The RoboSense Challenge 2025

👋 Welcome to Track #5: Cross-Platform 3D Object Detection of the 2025 RoboSense Challenge!

With the rise of robotics, LiDAR-based 3D object detection has garnered significant attention in both academia and industry. However, existing datasets and methods predominantly focus on vehicle-mounted platforms, leaving other autonomous platforms underexplored.

This track encourages participants to develop novel cross-platform adaptation framework that transfers knowledge from the well-studied vehicle platform to other platforms, including drones and quadruped robots.

🏆 Prize Pool: $2,000 USD (1st: $1,000, 2nd: $600, 3rd: $400) + Innovation Awards

As robotics continues to advance, LiDAR-based 3D object detection has become a focal point in both academia and industry. However, most existing datasets and methods target vehicle platforms, overlooking quadrupeds and drones. This challenge, built on our benchmark, aims to:

1. Build on three robot platforms - vehicles, drones, and quadruped robots - to foster innovations in a unified LiDAR-based 3D object detection framework;
2. Bridge geometric and data distribution disparities to achieve rapid knowledge transfer and model adaptation across different robot platforms;
3. Lower annotation and deployment overhead, supporting collaborative sensing for heterogeneous robot teams in urban, disaster, and indoor scenarios.

Phase #1: Adaptation from Vehicle → Drone

Duration: June 15th, 2025 (anytime on earth) - August 15th, 2025 (anytime on earth)

Settings:

• Source platform: LiDAR scans with 3D bounding-box annotations from the Vehicle platform
• Target platform: Unlabeled LiDAR scans from the Drone platform

Ranking Metric: AP@0.50 (R40) for the Car class evaluated on Drone data.

Phase #2: Adaptation from Vehicle → Quadruped

Duration: August 15th, 2025 (anytime on earth) - September 15th, 2025 (anytime on earth)

Settings:

• Source platform: LiDAR scans with 3D bounding-box annotations from the Vehicle platform
• Target platforms: Unlabeled LiDAR scans from the Quadruped platform

Ranking Metric: A weighted score combining:

• AP@0.50 (R40) for the Car class
• AP@0.50 (R40) for the Pedestrian class

Note: Scores computed on Quadruped platforms.

In this track, we adopt PV-RCNN as the base 3D object detector, and leverage ST3D/++ as our baseline adaptation framework. Detailed environment setup and experimental protocols can be found in the Track5 GitHub repository .

Beyond the provided baseline, participants are encouraged to explore alternative strategies to further boost cross-platform performance:

• Treat the cross-platform challenge as a domain adaptation problem by improving pseudo-label quality and fine-tuning on target-platform data.
• Design novel data augmentation techniques to bridge geometric and feature discrepancies across platforms.
• Adopt geometry-agnostic 3D detectors, such as point-based architectures, that are less sensitive to platform-specific point-cloud characteristics.

Phase 1 Results

Phase 2 Results

We provide the following resources to support the development of models in this track:

Here, we provide a list of Frequently Asked Questions (FAQs) below for better clarity. If you have additional questions on the details about this competition, please reach out at robosense2025@gmail.com.

📖 References

@misc{robosense2025track5,
  title        = {RoboSense Challenge 2025: Track 5 - Cross-Platform 3D Object Detection},
  author       = {RoboSense Challenge 2025 Organizers},
  year         = {2025},
  howpublished = {https://robosense2025.github.io/track5}
}