Bench2FreeAD: A Benchmark for Vision-based End-to-end Navigation in Unstructured Robotic Environments

2025-03-15

AI Summary

Existing end-to-end (E2E) autonomous driving methods are largely confined to structured road environments and lack support for navigation in unstructured settings—such as campuses, indoor spaces, and service roads. To address this limitation, we introduce FreeWorld, the first E2E robot navigation benchmark explicitly designed for unstructured environments, comprising multi-source, heterogeneous data from both Isaac Sim simulation and real-robot deployments. We propose a dedicated E2E navigation evaluation framework tailored to unstructured scenarios and conduct cross-domain generalization validation using vision-driven autonomous navigation (VAD) models. Experimental results demonstrate that FreeWorld significantly enhances model robustness and adaptability in complex, structure-free environments. All datasets, source code, and evaluation tools are publicly released on GitHub, establishing a reproducible, scalable, and standardized benchmark for logistics and service robotics research.

Technology Category

• Intelligent Robots: Localization, Mapping, and Navigation
• Computer Vision: Vision for Robotics & Autonomous Driving
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Most current end-to-end (E2E) autonomous driving algorithms are built on standard vehicles in structured transportation scenarios, lacking exploration of robot navigation for unstructured scenarios such as auxiliary roads, campus roads, and indoor settings. This paper investigates E2E robot navigation in unstructured road environments. First, we introduce two data collection pipelines - one for real-world robot data and another for synthetic data generated using the Isaac Sim simulator, which together produce an unstructured robotics navigation dataset -- FreeWorld Dataset. Second, we fine-tuned an efficient E2E autonomous driving model -- VAD -- using our datasets to validate the performance and adaptability of E2E autonomous driving models in these environments. Results demonstrate that fine-tuning through our datasets significantly enhances the navigation potential of E2E autonomous driving models in unstructured robotic environments. Thus, this paper presents the first dataset targeting E2E robot navigation tasks in unstructured scenarios, and provides a benchmark based on vision-based E2E autonomous driving algorithms to facilitate the development of E2E navigation technology for logistics and service robots. The project is available on Github.

Problem

• Explores end-to-end robot navigation in unstructured environments.
• Introduces FreeWorld Dataset for unstructured robotics navigation.
• Validates E2E autonomous driving models in unstructured scenarios.

Innovation

• Developed real-world and synthetic data collection pipelines
• Fine-tuned VAD model for unstructured environments
• Created FreeWorld Dataset for E2E navigation benchmarks