AI & ML
A Hierarchical Onboard Vision-Language-Navigation Framework for Aerial Robots in Complex Environments
arXiv cs.RO0 views
arXiv:2611.23930v1 Announce Type: new Abstract: Unmanned Aerial Vehicles (UAVs) are expected to achieve semantic reasoning, intelligent decision-making, and long-horizon navigation under strict onboard constraints. While recent foundation-model-driven approaches have shown promising progress toward this goal, existing methods remain limited by challenges in 3D data availability, trajectory feasibility, long-horizon reasoning, and computational efficiency in complex environments. To bridge these gaps, we propose VLN-AR, a hierarchical vision-language-navigation (VLN) framework for aerial robots. VLN-AR decouples multimodal reasoning and low-level control, enabling structured prediction of 3D pose commands and navigation intentions that are translated into dynamically feasible actions. To mitigate data scarcity and domain gaps, we introduce a fully automated 3D Gaussian Splatting (3DGS)-based pipeline to generate scalable photorealistic UAV navigation data. A progressive three-stage training strategy further enhances grounding, scene reasoning, and complex navigation. In addition, a lightweight cognitive agent improves long-horizon reasoning via memory-aware task decomposition. Finally, we design onboard optimization techniques, including token sparsification and model lightweighting, achieving 7.1x-9.4x speedup while maintaining high navigation performance on resource-limited UAV platforms. Experiments show great improvements over existing methods in both simulation and real-world UAV navigation. Project website: https://wyz000.github.io/VLN-AR/
Source:https://arxiv.org/abs/2611.23930PreviousGTI: Grounding Task Intent Throughout Long-Horizon VLA ManipulationNextSecuring and Auditing Long-Horizon LLM Agents: State-Transfer Auditing and Controls in the Agent Harness
Related posts
Introducing CustomerLake Conversions: Improve advertising performance with first-party customer context
0Databricks Blog
Bulk migrate your dashboards to AI/BI with Genie Code
0Databricks Blog
Four AI Architecture Patterns for Regulated Industries
0Databricks Blog

lululemon pilots store feedback with Amazon Bedrock AgentCore and AWS Quick
0AWS for Industries
Thống kê đối đầu Genoa vs Sassuolo: Sự vượt trội về hiệu suất và tính ổn định
0dev.to Data
Thách thức dữ liệu phòng ngự của HLV Thomas Frank tại Crystal Palace
0dev.to Data