bipedal_wheel

Introduction

This software contains simulation and basic control framework of bipedal wheel robot. In addition, navigation and trajectory tracker were added to the underlying control system.

The specific implementation method is as follows:

1. Calculating leg kinematics by VMC, Implement balance control by LQR and adding PID to achieve comprehensive motion.
2. Location is achieved using ICP and EKF, and navigation is achieved using ROS navigation.
3. Implement A-star search, spline trajectory optimization, and NMPC trajectory tracker using OCS2.

Installation

Source code

# Clone repository bipedal_wheel 
git clone https://github.com/1moule/bipedal_wheel

OCS2

1. Clone OCS2 into your workspace

   # Clone OCS2
   git clone git@github.com:leggedrobotics/ocs2.git
   

2. Compile these packages with catkin tools

   catkin config -DCMAKE_BUILD_TYPE=RelWithDebInfo
   catkin build ocs2_ros_interfaces ocs2_robotic_tools ocs2_sqp
   

build

Build the source code by catkin tools

catkin build bipedal_wheel_* velodyne_*

Quick Start

• Run basic control in empty world

  1. Run the simulation:

     roslaunch bipedal_wheel_gazebo empty_world.launch
     

  2. Load the controller

     roslaunch bipedal_wheel_controller load_controller.launch
     

  3. Start controller by rqt_controller_managerGui:

     sudo apt install ros-noetic-rqt-controller-manager
     rosrun rqt_controller_manager rqt_controller_manager
     

  4. Then you can control the robot by topic /cmd_vel

• Run navigation

  1. Run the simulation:

     roslaunch bipedal_wheel_gazebo test_world.launch
     

  2. Load and start controller as the previous steps and DO NOT publish topic /cmd_vel

  3. Run navigation

     roslaunch bipedal_wheel_navigation navigation.launch
     

Reference

[1] https://github.com/qiayuanl/legged_control

[2] https://github.com/SCAU-RM-NAV/rm2023_auto_sentry_ws

[3] https://github.com/MarineRock10/ST-opt-tools