Democritus University of Thrace (DUTH)
Multiple unmanned aerial vehicles motion planning and routing for optimized remote sensing operations
Abstract
dc:descriptionIn the context of this dissertation we deal with the problem of motion planning and routing for multiple unmanned aerial vehicles (UAVs) to perform effective and efficient remote sensing operations (RS). The largest part of the presented work focuses on Coverage Path Planning (CPP) and multi-agent CPP (mCPP), that is the problem of having to completely cover a designated region of interest (ROI) with one or more unmanned vehicles, while simultaneously respecting any restrictions enforced by the ROI, vehicle(s), user(s), and operational or legal frameworks. Additionally, a gap in literature for a specific type of UAV RS operations is identified, the problem of Fast Inspection of Scattered Regions (FISR) is formulated and described in detail, and a novel method to deal with it is introduced. As part of this work, a novel mCPP method tailored to real-world UAV RS operations has been introduced, that supports very complex-shaped, non-convex ROIs with additional no-fly zones (NFZs) inside them, manages operational resources like time and energy wisely, and manages to achieve almost complete coverage in all cases. To realize that, an optimization scheme based on simulated annealing has been introduced to efficiently apply grid-based path planning methodologies in real-world ROIs. Moreover, a heuristic is applied for the Spanning Tree Coverage (STC) path planning method to reduce the number of turns in the generated trajectories, and a modification to a state-of-the-art (SoA) method for area allocation is performed to support proportional area division and enable the simultaneous utilization of heterogeneous vehicles. The introduced method is extensively evaluated through a series of simulated runs and a comparison with an alternative SoA method. A multi-agent marginal utility study is performed to provide information on how and when the deployment of multiple UAVs can benefit coverage missions. Two indicative real-world scenarios regarding precision agriculture and search and rescue (SAR) operations are deployed, collecting and processing data to provide relative results. As a next step, building upon the aforementioned work, a CPP method with three ad-hoc coverage modes that intend to fulfill different objectives—each of them more appropriate for different applications, operational demands, and vehicles—is introduced. This work focuses on resolving any remaining issues when applying grid-based path planning missions to real-world operations. While the introduced heuristics are tailored to the STC method, this work can serve as a practical guide for most grid-based methods. The presented CPP solution is evaluated using the same simulation setup and benchmark ROIs as in the previous work, providing a direct comparison with it and the alternative SoA method. The evaluations indicate that all claimed benefits are achieved, while simultaneously the method manages to overcome the issues inherited from its grid-based nature. Finally, a new problem of UAV path planning for RS operations has been identified, described in detail, and resolved by introducing a novel method tailored to this purpose. The FISR problem, as stated above, is about the inspection of multiple disjoint ROIs with one or more UAVs in the shortest amount of time possible. The methodology introduced to resolve it is the multi-UAV Disjoint Areas' Inspection (mUDAI). mUDAI treats the FISR problem as two discrete optimization sub-problems: first, the optimal viewpoints for capturing each defined ROI are calculated, and second, a set of trajectories that minimize time and respect the specifications and limitations of the vehicles used is generated. The introduced method is validated via two sets of real-world UAV deployments for FISR missions. To enable easy and user-friendly deployment of UAVs for testing and evaluating the presented methodologies in real-world conditions, an end-to-end mission planning platform for RS operations has been developed. This platform, introduced in the first part of this work and updated continuously since then, includes everything needed to design, store, manage, and deploy multi-UAV missions for RS operations with commercial devices. The combination of state-of-the-art performance with increased levels of readiness and ease of use takes steps toward closing the gap between commercial platforms that use simplistic approaches but provide robust and trusted solutions for real-world operations, and research methods that introduce novel features and benefits but are usually not ready or easy to deploy in the field. Additionally, the development of solutions that allow the efficient and safe deployment of multiple UAVs working cooperatively to perform RS operations contributes to overcoming one of the major current limitations of multi-copter UAVs: their limited operational duration. In this way, the inspection of larger areas during the nominal operational duration of the vehicles is achieved, and time-critical operations and applications involving the detection of moving objects of interest are handled much more efficiently.
Degree
thesis:*- Grantor dc:publisher
- Democritus University of Thrace (DUTH)
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Apostolidis, Savvas
- Αποστολίδης, Σάββας
Subjects
dc:subject × 24- Συστήματα Μη Επανδρωμένων Αεροσκαφών
- Τηλεπισκόπηση
- Συστήματα πολλαπλών πρακτόρων
- Σχεδιασμός κίνησης
- Δρομολόγηση
- Σχεδιασμός Τροχιάς Κάλυψης
- Γρήγορη Επίβλεψη Διασκορπισμένων Περιοχών
- Επίβλεψη μη Συνεκτικών Περιοχών με πολλαπλά ΣμηΕΑ
- Unmanned aerial vehicles
- Remote sensing
- Multi-agent systems
- Motion planning
- Routing
- Coverage path planning
- Fast Inspection of Scattered Regions
- multi-UAV Disjoint Areas Inspection
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Επιστήμες Μηχανικού και Τεχνολογία
- Αυτοματισμοί και Συστήματα ελέγχου
- Ρομποτική
- Electrical Engineering, Electronic Engineering, Information Engineering
- Engineering and Technology
- Automation and control systems
- Robotics
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/59050