Summary
The integration of Unmanned Aerial Systems (UAS) with Artificial Intelligence (AI) has become crucial in addressing the challenges associated with civilian applications of UAVs, particularly within urban environments that pose significant risks of collision with obstacles. These UAVs are equipped with advanced sensor arrays, including LiDAR and computer vision technologies, which aid in developing a robust spatial perception model through comprehensive AI algorithm training on an embedded machine. This model empowers the UAV to navigate complex urban landscapes with a human-like understanding of its surroundings. During missions, the AI-driven perception system ensures real-time awareness by detecting and localizing objects. An innovative aspect of this research is the introduction of a real-time 3D path planner designed to optimize UAV trajectories amidst obstacle-heavy environments. This path planner utilizes a heuristic A* algorithm without the need to store frontier nodes in memory, distinguishing it from traditional implementations. It uses relative object positions from the perception system, incorporating techniques from simultaneous localization and mapping (SLAM) to ensure the creation of collision-free paths and enhancing navigational efficiency. Rigorous validation of the proposed path planner is performed through Software-In-The-Loop (SITL) simulations in constrained environments using high-fidelity UAV dynamics models. Preliminary real flight tests further assess the system's applicability in real-world scenarios, considering factors such as wind disturbances and dynamic obstacles. The results demonstrated the path planner's effectiveness in providing swift and accurate guidance, confirming its viability for real-time UAV missions in complex urban settings.
AI-powered UAVs use advanced sensors and a novel path planner to navigate urban environments efficiently. The planner optimizes trajectories without storing nodes, validated by simulations and real tests.
Cited By:
5
AI advances enable detailed 3D city models via generative approaches for better urban modeling. This method addresses data challenges and adapts to evolving city dynamics.
Published By:
Comput Urban Sci - Computational Urban Science
2024
Cited By:
7
3D dynamic landscape models using AI and GIS improve landscape planning with enhanced accuracy. The IDS-AI system achieves a 98.3% gradient sensitivity level and 93.4% area coverage ratio.
Published By:
Heliyon
2024
Cited By:
8
The study integrates SD, AI, and CityGML for urban planning, using GA for optimizing dorm allocations. It proposes strategic dorm placements in Berlin, balancing cost, accessibility, and transport.
Published By:
ISPRS Ann Photogramm Remote Sens Spat Inf Sci - ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
2024
Cited By:
2
A 3D urban traffic simulation platform was developed using Python and JavaScript, integrating with software like Sumo and Vissim. Users can easily access the platform to evaluate real-time traffic simulations, enhancing domestic traffic planning.
Published By:
Int J Front Eng Technol - International Journal of Frontiers in Engineering Technology
2022
Cited By:
4
CURB-OSG is a dynamic 3D scene graph engine for accurate multi-agent mapping and object prediction. It improves upon previous methods by collaborating multiple agents and doesn't rely solely on simulations.
Published By:
IROS - IEEE/RJS International Conference on Intelligent RObots and Systems
2025
Cited By:
8
A 3D animation system simulates facial expressions effectively, achieving a 95.40% simulation degree. Skin texture processing fidelity is 97.60%, enhancing virtual character authenticity.
Published By:
Mob Inf Syst - Mobile Information Systems
2022
Cited By:
3
Urban Digital Twins enhance urban efficiency by integrating AI, IoT, and Digital Twins. Challenges include interoperability, predictive modeling, and privacy.
Published By:
Smart City - Smart Cities
2025
Cited By:
29
IoT, AI, and digital twins enhance 3D scene completion and metaverse transactions. Collaborative 3D visualization boosts manufacturing processes for sustainable growth in IoT environments.
Published By:
Oeconomia Copernic - Oeconomia Copernicana
2025
Cited By:
37
Markerless 3D human motion tracking using AI and multi-camera views enhances accuracy and real-time performance. The framework integrates the WLS algorithm, showing readiness for real-world applications.
Published By:
ArXiv - arXiv.org
2023
Cited By:
13
The research develops a model for identifying potential rockfall zones using AI and 3D algorithms. Results show high-resolution susceptibility maps with up to 75% accuracy.
Published By:
Remote Sens - Remote Sensing
2023
Cited By:
11
3D reality captures ease digitization but lack full automation in modeling. AI integration could enhance architectural and cultural-heritage modeling using deep learning and real-time applications.
Published By:
heritage - The Heritage
2025
Cited By:
9
3D GIS flood visualisation helps identify vulnerable structures, emphasizing low-rise building risk. This approach aids in flood awareness and disaster response in urban areas.
Published By:
International journal of geoinformatics - International Journal of Geoinformatics
2026
Cited By:
0
AI4FOREST enhances autonomous forest monitoring using SLAM, 3D mapping, and deep learning for tree detection. It also introduces a model for improved navigation and examines point cloud density's impact on data accuracy.
Published By:
Machines
2026
Cited By:
1
The study presents a simulation-to-simulation framework for predicting path loss using ray tracing and geographic data, achieving an MAE of 4.8536 dB and RMSE of 9.3504 dB in a cross-city test. This method efficiently approximates RT-generated path loss maps under various conditions without relying on real-world measurements.
Published By:
Electronics
2026
Cited By:
0