision-making agents, wireless sensor networks (WSNs) emerged as a critical infrastructure for real-time monitoring in resource-constrained environments [ 3 , 4 ] . Within the broader class of WSNs, directional sensor networks (DSNs) have attracted significant attention due to their use of limited-angle sensors, such as cameras, infrared, ultrasonic, and LiDAR, offering focused coverage with higher precision [ 5 , 6 ] . These directional sensors necessitate not only optimal placement but also strategic orientation to ensure maximum effective coverage, making the deployment problem inherently more complex than in traditional omnidirectional WSNs [ 7 ] . This complexity is further amplified in aerial sensor networks (ASNs), wherein sensors are mounted on unmanned aerial vehicles (UAVs) or drones and deployed in three-dimensional space. Their sensing footprints, projected onto the ground, are typically fan-shaped and depend heavily on altitude and yaw angle [ 2 , 3 ] . However, a major challenge across these settings is uncertainty, stemming from sensor placement errors, localization drift, environmental disturbances, or communication inaccuracies. In cluttered or GPS-denied environments such as dense forests or disaster zones, such uncertainties can significantly degrade sensing performance, resulting in coverage gaps, misaligned orientation, energy inefficiency, and compromised communication [ 8 , 9 ] . Although grid-based and geometric deployment methods have attempted to optimize coverage [ 6 , 10 , 11 ] , they often assume precise sensor positioning, rendering them fragile under even small perturbations. While many efforts have focused on energy-efficient design and network longevity [ 12 , 13 ] , the deployment strategies themselves must also be resilient to uncertainty. Traditional probabilistic methods such as stochastic programming require accurate distributional assumptions, which are often unavailable in field deployments. This motivates the need for robust optimization (RO), a paradigm that accounts for uncertainty in a deterministic yet conservative manner. [ 14 , 15 ] Robust optimization explicitly incorporates uncertainty sets within optimization models, producing solutions that remain feasible across all realizations of uncertainty within those sets. Originating from Soyster’s foundational work [ 16 ] and advanced through tractable frameworks by Ben-Tal, Nemirovski, and El Ghaoui [ 14 , 17 , 18 ] , RO has matured into a powerful methodology applicable across continuous, discrete, and mixed-integer domains [ 19 , 20 , 21 , 22 ] . Its value is particularly evident in high-stakes or data-sparse environments, where conservative feasibility is prioritized over nominal optimality [ 23 ] . RO enables declarative statements like, “I do not know exactly where the sensor is, but I require the coverage constraint to be satisfied regardless of its exact position within an uncertainty region.” Yet a persistent challenge within RO is determining the threshold of uncertainty that a given solution can tolerate. Here, the radius of robust feasibility (RRF) plays a crucial role. RRF encapsulates the greatest size of an uncertainty set such that a desirable property, such as constraint feasibility, is preserved regardless of the parameter realization. Originally proposed by Goberna et al. [ 24 ] for linear semi-infinite programs (LSIPs), this concept has significantly evolved to encompass convex, nonconvex, and mixed-integer frameworks, see [ 23 ] and the references therein. The RRF offers not just a theoretical foundation but also a practical tool for designing systems that retain functionality under worst-case scenarios, which is often more critical than optimizing for ideal conditions. The radius essentially delineates the largest possible uncertainty set, typically modeled as a norm ball or more general convex set, within which the robust counterpart remains feasible, thereby ensuring system reliability, see [ 25 ] . Furthermore, the distinction between different types of radii, such as the radius of highly robust weakly efficient solutions for multi-objective convex programs [ 24 ] , and the radius of robust feasibility for constraint-focused optimization [ 26 ] , underscores the versatility of this concept. The latter, which forms the basis of our work, serves as a computational threshold ensuring that robust feasible solutions exist across all admissible perturbations. Applications of RRF now span from gas network control and facility location design to wireless and distributed sensor networks, proving its value in domains where preserving system operability under uncertainty is paramount [ 27 ] . In the context of distributed sensor networks, where sensor deployment often faces imprecision due to terrain irregularities, environmental noise, or mobility, the integration of RRF provides a rigorous mechanism to tolerate spatial uncertainty. Instead of assuming perfect placement, RRF enables one to model a tolerance region, allowing small displacements without violating feasibility [ 3 ] . This leads to robust strategies that enhance system resilience while preserving coverage quality. As Ridolfi et al. [ 28 ] have recently demonstrated, transitioning from uncertain coefficients to uncertain locations marks a significant breakthrough, enabling new insights in location-based optimization. The RRF offers a powerful geometric and optimization-based lens through which to examine robustness. Given the energy constraints of sensor networks, an uncertainty-aware approach is essential for optimizing data collection while ensuring long-term network sustainability. A robust design, rather than one solely focused on optimality under fixed conditions, enhances the adaptability of DSNs across varying uncertainty scenarios, reinforcing the importance of incorporating RRF into DSN modeling. In this paper, we extend the RRF paradigm to aerial directional sensor networks, a class of DSNs where each sensor is mounted at a fixed height and projects a directional sensing truncated cone onto the ground. This formulation is particularly applicable to UAV-based surveillance, environmental monitoring, and disaster response. Building upon the Voronoi-based greedy orientation framework introduced in prior work [ 7 ] , we propose an uncertainty-aware optimization model that strategically orients sensors to maximize coverage, while ensuring robust feasibility under bounded sensor displacement. Our contributions can be summarized as follows: • We formulate a nominal coverage maximization model for aerial DSNs and extend it into a robust optimization framework that explicitly accounts for location uncertainty using RRF. •

Read the complete text