AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

HTNav: A Hybrid Navigation Framework with Tiered Structure for Urban Aerial Vision-and-Language Navigation

Chengjie Fan, Cong Pan, Zijian Liu, Ningzhong Liu, Jie Qin

CVPR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Pretraining and data augmentation for VLN · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add IL-RL staged training (ablation)MGP · CityNav Test Unseen (revised)SR (%): 9.70 -> 18.95Improved
add tiered MacroPlanner/MicroActor structure (ablation)MGP+IL-RL · CityNav Test Unseen (revised)SR (%): 18.95 -> 24.03Improved
add tiered structure (ablation)MGP+IL-RL · CityNav Test Unseen (revised)NE (m): 47.8 -> 44.0Improved
add residual map encoder (ablation)HTNav tiered · CityNav Test Unseen (revised)SR (%): 24.03 -> 24.18No clear effect
add SCConv module (ablation)HTNav tiered · CityNav Test Unseen (revised)SR (%): 24.03 -> 24.16No clear effect
add residual encoder and SCConv together (ablation)HTNav tiered · CityNav Test Unseen (revised)SR (%): 24.03 -> 25.49Improved
add residual encoder and SCConv together (ablation)HTNav tiered · CityNav Validation Unseen (revised)SR (%): 15.09 -> 17.69Improved
increase RL weight from 0 to 0.20 (ablation)HTNav · CityNav Test Unseen (revised)SR (%): 18.90 -> 25.49Improved
increase RL weight from 0.20 to 0.30 (ablation)HTNav · CityNav Test Unseen (revised)SR (%): 25.49 -> 23.56Hurt
add landmark map input (ablation)HTNav · CityNav (Table 5 split)SR (%): 1.86 -> 28.30Improved
add target and surroundings maps (ablation)HTNav · CityNav (Table 5 split)SR (%): 28.30 -> 28.62No clear effect
use manually revised CityNav annotationsMGP · CityNav Test Unseen (revised)SR (%): 6.38 -> 9.70Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy