Follow Everything: A Leader-Following and Obstacle Avoidance Framework with Goal-Aware Adaptation

2025-04-28

AI Summary

To address poor generalization—specifically, the inability to adapt to heterogeneous leaders (e.g., humans, UAVs, landmarks)—and insufficient robustness during visual occlusion in robot leader-follower navigation, this paper proposes a universal leader-following framework. Methodologically, it introduces (1) a novel distance-frame buffering mechanism that dynamically stores multi-scale distance-aware embeddings, and (2) a target-aware adaptive state machine that tightly couples graph-search-based trajectory planning with graph neural network-driven motion generation, enabling persistent tracking and real-time obstacle avoidance under visual interruption. The framework integrates segmentation-driven leader identification and end-to-end vision–motion co-control. Extensive simulations and real-robot experiments demonstrate significant improvements: 23.6% higher follower success rate, 57.1% reduction in visual occlusion duration, 41.3% lower collision rate, and 19.8% shorter average leader–follower distance.

Technology Category

• Intelligent Robots: Localization, Mapping, and Navigation
• Computer Vision: Motion & Tracking
• Planning, Routing, and Scheduling: Replanning and Plan Repair

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Robust and flexible leader-following is a critical capability for robots to integrate into human society. While existing methods struggle to generalize to leaders of arbitrary form and often fail when the leader temporarily leaves the robot's field of view, this work introduces a unified framework addressing both challenges. First, traditional detection models are replaced with a segmentation model, allowing the leader to be anything. To enhance recognition robustness, a distance frame buffer is implemented that stores leader embeddings at multiple distances, accounting for the unique characteristics of leader-following tasks. Second, a goal-aware adaptation mechanism is designed to govern robot planning states based on the leader's visibility and motion, complemented by a graph-based planner that generates candidate trajectories for each state, ensuring efficient following with obstacle avoidance. Simulations and real-world experiments with a legged robot follower and various leaders (human, ground robot, UAV, legged robot, stop sign) in both indoor and outdoor environments show competitive improvements in follow success rate, reduced visual loss duration, lower collision rate, and decreased leader-follower distance.

Problem

• Generalizing leader-following to arbitrary leader forms
• Enhancing robustness when leader leaves robot's view
• Ensuring efficient following with obstacle avoidance

Innovation

• Segmentation model replaces traditional detection methods
• Distance frame buffer enhances leader recognition robustness
• Goal-aware adaptation mechanism with graph-based planner