E-SocialNav: Efficient Socially Compliant Navigation with Language Models

2026-03-21

AI Summary

This work addresses the limitations of existing language models in robot navigation, which often neglect social compliance and incur high computational costs, hindering deployment on resource-constrained platforms. To overcome these challenges, the authors propose E-SocialNav—a lightweight, efficient language model specifically designed for socially compliant navigation. Leveraging a two-stage training strategy combining supervised fine-tuning and direct preference optimization, E-SocialNav achieves superior performance over zero-shot baselines with only limited data, excelling in both semantic plausibility and action accuracy. Experimental results demonstrate that E-SocialNav significantly outperforms large foundation models such as GPT-4o and Claude in generating socially appropriate behaviors, while maintaining high efficiency and practicality for real-world robotic deployment.

Technology Category

• Natural Language Processing: Safety and Robustness
• Planning, Routing, and Scheduling: Planning with Language Models
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Search and Retrieval-Augmented AI: Large language models for search
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Language models (LMs) are increasingly applied to robotic navigation; however, existing benchmarks primarily emphasize navigation success rates while paying limited attention to social compliance. Moreover, relying on large-scale LMs can raise efficiency concerns, as their heavy computational overhead leads to slower response times and higher energy consumption, making them impractical for real-time deployment on resource-constrained robotic platforms. In this work, we evaluate the social compliance of GPT-4o and Claude in robotic navigation and propose E-SocialNav, an efficient LM designed for socially compliant navigation. Despite being trained on a relatively small dataset, E-SocialNav consistently outperforms zero-shot baselines in generating socially compliant behaviors. By employing a two-stage training pipeline consisting of supervised fine-tuning followed by direct preference optimization, E-SocialNav achieves strong performance in both text-level semantic similarity to human annotations and action accuracy. The source code is available at https://github.com/Dr-LingXiao/ESocialNav.

Problem

• socially compliant navigation
• language models
• robotic navigation
• computational efficiency
• real-time deployment

Innovation

• Socially Compliant Navigation
• Efficient Language Model
• Direct Preference Optimization
• Two-stage Training
• Robotic Navigation