Hector Tovanche-Picon, Javier González-Trejo, Ángel Flores-Abad, Miguel Ángel García-Terán, Diego Mercado-Ravell
{"title":"Real-time safe validation of autonomous landing in populated areas: from virtual environments to Robot-In-The-Loop","authors":"Hector Tovanche-Picon, Javier González-Trejo, Ángel Flores-Abad, Miguel Ángel García-Terán, Diego Mercado-Ravell","doi":"10.1007/s10055-024-00965-6","DOIUrl":null,"url":null,"abstract":"<p>Safe autonomous landing for Unmanned Aerial Vehicles (UAVs) in populated areas is a crucial aspect for successful integration of UAVs in populated environments. Nonetheless, validating autonomous landing in real scenarios is a challenging task with a high risk of injuring people. In this work, we propose a framework for safe real-time and thorough evaluation of vision-based autonomous landing in populated scenarios, using photo-realistic virtual environments and physics-based simulation. The proposed evaluation pipeline includes the use of Unreal graphics engine coupled with AirSim for realistic drone simulation to evaluate landing strategies. Then, Software-/Hardware-In-The-Loop can be used to test beforehand the performance of the algorithms. The final validation stage consists in a Robot-In-The-Loop evaluation strategy where a real drone must perform autonomous landing maneuvers in real-time, with an avatar drone in a virtual environment mimicking its behavior, while the detection algorithms run in the virtual environment (virtual reality to the robot). This method determines the safe landing areas based on computer vision and convolutional neural networks to avoid colliding with people in static and dynamic scenarios. To test the robustness of the algorithms in adversary conditions, different urban-like environments were implemented, including moving agents and different weather conditions. We also propose different metrics to quantify the performance of the landing strategies, establishing a baseline for comparison with future works on this challenging task, and analyze them through several randomized iterations. The proposed approach allowed us to safely validate the autonomous landing strategies, providing an evaluation pipeline, and a benchmark for comparison. An extensive evaluation showed a 99% success rate in static scenarios and 87% in dynamic cases, demonstrating that the use of autonomous landing algorithms considerably prevents accidents involving humans, facilitating the integration of drones in human-populated spaces, which may help to unleash the full potential of drones in urban environments. Besides, this type of development helps to increase the safety of drone operations, which would advance drone flight regulations and allow their use in closer proximity to humans.</p>","PeriodicalId":23727,"journal":{"name":"Virtual Reality","volume":null,"pages":null},"PeriodicalIF":4.4000,"publicationDate":"2024-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Virtual Reality","FirstCategoryId":"94","ListUrlMain":"https://doi.org/10.1007/s10055-024-00965-6","RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS","Score":null,"Total":0}
引用次数: 0
Abstract
Safe autonomous landing for Unmanned Aerial Vehicles (UAVs) in populated areas is a crucial aspect for successful integration of UAVs in populated environments. Nonetheless, validating autonomous landing in real scenarios is a challenging task with a high risk of injuring people. In this work, we propose a framework for safe real-time and thorough evaluation of vision-based autonomous landing in populated scenarios, using photo-realistic virtual environments and physics-based simulation. The proposed evaluation pipeline includes the use of Unreal graphics engine coupled with AirSim for realistic drone simulation to evaluate landing strategies. Then, Software-/Hardware-In-The-Loop can be used to test beforehand the performance of the algorithms. The final validation stage consists in a Robot-In-The-Loop evaluation strategy where a real drone must perform autonomous landing maneuvers in real-time, with an avatar drone in a virtual environment mimicking its behavior, while the detection algorithms run in the virtual environment (virtual reality to the robot). This method determines the safe landing areas based on computer vision and convolutional neural networks to avoid colliding with people in static and dynamic scenarios. To test the robustness of the algorithms in adversary conditions, different urban-like environments were implemented, including moving agents and different weather conditions. We also propose different metrics to quantify the performance of the landing strategies, establishing a baseline for comparison with future works on this challenging task, and analyze them through several randomized iterations. The proposed approach allowed us to safely validate the autonomous landing strategies, providing an evaluation pipeline, and a benchmark for comparison. An extensive evaluation showed a 99% success rate in static scenarios and 87% in dynamic cases, demonstrating that the use of autonomous landing algorithms considerably prevents accidents involving humans, facilitating the integration of drones in human-populated spaces, which may help to unleash the full potential of drones in urban environments. Besides, this type of development helps to increase the safety of drone operations, which would advance drone flight regulations and allow their use in closer proximity to humans.
期刊介绍:
The journal, established in 1995, publishes original research in Virtual Reality, Augmented and Mixed Reality that shapes and informs the community. The multidisciplinary nature of the field means that submissions are welcomed on a wide range of topics including, but not limited to:
Original research studies of Virtual Reality, Augmented Reality, Mixed Reality and real-time visualization applications
Development and evaluation of systems, tools, techniques and software that advance the field, including:
Display technologies, including Head Mounted Displays, simulators and immersive displays
Haptic technologies, including novel devices, interaction and rendering
Interaction management, including gesture control, eye gaze, biosensors and wearables
Tracking technologies
VR/AR/MR in medicine, including training, surgical simulation, rehabilitation, and tissue/organ modelling.
Impactful and original applications and studies of VR/AR/MR’s utility in areas such as manufacturing, business, telecommunications, arts, education, design, entertainment and defence
Research demonstrating new techniques and approaches to designing, building and evaluating virtual and augmented reality systems
Original research studies assessing the social, ethical, data or legal aspects of VR/AR/MR.