Autonomous Ground Navigation in Highly Constrained Spaces: Lessons learned from The 3rd BARN Challenge at ICRA 2024

TL;DR

The paper investigates autonomous ground navigation in highly constrained environments through The 3rd BARN Challenge at ICRA 2024, combining a simulation qualifier and physical finals to benchmark current state-of-the-art stacks. It highlights three top approaches: LiCS-KI's Transformer-based end-to-end local planning with a safety layer and BC training, MLDA_EEE's MPC with mode-switching, and AIMS's geometry-based obstacle detection with backward sampling; collectively, these demonstrate a strong trend toward hybrid architectures and practical safety enhancements. Key findings include a first physical win for end-to-end learning, the inaugural use of Transformers in the challenge, and demonstrated successful sim-to-real transfer, underscoring the practical viability of learned components when coupled with robust safety and planning layers. The work provides actionable insights for robust navigation in cluttered settings and informs future competition design, particularly regarding participation diversity and dynamic obstacle handling.

Abstract

The 3rd BARN (Benchmark Autonomous Robot Navigation) Challenge took place at the 2024 IEEE International Conference on Robotics and Automation (ICRA 2024) in Yokohama, Japan and continued to evaluate the performance of state-of-the-art autonomous ground navigation systems in highly constrained environments. Similar to the trend in The 1st and 2nd BARN Challenge at ICRA 2022 and 2023 in Philadelphia (North America) and London (Europe), The 3rd BARN Challenge in Yokohama (Asia) became more regional, i.e., mostly Asian teams participated. The size of the competition has slightly shrunk (six simulation teams, four of which were invited to the physical competition). The competition results, compared to last two years, suggest that the field has adopted new machine learning approaches while at the same time slightly converged to a few common practices. However, the regional nature of the physical participants suggests a challenge to promote wider participation all over the world and provide more resources to travel to the venue. In this article, we discuss the challenge, the approaches used by the three winning teams, and lessons learned to direct future research and competitions.

Figures (6)

• Figure 1: Four example BARN environments in the Gazebo simulator (ordered by ascending relative difficulty level).
• Figure 2: Final physical competition participants and organizers at The 3rd BARN Challenge in Yokohama, Japan.
• Figure 3: Transformer-based neural network used in LiCS damanik2024lics.
• Figure 4: Safety zone illustration for the safety check layer of LiCS during linear and radial movement.
• Figure 5: Rviz visualization of obstacles in the MPC. White squares: obstacle coordinates sampled from raw LiDAR scan. Yellow squares: blind spot obstacle coordinates obtained from local_costmap.