{"id":{"repo_id":"york","oai_identifier":"oai:yorkspace.library.yorku.ca:10315/43306"},"canonical_url":"https://search.dev.ndltd.org/etd/york/oai:yorkspace.library.yorku.ca:10315/43306","repository":{"repo_id":"york","name":"York University","base_url":"https://yorkspace.library.yorku.ca/oai/request"},"display":{"title":"Digital Twin Platform for Drone Applications","abstract":"Simulation tools and digital twins are vital to UAV research, enabling efficient prototyping, testing, and deployment of algorithms in dynamics, control, computer vision, and deep learning. Despite their importance, existing simulators face limitations in modularity, scalability, and multidisciplinary integration, while the persistent sim-to-real gap remains a major challenge. Nevertheless, simulators are indispensable, particularly for deep reinforcement learning (DRL), where extensive data collection in real-world settings is infeasible due to hardware and safety constraints. This thesis presents a modular, scalable, and transferable simulation platform for quadrotor UAVs, designed to support research in control systems, DRL, and embodied AI. The platform incorporates a high-fidelity digital twin, modeled using Blade Element Theory, with parameterized aerodynamic forces tailored to the Quanser QDrone2 and adaptable to other quadrotor architectures. A dedicated Control library, built within the Gazebo ecosystem, facilitates rapid prototyping and benchmarking of UAV controllers under a unified framework. To extend applicability to autonomous navigation, computer vision, and state estimation, Gazebo plugins are developed to enable efficient same-process interfacing and integration with both ROS1 and ROS2, thereby ensuring seamless transfer of algorithms from simulation to hardware. To advance DRL research, a suite of deterministic and stochastic environments is introduced. These environments balance reproducibility and realism by modeling communication delays, timing variations, and other uncertainties, while remaining compatible with widely used DRL libraries. Their parallelizable design enables large-scale data collection within hours, significantly accelerating training. Experimental validation confirms the platform’s fidelity and effectiveness in bridging the gap between simulation and real-world UAV deployment","abstract_html":"Simulation tools and digital twins are vital to UAV research, enabling efficient prototyping, testing, and deployment of algorithms in dynamics, control, computer vision, and deep learning. Despite their importance, existing simulators face limitations in modularity, scalability, and multidisciplinary integration, while the persistent sim-to-real gap remains a major challenge. Nevertheless, simulators are indispensable, particularly for deep reinforcement learning (DRL), where extensive data collection in real-world settings is infeasible due to hardware and safety constraints. This thesis presents a modular, scalable, and transferable simulation platform for quadrotor UAVs, designed to support research in control systems, DRL, and embodied AI. The platform incorporates a high-fidelity digital twin, modeled using Blade Element Theory, with parameterized aerodynamic forces tailored to the Quanser QDrone2 and adaptable to other quadrotor architectures. A dedicated Control library, built within the Gazebo ecosystem, facilitates rapid prototyping and benchmarking of UAV controllers under a unified framework. To extend applicability to autonomous navigation, computer vision, and state estimation, Gazebo plugins are developed to enable efficient same-process interfacing and integration with both ROS1 and ROS2, thereby ensuring seamless transfer of algorithms from simulation to hardware. To advance DRL research, a suite of deterministic and stochastic environments is introduced. These environments balance reproducibility and realism by modeling communication delays, timing variations, and other uncertainties, while remaining compatible with widely used DRL libraries. Their parallelizable design enables large-scale data collection within hours, significantly accelerating training. Experimental validation confirms the platform’s fidelity and effectiveness in bridging the gap between simulation and real-world UAV deployment","abstract_has_math":false,"creators":["Haridevan, Amal Dev"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Shan, Jinjun"],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025-11-11","date_published":"2025-11-11","updated_at":"2026-07-24T06:33:55Z","subjects":["Aerospace engineering","Artificial intelligence","Robotics"],"languages":["en"],"rights":["Author owns copyright, except where explicitly noted. Please contact the author directly with licensing requests."],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/10315/43306","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Shan, Jinjun"]},{"key":"dc:creator","label":"Author","values":["Haridevan, Amal Dev"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2025-11-11T20:05:13Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2025-11-11T20:05:13Z"]},{"key":"dc:date.issued","label":"Date","values":["2025-11-11"]},{"key":"dc:type","label":"Dc Type","values":["Electronic Thesis or Dissertation"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Aerospace engineering","Artificial intelligence","Robotics"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Author owns copyright, except where explicitly noted. Please contact the author directly with licensing requests."]