ROS 2 Social Locobot Development

This repository is a base for tests on the LoCoBot, both on the real robot and the simulation. The main repository of LoCoBot can be cloned from the repository Interbotix ROS Rovers.

Prerequisites

You need to have the following installed on your machine:

• ROS 2 Humble
• Alternative: Docker Desktop for development using docker container

Folder structure

├── 3rdparty
├── docker
├── nav2_soloco_controller
├── nav2_soloco_costmap_plugin
├── soloco_interfaces           
├── soloco_launch               
├── soloco_perception
├── LICENSE
├── README.md
├── .gitignore
└── .gitmodules

Installation

For first time installation, please refer to the installation guide.

1. Pulling new changes locally

   git pull --recurse-submodules
   git submodule update --init --recursive
   

2. Applying new changes inside docker container

   # Run container
   sudo docker run --runtime nvidia -it --rm --network host --privileged -v /dev:/dev -v /home/zedbox/ros2_ws:/home/zedbox/ros2_ws stephenadhi/ros2:humble-l4t-r35.2-zedsdk-4.0
   # Pull new changes
   cd home/zedbox/ros2_ws/src/human_ws
   git pull
   cd home/zedbox/ros2_ws
   colcon build --packages-select soloco_perception nav2_soloco_controller soloco_launch soloco_interfaces
   

Custom Launch Examples

We provide general launch files for perception and navigation. Our modified version of the LoCoBot have two computing device onboard: Intel NUC and ZED Box (Jetson Xavier NX).

First, SSH to the Intel NUC computer

ssh locobot@locobot

Optional: choose unique domain ID to prevent interference from other running ROS 2 system

export ROS_DOMAIN_ID=2 # Adjust to your settings, default=0

Go to the directory and source installation

cd ~/interbotix_ws
source install/setup.bash && source /opt/ros/humble/setup.bash

Default LoCoBot launch with navigation2 stack + SLAM, and visualization:

ros2 launch soloco_launch intelnuc_locobot_bringup.launch.py nav2_param_filename:=smac_mppi_nav2_params.yaml

LoCoBot launch with custom neural motion planner as controller:

ros2 launch soloco_launch intelnuc_locobot_bringup.launch.py use_nav2_slam:=True nav2_param_filename:=smac_soloco_nav2_params.yaml use_soloco_controller:=True

For launching ZED human perception and multi-tracking module:

SSH to the ZED Box

ssh [email protected]

Run the shell script to bringup our ROS modules inside a docker container

./docker_zedbox_bringup.sh

Visualization in remote PC:

Source installation and set the same ROS domain ID.

export ROS_DOMAIN_ID=2 # Adjust to your settings, default=0
source install/setup.bash

Launch RViZ:

ros2 launch soloco_launch remote_view.launch.py

Simulation Example with Pedsim

Launch Locobot in Gazebo, optionally with GUI, pedestrian simulator, and RViZ. Odometry is based on ground truth with gaussian noise=0.001. The default launch is the following command:

ros2 launch soloco_launch locobot_sim.launch.py use_gazebo_gui:=false use_pedsim:=true use_soloco_controller:=true launch_remote_view:=true

Rosbag Example

To record only robot data, run the following command ros2 bag record /locobot/commands/velocity /locobot/odom /map /local_costmap/costmap /global_costmap/costmap /tf /tf_static /plan /visualization/predicted_future /human/interpolated_history /robot/ego_trajectory /goal_pose /locobot/robot_description /zed2/zed_node/left_raw/image_raw_color

To record data for visualization purposes, run the following command

ros2 bag record /locobot/commands/velocity /locobot/odom /map /local_costmap/costmap /global_costmap/costmap /tf /tf_static /plan /visualization/predicted_future /visualization/human_tracks /visualization/robot_track /visualization/subgoal /visualization/human_bounding_boxes /goal_pose /locobot/robot_description /zed2/zed_node/left_raw/image_raw_color

To record simulation data, run the following command ros2 bag record /locobot/diffdrive_controller/cmd_vel_unstamped /locobot/odom /map /local_costmap/costmap /global_costmap/costmap /tf /tf_static /plan /human/simulated_agents /human/predicted_future /human/interpolated_history /robot/ego_trajectory /goal_pose /locobot/robot_description