cs4home_person_tracker

Person tracking system using multi-sensor fusion (laser-based leg detection + YOLO vision) with the cs4home architecture prototype.

Prerequisites

• YOLO ROS
• State Observers
• UPO Laser People Detector
• CS4Home Architecture

Installation

cd ~/ros2_ws/src
git clone https://github.com/CoreSenseEU/cs4home_vision_module.git
vcs import --recursive < cs4home_vision_module/thirparty.repos
cd ~/ros2_ws
colcon build

Launch

Step 1: Launch YOLO ROS

ros2 launch yolo_bringup yolo.launch.py

Configure and activate the YOLO nodes:

ros2 lifecycle set /yolo_node configure
ros2 lifecycle set /yolo_node activate
ros2 lifecycle set /yolo_3d_node configure
ros2 lifecycle set /yolo_3d_node activate

Step 2: Launch Navigation or Provide Map Transform

You need to have the map transform available. Either:

• Launch your navigation stack, or
• Change target_frame_ in PersonTrackerCore.cpp from "map" to "base_link" or your preferred frame

Step 3: Launch UPO Laser People Detector

Follow the instructions from the UPO Laser People Detector repository.

Run the laser model node (adjust paths and topic names for your setup):

ros2 run upo_laser_people_detector lasermodelnode --ros-args \
  -p model_file:=/home/juan/Downloads/LFE-PPN.onnx \
  -p laser_topic:=/scan_raw

Note: Replace /home/juan/Downloads/LFE-PPN.onnx with your actual model file path and /scan_raw with your laser topic.

Step 4: Launch Person Tracker

ros2 launch cs4home_person_tracker person_tracker.launch.py

Configure and activate the person_tracker cognitive module:

ros2 lifecycle set /person_tracker configure
ros2 lifecycle set /person_tracker activate

Configuration

The configuration for the person tracker cognitive module is located in the package's config directory.