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/10315/43306"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Simulation tools and digital twins are vital to UAV research, enabling efficient prototyping, testing, and deployment of algorithms in dynamics, control, computer vision, and deep learning. Despite their importance, existing simulators face limitations in modularity, scalability, and multidisciplinary integration, while the persistent sim-to-real gap remains a major challenge. Nevertheless, simulators are indispensable, particularly for deep reinforcement learning (DRL), where extensive data collection in real-world settings is infeasible due to hardware and safety constraints. This thesis presents a modular, scalable, and transferable simulation platform for quadrotor UAVs, designed to support research in control systems, DRL, and embodied AI. The platform incorporates a high-fidelity digital twin, modeled using Blade Element Theory, with parameterized aerodynamic forces tailored to the Quanser QDrone2 and adaptable to other quadrotor architectures. A dedicated Control library, built within the Gazebo ecosystem, facilitates rapid prototyping and benchmarking of UAV controllers under a unified framework. To extend applicability to autonomous navigation, computer vision, and state estimation, Gazebo plugins are developed to enable efficient same-process interfacing and integration with both ROS1 and ROS2, thereby ensuring seamless transfer of algorithms from simulation to hardware. To advance DRL research, a suite of deterministic and stochastic environments is introduced. These environments balance reproducibility and realism by modeling communication delays, timing variations, and other uncertainties, while remaining compatible with widely used DRL libraries. Their parallelizable design enables large-scale data collection within hours, significantly accelerating training. Experimental validation confirms the platform’s fidelity and effectiveness in bridging the gap between simulation and real-world UAV deployment"]},{"key":"dc:title","label":"Title","values":["Digital Twin Platform for Drone Applications"]}]}],"canonical_facts":{"dc:contributor.advisor":["Shan, Jinjun"],"dc:creator":["Haridevan, Amal Dev"],"dc:date.accessioned":["2025-11-11T20:05:13Z"],"dc:date.available":["2025-11-11T20:05:13Z"],"dc:date.issued":["2025-11-11"],"dc:description.abstract":["Simulation tools and digital twins are vital to UAV research, enabling efficient prototyping, testing, and deployment of algorithms in dynamics, control, computer vision, and deep learning. Despite their importance, existing simulators face limitations in modularity, scalability, and multidisciplinary integration, while the persistent sim-to-real gap remains a major challenge. Nevertheless, simulators are indispensable, particularly for deep reinforcement learning (DRL), where extensive data collection in real-world settings is infeasible due to hardware and safety constraints. This thesis presents a modular, scalable, and transferable simulation platform for quadrotor UAVs, designed to support research in control systems, DRL, and embodied AI. The platform incorporates a high-fidelity digital twin, modeled using Blade Element Theory, with parameterized aerodynamic forces tailored to the Quanser QDrone2 and adaptable to other quadrotor architectures. A dedicated Control library, built within the Gazebo ecosystem, facilitates rapid prototyping and benchmarking of UAV controllers under a unified framework. To extend applicability to autonomous navigation, computer vision, and state estimation, Gazebo plugins are developed to enable efficient same-process interfacing and integration with both ROS1 and ROS2, thereby ensuring seamless transfer of algorithms from simulation to hardware. To advance DRL research, a suite of deterministic and stochastic environments is introduced. These environments balance reproducibility and realism by modeling communication delays, timing variations, and other uncertainties, while remaining compatible with widely used DRL libraries. Their parallelizable design enables large-scale data collection within hours, significantly accelerating training. Experimental validation confirms the platform’s fidelity and effectiveness in bridging the gap between simulation and real-world UAV deployment"],"dc:identifier.uri":["https://hdl.handle.net/10315/43306"],"dc:language":["en"],"dc:rights":["Author owns copyright, except where explicitly noted. Please contact the author directly with licensing requests."],"dc:subject":["Aerospace engineering","Artificial intelligence","Robotics"],"dc:title":["Digital Twin Platform for Drone Applications"],"dc:type":["Electronic Thesis or Dissertation"]},"updated_at":"2026-07-24T06:33:55Z"}