Skip to content

Baey/bernard_hw_if

BranchesTags

Repository files navigation

BERNARD ROS2 Package Test Build node for arm64

BERNARD Hardware Interface (bernard_hw_if)

Overview

This ROS 2 package provides the hardware interface for the BERNARD bipedal robot. It handles real-time communication with the MD80 servo actuators via the CANdle CAN interface and publishes sensor data (joint states, IMU, foot pressure) received from the STM32 microcontroller. This package is part of the larger BERNARD Bipedal Robot Project.

Key Features:

  • CAN Communication: Real-time control of 6 MD80 servo actuators via CANdle interface
  • Robot Control Modes: OFF, MANUAL (joystick), RL_POLICY (reinforcement learning), and HOLD_POSITION
  • Sensor Integration: Publishes IMU data, joint states, and foot pressure sensor readings from STM32 microcontroller
  • JointState Interface: Standard ROS 2 JointState messages for compatibility with control stacks
  • Robust Error Handling: Hardware fault detection and graceful error recovery

Project Structure:

This repository is part of a larger ecosystem:

System Architecture

┌─────────────────────┐
│   RL Policy /       ├───────────────┐
│ Joystick Input      │               │
└──────────▲──────────┘               │
           │                          │
      ┌────▼──────────────────────┐   │
      │   bernard_hw_if (ROS 2)   │   │
      │  - MD80 actuator control  │   │
      │  - Sensor data fusion     │   │
      └───▲────────────▲──────────┘   │
          │ CAN Bus    │ Serial/TCP   │
      ┌───▼────┐  ┌────┴──────────┐   │
      │ MD80   │  │   STM32 uC    │   │
      │ Drives │  │ - IMU (BNO055)├───┘
      └────────┘  │ - Foot sensors│
                  │ - micro-ROS   │
                  └───────────────┘

Requirements

Hardware

  • STM32 Microcontroller (e.g., Nucleo L476RG or similar)

    • BNO055 IMU for orientation/acceleration data
    • Analog pressure sensors for foot contact detection
    • TFT display for GUI (optional)

  • MD80 Servo Actuators (6x)

    • CAN interface compatible

  • CANdle CAN Interface for CAN communication

    • Supports real-time actuator control at 200 Hz+

Software

  • ROS 2 (Foxy or later)
  • C++20 or later
  • ament_cmake build system
  • rclcpp - ROS 2 C++ client library
  • CANdle SDK (included as submodule)

Installation

1. Clone the Repository

# Clone the main repository with all submodules
git clone --recurse-submodules https://github.com/Baey/bernard-bipedal-robot.git
cd bernard-bipedal-robot

2. Install Dependencies

On Ubuntu/Debian:

# Install ROS 2 (if not already installed)
# Follow: https://docs.ros.org/en/iron/Installation.html

# Source ROS 2 environment
source /opt/ros/$ROS_DISTRO/setup.bash

# Install build dependencies
sudo apt-get install -y \
    ros-$ROS_DISTRO-rclcpp \
    ros-$ROS_DISTRO-sensor-msgs \
    build-essential \
    cmake

# Install colcon build tool
sudo apt-get install -y python3-colcon-common-extensions

3. Build the Package

# Navigate to the workspace root
cd bernard-bipedal-robot

# Source ROS 2
source /opt/ros/$ROS_DISTRO/setup.bash

# Build the package using colcon
colcon build --packages-select bernard_hw_if

# Source the install space
source install/local_setup.bash

Usage

Running the Node

# Source the ROS 2 environment
source /opt/ros/$ROS_DISTRO/setup.bash
source install/local_setup.bash

# Launch the actuators node
ros2 run bernard_hw_if actuators_node

# Or with a launch file (if available)
ros2 launch bernard_hw_if actuators.launch.py

Control Modes

The hardware interface supports multiple control modes:

  1. OFF: All actuators disabled (safe state)
  2. MANUAL: Real-time joystick/gamepad control via joy input
  3. RL_POLICY: Receives commanded actions from RL policy running in simulation
  4. HOLD_POSITION: Maintains current joint positions (compliant control)

Published Topics

Topic Message Type Description
/joint_states sensor_msgs/JointState Current joint positions, velocities, and torques
/imu sensor_msgs/Imu IMU accelerations and angular velocities
/feet_pressure std_msgs/Float32MultiArray Foot contact pressure sensor readings
/status Custom msg (optional) Robot system status

