🤖 AI Summary
To address safety navigation and social compliance challenges for robots in dynamic social environments, this paper proposes an end-to-end safe policy learning framework integrating Adaptive Conformal Inference (ACI) with Constrained Reinforcement Learning (CRL). The method pioneers the incorporation of ACI into CRL, enabling real-time quantification of perception uncertainty and risk-aware policy optimization, thereby significantly enhancing out-of-distribution robustness. Leveraging multimodal observation enhancement and ROS2-based real-world deployment, the approach achieves a 96.93% success rate on the CrowdNav benchmark—surpassing the state-of-the-art by 11.67%—while reducing collision frequency by 4.5× and socially intrusive trajectory violations by 2.8×. Comprehensive evaluation across sparse and dense crowd scenarios validates its high safety assurance and social courtesy.
📝 Abstract
Reinforcement learning (RL) enables social robots to generate trajectories without relying on human-designed rules or interventions, making it generally more effective than rule-based systems in adapting to complex, dynamic real-world scenarios. However, social navigation is a safety-critical task that requires robots to avoid collisions with pedestrians, whereas existing RL-based solutions often fall short of ensuring safety in complex environments. In this paper, we propose SoNIC, which to the best of our knowledge is the first algorithm that integrates adaptive conformal inference (ACI) with constrained reinforcement learning (CRL) to enable safe policy learning for social navigation. Specifically, our method not only augments RL observations with ACI-generated nonconformity scores, which inform the agent of the quantified uncertainty but also employs these uncertainty estimates to effectively guide the behaviors of RL agents by using constrained reinforcement learning. This integration regulates the behaviors of RL agents and enables them to handle safety-critical situations. On the standard CrowdNav benchmark, our method achieves a success rate of 96.93%, which is 11.67% higher than the previous state-of-the-art RL method and results in 4.5 times fewer collisions and 2.8 times fewer intrusions to ground-truth human future trajectories as well as enhanced robustness in out-of-distribution scenarios. To further validate our approach, we deploy our algorithm on a real robot by developing a ROS2-based navigation system. Our experiments demonstrate that the system can generate robust and socially polite decision-making when interacting with both sparse and dense crowds. The video demos can be found on our project website: https://sonic-social-nav.github.io/.