# uavs Published articles for uavs. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Symmetry-Aware Surrogate-Assisted Testing of UAVs DevFeed: [Symmetry-Aware Surrogate-Assisted Testing of UAVs]() Original publisher: [Read original article]() Author: Anas Shrinah, Chris Bennett, Kerstin Eder Published: 2026-11-06T05:00:00Z Content type: article Language: en Sources: [arXiv: Software Engineering]() Topics: [Test automation](), [Neural Network](), [reproducibility]() Tags: [canonicalization](), [cs-ro](), [cs-se](), [dataset](), [diversity](), [neural-network](), [safety](), [simulation](), [simulations](), [symmetry](), [test](), [testing](), [training](), [uavs]() ### AI overview The approach reuses a neural network trained on one reference mission to guide UAV autopilot testing across missions. It maps obstacle configurations into a shared canonical coordinate frame and predicts minimum clearance to steer test generation toward safety-critical cases, reducing the high-fidelity simulation data needed for training. On the Aerialist test bench, it achieved a 42% average critical-test-case hit rate across evaluated missions, versus about 10% for the constrained-random training-data baseline, with high input diversity and competitive benchmark performance. Prediction transfer was limited when flight-path characteristics differed. The authors provide the trained model and a dataset of more than 2,500 high-fidelity simulations. ### Source excerpt arXiv:2611.08012v1 Announce Type: cross Abstract: We investigate a symmetry-aware surrogate-assisted testing approach for UAV autopilots. The proposed approach exploits geometric symmetries between flight missions by trans- forming obstacle configurations into a shared canonical coordi- nate frame. This enables a neural-network surrogate trained on a single reference mission to be reused across arbitrary missions, reducing the amount of high-fidelity simulation data required for training. The surrogate model is trained to predict the minimum clearance distance between a UAV and surrounding obstacles and is used to guide test generation towards safety-critical scenarios. We evaluate this approach using the Aerialist UAV test bench and benchmark its performance against techniques from the 2024 to 2026 UAV Testing Competitions. Results show that the approach achieves an average critical-test-case hit rate of 42% across the evaluated missions, compared with approximately 10% for the constrained-random baseline used to generate the training data. This approach also achieves high input diversity and competitive benchmark performance. However, our findings reveal limitations in transferring surrogate predictions across missions with differing flight-path characteristics, highlighting the boundaries of geometric canonicalization. To support reproducibility and future research, we provide the trained surrogate model and a dataset comprising more than 2,500 high-fidelity simulations. ## How Drones Navigate Without GPS: A Builder's Guide to Visual SLAM in 2026 DevFeed: [How Drones Navigate Without GPS: A Builder's Guide to Visual SLAM in 2026]() Original publisher: [Read original article]() Author: vishwa vimukthi gammuduwaththage Published: 2026-10-14T19:45:01Z Content type: tutorial Language: en Sources: [HackerNoon]() Topics: [Robot Manipulation](), [Robot Navigation](), [Holoscan]() Tags: [autonomous-drones](), [computer-vision](), [drone-technology](), [drones](), [edge-ai](), [edge-ai-for-drones](), [gps-denied-navigation](), [slam](), [uavs](), [visual-inertial-odometry](), [visual-slam]() ### AI overview This builder's guide explains how visual SLAM can support drone navigation when GPS is unavailable. It covers camera-based feature tracking, visual-inertial fusion, mapping and loop closure, compute and path-planning choices, and challenges such as motion blur, poor lighting, vibration, and weak visual texture. It describes a reference architecture while noting it has not been flight-tested, and discusses applications, recent developments, and the case for combining vision with other navigation modes. ### Source excerpt Explore how visual SLAM enables GPS-denied drone navigation, from visual-inertial odometry and mapping to edge AI, path planning, and limitations. ## Seeing the Threat: Why Counter-UAS Needs an AI Brain That Detects Aerial Threats With Vision Alone DevFeed: [Seeing the Threat: Why Counter-UAS Needs an AI Brain That Detects Aerial Threats With Vision Alone]() Original publisher: [Read original article]() Author: vishwa vimukthi gammuduwaththage Published: 2026-10-13T20:18:47Z Content type: opinion Language: en Sources: [HackerNoon]() Topics: [Computer vision](), [Artificial Intelligence](), [object-detection](), [Physical AI]() Tags: [ai](), [ai-in-warfare](), [artificial-intelligence](), [computer-vision](), [counter-uas](), [drone](), [drone-technology](), [drones](), [uavs](), [vision]() ### AI overview The article argues that computer vision can serve as the primary detection layer for counter-UAS systems, particularly as autonomous and fiber-optic drones evade radio-based detection. It describes requirements for real-time edge systems, including detecting, tracking, and classifying aerial objects while keeping false alarms low and integrating results with command-and-control systems. ### Source excerpt Why counter drone drone and modern aerial threat need a AI brain to detect and neutralize them by using computer vision, mainly