Navigating the Proximity-Safety Balance: Constraint Decomposition for Human Following in Pedestrian Crowds

2026-08-10

AI Summary

This work addresses the challenge of target following in dense crowds, where maintaining proximity to the target must be balanced against collision avoidance—a trade-off that existing methods struggle to control explicitly. To this end, the authors propose a multi-constrained reinforcement learning framework that decomposes the task into a sparse reward and multiple cost constraints, each endowed with clear behavioral semantics. By introducing tunable thresholds on these constraints, the approach enables explicit and interpretable balancing between safety and closeness to the target. Furthermore, it incorporates uncertainty estimates from human motion prediction to enhance decision robustness. Experimental results demonstrate that the method achieves superior performance trade-offs both in-distribution and out-of-distribution, and real-world deployment validates its effectiveness in practical environments.

Abstract

Following a target human in crowded environments involves an inherent conflict between staying close to the target and navigating safely among surrounding pedestrians and obstacles. This conflict becomes more severe in dense scenarios, where aggressive following risks collisions and conservative margins lead to target loss, especially when pedestrian behaviors are unfamiliar or unpredictable. Existing reinforcement learning (RL) methods typically encode these competing objectives into a single dense reward, but the resulting proximity-safety balance is implicit and difficult to adjust across conditions. To address this, we decompose the human-following task into a sparse task reward and independent cost constraints within a multi-constraint RL formulation, where each constraint is managed through cost thresholds with direct behavioral meaning rather than implicit reward weight ratios, allowing explicit and tunable control over the trade-off. We further quantify the prediction uncertainty of human motions and integrate these estimates into the RL costs to enhance safety under unpredictable conditions. Extensive experiments across both in-distribution and out-of-distribution settings demonstrate that our method achieves an effective proximity-safety balance compared to baselines. Real-robot deployment further validates the feasibility of our method in real-world scenarios. More details are available on our project page: https://nav-ps-balance.github.io/.

Problem

• human following
• proximity-safety balance
• pedestrian crowds
• collision avoidance
• target loss

Innovation

• constraint decomposition
• multi-constraint reinforcement learning
• proximity-safety trade-off
• prediction uncertainty
• human-following navigation