Subscribed Topics

Topic Message Type Description
/joy sensor_msgs/Joy Joystick input for manual control
/actions std_msgs/Float32MultiArray Action commands from RL policy
/mode_command Custom msg (optional) Robot control mode selection

Configuration

Configuration parameters are defined in include/config.hpp:

// Number of actuators
constexpr uint8_t ACTUATORS_NUM = 6;

// Timing parameters
constexpr std::chrono::milliseconds ZEROING_BLINK_INTERVAL = std::chrono::milliseconds(1500);
constexpr std::chrono::milliseconds MANUAL_SELECTION_BLINK_INTERVAL = std::chrono::milliseconds(2500);

// CAN/actuator specific parameters

Modify these values to match your hardware configuration.

Hardware Interface Details

MD80 Actuators

The interface communicates with 6 MD80 servo actuators over CAN bus. Each actuator provides:

  • Joint position feedback
  • Velocity feedback
  • Torque feedback
  • Temperature monitoring

STM32 Microcontroller Integration

The STM32 node (from bernard-stm32-ros-node) communicates IMU and foot pressure data via:

  • Serial connection for real-time sensor data
  • micro-ROS bridge for ROS topic integration

Joystick Control

In MANUAL mode, joystick input is processed from Joy messages. Map your gamepad axes/buttons as needed in the code.

Building and Testing

Build the Package

colcon build --packages-select bernard_hw_if

Run Tests

# Run unit tests
colcon test --packages-select bernard_hw_if --executor sequential

# View test results
colcon test-result --verbose

Build Options

You can customize the build with CMake flags:

colcon build --packages-select bernard_hw_if \
    -DCMAKE_BUILD_TYPE=Release \
    -DBUILD_TESTING=ON

Troubleshooting

CAN Interface Not Found

Ensure the CANdle interface is properly connected and recognized by the system:

# List connected USB devices
lsusb

# Check for CAN interfaces
ip link show

Actuators Not Responding

  1. Verify CAN bus wiring and termination
  2. Check actuator power supply
  3. Run diagnostic command to test communication

Sensor Data Not Received

  1. Verify STM32 is running and connected
  2. Check serial port permissions: ls -l /dev/ttyUSB*
  3. Ensure micro-ROS agent is running

Related Repositories

Project Information

Author: Błażej Szargut
License: BSD-3-Clause
Affiliation: AGH University of Krakow, Master's Thesis Project

This hardware interface is part of a comprehensive bipedal robot platform combining:

  • Real-time hardware control via ROS 2
  • Advanced reinforcement learning in simulation (Isaac Lab)
  • Sensor fusion from multiple sources (IMU, joint feedback, foot pressure)

Contributing

To contribute improvements or bug fixes:

  1. Fork the repository
  2. Create a feature branch
  3. Make your changes
  4. Test thoroughly
  5. Submit a pull request

Please follow the existing code style and include appropriate tests for new features.

Acknowledgments

This project is developed as a Master's thesis at AGH University of Krakow. Special thanks to:

  • The Isaac Lab and IsaacSim teams for simulation framework
  • The ROS 2 community for middleware and tools
  • Contributors to the CANdle SDK

License

This project is licensed under the BSD-3-Clause License. See LICENSE file for details.

Quick Start Example

# Terminal 1: Start the hardware interface node
source /opt/ros/$ROS_DISTRO/setup.bash
source ~/ws/install/local_setup.bash
ros2 run bernard_hw_if actuators_node

# Terminal 2: Publish a manual control command
ros2 topic pub /actions std_msgs/msg/Float32MultiArray \
    '{data: [0.5, -0.3, 0.2, 0.1, -0.4, 0.25]}'

# Terminal 3: Monitor joint states
ros2 topic echo /joint_states

For more advanced usage and RL training, see the bernard-rl repository.

About

ROS 2 node for controlling MD80 actuators and publishing their states. Designed for use on embedded controllers (Kria/ARM64) as part of the BERNARD bipedal robot.

Topics

arm64 ros2 ros2-humble ros2-jazzy kria-kv260

Resources

Readme

License

BSD-3-Clause license
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages