Deploying SICNav in the Field: Safe and Interactive Crowd Navigation using MPC and Bilevel Optimization

2025-06-10

AI Summary

To address robot stalling and entrapment in dense human crowds—caused by neglecting closed-loop human-robot interaction—this paper proposes a two-layer Model Predictive Control (MPC) framework: an upper layer optimizes robot trajectory planning, while a lower layer explicitly models human responses to robot actions, enabling joint prediction and planning. The method integrates multi-agent interaction modeling with real-time embedded motion planning, and is validated in unknown indoor and outdoor environments over a 7-kilometer, 2-hour end-to-end autonomous navigation mission. Experiments demonstrate significant improvements in passage efficiency and more natural, socially compliant avoidance behavior compared to conventional decoupled approaches. Critically, the system achieves zero collisions and zero prolonged standstills throughout all trials. This work marks the first systematic deployment of truly mutual human–robot adaptive navigation in real-world dynamic crowd settings.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Intelligent Robots: Human-Robot Interaction
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Safe and efficient navigation in crowded environments remains a critical challenge for robots that provide a variety of service tasks such as food delivery or autonomous wheelchair mobility. Classical robot crowd navigation methods decouple human motion prediction from robot motion planning, which neglects the closed-loop interactions between humans and robots. This lack of a model for human reactions to the robot plan (e.g. moving out of the way) can cause the robot to get stuck. Our proposed Safe and Interactive Crowd Navigation (SICNav) method is a bilevel Model Predictive Control (MPC) framework that combines prediction and planning into one optimization problem, explicitly modeling interactions among agents. In this paper, we present a systems overview of the crowd navigation platform we use to deploy SICNav in previously unseen indoor and outdoor environments. We provide a preliminary analysis of the system's operation over the course of nearly 7 km of autonomous navigation over two hours in both indoor and outdoor environments.

Problem

• Safe navigation in crowded environments for service robots
• Modeling human-robot interactions in motion planning
• Combining prediction and planning into one optimization problem

Innovation

• Bilevel MPC for crowd navigation
• Combines prediction and planning optimization
• Explicitly models human-robot interactions