SAGA: A Robust Self-Attention and Goal-Aware Anchor-based Planner for Safe UAV Autonomous Navigation

2026-05-04

AI Summary

This work addresses the challenge of enabling high-speed, safe, and efficient autonomous navigation for unmanned aerial vehicles in complex environments by proposing a one-stage joint regression-and-ranking planning framework based on fixed motion anchors. The method simultaneously predicts optimized terminal states and planning scores for each anchor in a single forward pass and decodes dynamically feasible trajectories. It introduces geometric-aware tokens and a self-attention mechanism to enable global cross-anchor reasoning, employs polar coordinate-based positional encoding to preserve directional structure, and integrates a goal-aware modulation module that injects velocity, acceleration, and target information. Experimental results demonstrate that the approach achieves a 100% task success rate at a maximum speed of 4.0 m/s, maintains a minimum safety distance of 0.7576 m, and reduces average flight time to 27.49 s, significantly outperforming existing methods such as YOPO.

Abstract

Agile unmanned aerial vehicle (UAV) navigation in cluttered environments demands a planning architecture that is both computationally efficient and structurally expressive enough to reason over multiple feasible motions. This paper presents SAGA, a robust self-attention and goal-aware anchor-based planner for safe UAV autonomous navigation. SAGA formulates local planning as a one-stage joint regression-and-ranking problem over a fixed lattice of motion anchors. Given a depth image and a body-frame motion state, the planner predicts refined terminal states and planning scores for all anchors in a single forward pass, after which the best candidate is decoded into a dynamically feasible trajectory. The key idea of SAGA is to transform anchor-aligned features into geometry-aware tokens and perform cross-anchor global reasoning with self-attention. To preserve directional structure in the token space, we further introduce a polar positional encoding derived from anchor yaw and pitch. In addition, a goal-aware modulation module injects velocity, acceleration, and target information into the token representation before final score prediction. Experiments in cluttered pillar-map environments under maximum speed settings of 2.0, 3.0, and 4.0~m/s show that SAGA consistently achieves a 100\% success rate, while YOPO drops from 90.91\% to 62.50\%, Ego-planner from 71.43\% to 52.63\%, and Fast-planner from 52.63\% to 38.46\%. Under the 4.0~m/s maximum speed setting, SAGA also improves average safety from 1.9843~m to 2.3888~m and minimum safety from 0.4390~m to 0.7576~m over YOPO, while reducing total flight time from 40.4631~s to 27.4901~s. The comparison with SAGA w/o PPE further shows that explicit polar positional encoding is critical for stable cross-anchor reasoning and safe passage selection in cluttered scenes.

Problem

• UAV autonomous navigation
• motion planning
• cluttered environments
• safety
• high-speed navigation

Innovation

• self-attention
• goal-aware planning
• motion anchors
• polar positional encoding
• UAV navigation