Cognitive modules with cs4home_architecture

This tutorial builds and runs the example cognitive modules of cs4home_examples, which use the ROS 2 implementation of the CoreSense architecture, cs4home_architecture. See CoreSense design for the structure of a cognitive module.

The repository contains two examples:

• Face module: face identities from hri_face_detect are converted into knowledge-graph updates.

• YOLO module: camera images are processed by YOLO and published as detections.

Build

Requires ROS 2 and vcstool.

mkdir -p ~/cs4home_examples_ws/src
cd ~/cs4home_examples_ws/src
git clone https://github.com/CoreSenseEU/cs4home_examples.git
vcs import --recursive < cs4home_examples/thirdparty.repos
cd ..
python3 -m venv --system-site-packages py_deps
source py_deps/bin/activate
pip install -r src/thirdparty/hri_face_detect/requirements.txt
pip install -r src/thirdparty/yolov8_ros/requirements.txt
rosdep install --from-paths src --ignore-src -r -y
colcon build --symlink-install
source install/setup.bash

Run the YOLO module

Start a camera driver, then:

ros2 launch cs4home_simple_project yolo_example.launch.py

Run the face module

Start hri_face_detect, then run the module with its parameters:

ros2 run cs4home_simple_project face_module --ros-args \
  --params-file src/cs4home_examples/cs4home_simple_project/config/params_simple.yaml

The result can be inspected with the knowledge-graph RQt plugin.

Create your own module

cs_functional_module_template is a starting point for a new module. More complete examples are the perception modules of the social testbed: cs4home_sound_module and cs4home_vision_module.