Kaia.ai is a software platform for intelligent robots. Please read our blog for more.
Please sign up for a Kickstarter pre-launch invite here. We plan to launch on Kickstarter in 2nd half of 2024 via our Maker's Pet hardware partner's campaign.
Please visit the Support Forum!
Kaia.ai robotics software platform is actively evolving and currently consists of these parts:
- Micro-ROS Arduino library for Kaia.ai-compatible robots repo
- Micro-ROS LiDAR telemetry receiver package
- End-user and development ROS2 Docker images repo
- Robot Gazebo simulation ROS2 package
- Kaia.ai Python ROS2 software wrapper package
- Robot operation ROS2 repo, including SLAM mapping, navigation, frontier exploration, etc.
- Robot keyboard teleoperation package
- WebRTC-based image/video/data streaming
- Python-based image/audio sensing, processing (ML), decision making (ML/AI), robot face animation (TODO)
- Cloud software infrastructure (TODO)
- robot skills store (TODO)
- Maker's Pet Loki 200mm 3D-printed pet robot
- Maker's Pet Fido 250mm 3D-printed pet robot
- Maker's Pet Snoopy 300mm 3D-printed pet robot
- Maker's Pet Mini 125mm 3D-printed educational robot
- Add your own version to the list
- YDLIDAR X4 (default), X2/X2L, X3, X3PRO, SCL
- Neato XV11
- Xiaomi Roborock 1st gen LDS02RR (~$16 off AliExpress including shipping)
- SLAMTEC RPLIDAR A1
- 3irobotix Delta-2A, Delta-2B, Delta-2G
- LDROBOT LD14P
- Camsense X1
The entire up-to-date list of supported LiDAR is here.
There are two ways to install ROS2 and Kaia.ai on your PC - using Docker (Windows, Linux) or Ubuntu 24.04 (physical PC of virtual machine). Follow the installation instructions here.
Watch build, setup and bringup videos
- Note: these videos are outdated; updated videos will be published shortly for Maker's Pet Mini
The Kaia.ai Docker image contains ROS2 and micro-ROS pre-configured with additional Kaia.ai ROS2 packages.
Open a Windows command shell or Windows PowerShell window and type the command below. This should give you a bash prompt.
docker pull kaiaai/kaiaai:humble
docker run --name makerspet -it --rm -p 8888:8888/udp -p 4430:4430/tcp -e DISPLAY=host.docker.internal:0.0 -e LIBGL_ALWAYS_INDIRECT=0 kaiaai/kaiaai:humble
Get an aditional bash prompt by opening another Windows command shell or Windows PowerShell window and typing:
docker exec -it makerspet bash
If you installed ROS2/Kaia.ai without Docker directly on a Ubuntu PC/VM, just boot your Ubuntu PC to a bash prompt.
makerspet_mini is the default robot model. When using another robot model, change the default model using Kaia.ai CLI.
For example, the command below sets the default robot model to makerspet_loki.
kaia config robot.model makerspet_loki
# Launch the physical robot
ros2 launch kaiaai_bringup physical.launch.py
# Drive robot manually
ros2 run kaiaai_teleop teleop_keyboard
# Monitor robot sensors
ros2 launch kaiaai_bringup monitor_robot.launch.py
# Create a map while driving manually
ros2 launch kaiaai_bringup cartographer.launch.py
# Save the newly-created map
ros2 run nav2_map_server map_saver_cli -f ~/map --ros-args -p save_map_timeout:=60.0
# Launch the physical robot
ros2 launch kaiaai_bringup physical.launch.py
# Specify target location;; robot self-drives using an existing map
ros2 launch kaiaai_bringup navigation.launch.py map:=$HOME/map
# Launch SLAM (simultaneous localization and mapping) - navigate and map simultaneously
ros2 launch kaiaai_bringup navigation.launch.py slam:=True
# Robot automatically seeks out, self-drives to unknown locations
ros2 launch explore_lite explore.launch.py
# Save the newly-created map
ros2 run nav2_map_server map_saver_cli -f ~/map --ros-args -p save_map_timeout:=60.0
# View parameters
ros2 node list
ros2 node info /pet
ros2 param list /pet
ros2 param dump /pet
# Get the current laser scan frequency
ros2 param get /pet lidar.scan.freq.now
# Set the desired laser scan frequency to 7 Hz
ros2 param set /pet lidar.scan.freq.target 7.0
# Get the current desired laser scan frequency
ros2 param get /pet lidar.scan.freq.target
# Reset the desired laser scan frequency to default
ros2 param set /pet lidar.scan.freq.target 0.0
# List available topics
ros2 topic list
# Get WiFi strength
ros2 topic echo /wifi_state --once
# View raw telemetry
ros2 topic echo /telemetry --once
# Get LiDAR scan data
ros2 topic echo /scan --once
# View current odometer value
ros2 topic echo /odom --once
# View current wheel rotation angles
ros2 topic echo /joint_states --once
# View current battery voltage, charge percentage
ros2 topic echo /battery_state --once
# View target velocity sent by kaiaai_telem or navigation
ros2 topic echo /cmd_vel --once
# Launch the robot in a simulation - drive manually
ros2 launch kaiaai_gazebo world.launch.py
ros2 run kaiaai_teleop teleop_keyboard
ros2 launch kaiaai_bringup monitor_robot.launch.py
# Launch the robot in a simulation - robot self-drives around
ros2 launch kaiaai_gazebo world.launch.py
ros2 launch kaiaai_gazebo self_drive_gazebo.launch.py
ros2 launch kaiaai_bringup monitor_robot.launch.py
