HumanDiffusion: A Vision-Based Diffusion Trajectory Planner with Human-Conditioned Goals for Search and Rescue UAV

2026-01-21

AI Summary

This work addresses the challenge of reliable human detection and safe navigation for drones in mapless, dynamic search-and-rescue scenarios. The authors propose a lightweight, image-conditioned diffusion-based trajectory planner that, for the first time, leverages human-conditioned guidance within a diffusion model to generate smooth, collision-free navigation trajectories directly in pixel space. The approach operates end-to-end without requiring prior maps or high computational resources, integrating YOLOv11 for human detection with the diffusion model to enable human-centric visual navigation. Evaluated on a 300-sample test set, the method achieves a trajectory reconstruction mean squared error of 0.02 and demonstrates an 80% success rate in real-world indoor disaster simulation tasks, confirming its practicality and robustness.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Search and Optimization: Learning to Search

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Reliable human--robot collaboration in emergency scenarios requires autonomous systems that can detect humans, infer navigation goals, and operate safely in dynamic environments. This paper presents HumanDiffusion, a lightweight image-conditioned diffusion planner that generates human-aware navigation trajectories directly from RGB imagery. The system combines YOLO-11 based human detection with diffusion-driven trajectory generation, enabling a quadrotor to approach a target person and deliver medical assistance without relying on prior maps or computationally intensive planning pipelines. Trajectories are predicted in pixel space, ensuring smooth motion and a consistent safety margin around humans. We evaluate HumanDiffusion in simulation and real-world indoor mock-disaster scenarios. On a 300-sample test set, the model achieves a mean squared error of 0.02 in pixel-space trajectory reconstruction. Real-world experiments demonstrate an overall mission success rate of 80% across accident-response and search-and-locate tasks with partial occlusions. These results indicate that human-conditioned diffusion planning offers a practical and robust solution for human-aware UAV navigation in time-critical assistance settings.

Problem

• human-aware navigation
• search and rescue
• UAV trajectory planning
• vision-based planning
• emergency response

Innovation

• diffusion-based planning
• human-conditioned navigation
• vision-based UAV control
• pixel-space trajectory generation
• search and rescue robotics