2D LiDAR People Detection for ROS 2

This ROS package contains nodes for detecting people using 2D LiDAR.

See also:

• 2D Laser People Benchmark (including LFE-PPN and LFE-Peaks source code)
• FROG dataset
• Paper (arXiv)

System requirements

• Ubuntu 22.04 Jammy
• ROS 2 Humble
• CUDA Toolkit 12.1 (tested)
• ONNX Runtime 1.16.3 (tested):

wget https://robotics.upo.es/~famozur/onnx/onnxruntime-gpu_1.16.3_amd64.deb
sudo apt install ./onnxruntime-gpu_1.16.3_amd64.deb

Nodes

This package contains two nodes: lasermodelhost (implementing LFE-PPN) and lasermodelhost_peaks (implementing LFE-Peaks). Each node can be launched using the following command:

ros2 run upo_laser_people_detector lasermodelhost{_peaks} --ros-args -p model_file:=some_model.onnx -p other_param:=value ...

List of parameters:

• model_file (string): Path to the ONNX file containing model weights. This parameter must be explicitly provided. Please look at the pre-trained models section below.
• laser_topic (string): Name of the input ROS topic containing sensor_msgs/LaserScan messages from the 2D LiDAR. Defaults to /scanfront.
• output_topic (string): Name of the output ROS topic for upo_laser_people_msgs/PersonDetectionList messages. Defaults to detected_people (namespace relative).
• marker_topic (string): Name of the output ROS topic for visualization_msgs/MarkerArray messages for use with RViz. Defaults to detected_people_markers (namespace relative).
• scan_near (float): Minimum distance between the 2D LiDAR and the person, in meters. Defaults to 0.02 m.
• scan_far (float): Maximum distance between the 2D LiDAR and the person, in meters. Defaults to 10 m.
• score_threshold (float): Score threshold for considering a person detection. Defaults to an appropriate value for each model.
• person_radius (float): Radius of the person bounding circles in meters (only for LFE-Peaks). Defaults to 0.4 m.

Pre-trained models

• LFE-PPN (for use with lasermodelhost)
• LFE-Peaks (for use with lasermodelhost_peaks)

Reference

@misc{frog2023,
	author = {Fernando Amodeo and Noé Pérez-Higueras and Luis Merino and Fernando Caballero},
	title = {FROG: A new people detection dataset for knee-high 2D range finders},
	year = {2023},
	eprint = {arXiv:2306.08531},
}

Acknowledgements

This work is partially funded by the Programa Operativo FEDER Andalucía 2014-2020, Consejería de Economía, Conocimiento y Universidades (DeepBot, PY20_00817) and by the projects NHoA PLEC2021-007868 and NORDIC TED2021-132476B-I00, funded by MCIN/AEI/10.13039/501100011033 and the European Union "NextGenerationEU"/"PRTR".