# Launch the robot in a simulation - let it navigate automatically using an existing map
ros2 launch kaiaai_gazebo world.launch.py
ros2 launch kaiaai_bringup navigation.launch.py use_sim_time:=true \
map:=/ros_ws/src/kaiaai_gazebo/map/living_room.yaml
# Launch the robot in a simulation - navigate and create a map simultaneously; save the map
ros2 launch kaiaai_gazebo world.launch.py robot_model:=makerspet_loki
ros2 launch kaiaai_bringup navigation.launch.py use_sim_time:=true slam:=True
ros2 run nav2_map_server map_saver_cli -f ~/map --ros-args -p save_map_timeout:=60.0
# Launch the robot in a simulation - create, save a map; robot self-drives around trivially
ros2 launch kaiaai_gazebo world.launch.py
ros2 launch kaiaai_bringup cartographer.launch.py use_sim_time:=true
ros2 launch kaiaai_gazebo self_drive_gazebo.launch.py
ros2 run nav2_map_server map_saver_cli -f ~/living_room_map --ros-args -p save_map_timeout:=60.0
# Launch the robot in a simulation - navigate and create a map simultaneously
# Robot seeks out, self-drives to unknown locations to complete the mapping
ros2 launch kaiaai_gazebo world.launch.py
ros2 launch kaiaai_bringup navigation.launch.py use_sim_time:=true slam:=True
ros2 launch explore_lite explore.launch.py
ros2 run nav2_map_server map_saver_cli -f ~/map --ros-args -p save_map_timeout:=60.0
# Launch the robot in a simulation - navigate and create a map simultaneously
# Robot seeks out, self-drives to unknown locations to complete the mapping, saves map
ros2 run auto_mapper auto_mapper map_path:=/ros_ws/map.yaml
ros2 launch kaiaai_gazebo world.launch.py
ros2 launch kaiaai_bringup navigation.launch.py use_sim_time:=true slam:=True
# Launch the robot in a simulation - navigate and create a map simultaneously
# Robot seeks out, self-drives to unknown locations to complete the mapping
ros2 launch kaiaai_gazebo world.launch.py
ros2 launch kaiaai_bringup navigation.launch.py use_sim_time:=true slam:=True
ros2 run nav2_wfd explore
You can set lidar_model to any of these supported models: XIAOMI-LDS02RR, YDLIDAR-X4, YDLIDAR-X3,
YDLIDAR-X3-PRO, YDLIDAR-X2-X2L, YDLIDAR-SCL, NEATO-XV11, 3IROBOTIX-DELTA-2A, 3IROBOTIX-DELTA-2B,
3IROBOTIX-DELTA-2G, LDROBOT-LD14P, CAMSENSE-X1, SLAMTEC-RPLIDAR-A1.
# Inspect, edit robot's URDF model
ros2 launch kaiaai_bringup inspect_urdf.launch.py
ros2 launch kaiaai_bringup edit_urdf.launch.py
# Convert URDF robot model file into SDF Gazebo simulation model file
ros2 run kaiaai_gazebo urdf2sdf.sh /ros_ws/src/makerspet_loki
cd /ros_ws && colcon build --symlink-install --packages-select makerspet_loki
Commands physical.launch.py, teleop_keyboard, monitor_robot.launch.py, cartographer.launch.py,
navigation.launch.py, inspect_urdf.launch.py, edit_urdf.launch.py and world.launch.py accept
an optional robot_model argument to override the default robot package setting, per launch.
The physical.launch.py also accepts an optional lidar_model argument to override the default
LiDAR model choice.
For example, the command below will use the makerspet_loki robot model and YDLIDAR-X3 LiDAR instead of the defaults.
ros2 launch kaiaai_bringup physical.launch.py robot_model:=makerspet_loki lidar_model:=YDLIDAR-X3
Initial versions of packages in this repo are based on ROBOTIS Turtlebot3 code
v0.10.0 in debug
- converted kaiaa from metapackage to Pyhon package
- kaia CLI sets default robot model
- added YDLIDAR SCL
- added intensity telemetry publication
- added Maker's Pet Mini
- added LiDAR orientation_deg
v0.9.0
- added WebRTC, web server, OpenCV launch
- added Camsense X1 LiDAR
v0.8.0
- added kaiaai_python ROS2 package
- added 3irobotix Delta-2A 230400 baud (vs 115200)
- added 3irobotix Delta-2B
- bugfix missing first scan point
v0.7.0
- kaiaai_telemetry
- switched to KaiaTelemetry2 message
- publish /battery_state
- publish /wifi_state RSSI
- discard_broken_scans, needs more debug
- added LDROBOT LD14P
v0.6.0 2/11/2024
- added 3irobotix Delta-2A, Delta-2G
2/5/2024
- added LiDAR laser distance scan sensors support
- YDLIDAR X3, X3-PRO
- Neato XV11
- RPLIDAR A1
- split kaiaai_telemetry config into default and custom
kaiaai/kaiaai_telemetry/config/telem.yamlis the default configmakerspet_loki/config/telem.yamlis the custom config for the Loki robot modelmakerspet_fido/config/telem.yamlis the custom config for the Fido robot modelmakerspet_snoopy/config/telem.yamlis the custom config for the Snoopy robot model
1/28/2024
- added YDLIDAR X2 support to kaiaai_telemetry
1/21/2024
- kaiaai_telemetry now supports multiple LiDAR laser distance scan sensors
- added Xiaomi Mi LDS02RR; default is YDLIDAR X4
12/11/2024
- added fully automatic self-driving to map exploration (frontier exploration m-explore)
- I think this code needs debug
12/7/2024
- added Nav2 SLAM (as an alternative to Google Cartographer)
- Nav2 SLAM enables driving autonomously to a set goal while mapping (and while the map is still incomplete)
- increase the map saver default timeout
- fixed the map saving command timing out on my "slow" laptop