Lidar people detection and tracking libraries
Set of libraries and ROS nodes to detect and track people using 2D lidar
Main Characteristic
The set of libraries consists in:
- a people detection C++ library, with its ROS wrapper node, which subscribes to LaserScan messages and publishes detected people.
- a people tracking C++ library, with its ROS wrapper node, which subscribes to detected people and publishes tracked people with IDs
- an intermediate ROS node which filters detections using a map (OccupancyGrid in ROS) by neighbour distance to obstacles. It subscribes to detected people and a map, and publishes filtered people detections.
Technical Categories
Robotics and automation
Last updated
25.05.2021 - 16:23
Detailed Description
Hardware architecture: X64
Format: DockerFile based on osrf/ros:kinetic-desktop-full, adding ROS workspace with libraries and ROS packages
Additional information: All can be tested with an example rosbag included.
Trustworthy AI
The libraries included in this resource are, (1) lawful, as they respect all applicable laws and regulations, especially it is GDPR-compliant, (2) ethical, as they use anonymous data to identify and be aware of surrounding people position and behave accordingly, (3) and robust, from a technical perspective, especially as it is deployed in a "ready-to-use" Docker container.
GDPR Requirements
It is GDPR-compliant as all data used is anonymous (lidar ranges).