OpenDuck Mini V2: Smart Wheeled Platform

OpenDuck Mini V2 is a compact, open-source differential drive wheeled robot designed for autonomous navigation, SLAM mapping, and digital fabrication training. Engineered as a robust testbed, it coordinates low-level ESP32 motor loops with high-level ROS2 SLAM stack on a Raspberry Pi.

What's New in Mini V2

micro-ROS Control

Integrated micro-ROS agent coordinates actuator velocity directly over WiFi or serial.

LIDAR Navigation

A 360-degree LIDAR scanner sits on the top deck for real-time laser mapping.

3D Printed Chassis

Fully open-source STL files make repairing or modifying the robot simple.

Hardware & Specification

The system utilizes standard N20 encoder gearmotors, an ESP32 microcontroller for real-time motor PID speed control, and a Raspberry Pi host processor for mapping and logic.

Component	OpenDuck Mini V2 Specs	Connection Interface
Low-level controller	ESP32 DevKitC v4 (Dual-Core)	USB-Serial / UART
High-level controller	Raspberry Pi 4 (4GB) or Raspberry Pi 5	Host PC/WLAN
Motors	2x N20 Gearmotors with Hall-effect Encoders	Dual H-Bridge Driver (TB6612FNG)
LIDAR Module	RPLIDAR A1M8 360-degree Scanner	USB to UART module
IMU Sensor	MPU6050 (6-DOF Accelerometer + Gyro)	I2C Interface
Power Source	7.4V 2S Li-Po or 18650 Battery Pack	DC-DC buck converters (5V and 3.3V)

Important: Disconnect the 7.4V battery pack before linking the ESP32 to your PC via USB to prevent back-powering and damage to your USB port.

micro-ROS Firmware Example

The following C++ code fragment runs on the ESP32 low-level microcontroller, establishing velocity subscription topics using micro-ROS to control motors:

Arduino C++ (ESP32)

#include 
#include 
#include 
#include 

rcl_subscription_t subscriber;
geometry_msgs__msg__Twist msg;

void subscription_callback(const void *msgin) {
    const geometry_msgs__msg__Twist * twist_msg = (const geometry_msgs__msg__Twist *)msgin;
    
    // Command motor speed
    float linear_x = twist_msg->linear.x;
    float angular_z = twist_msg->angular.z;
    
    // Convert twist values to differential wheel velocities
    float right_wheel_target = linear_x + angular_z;
    float left_wheel_target = linear_x - angular_z;
    
    // Drive motors (PID velocity controller updates here)
    set_motor_speeds(left_wheel_target, right_wheel_target);
}

To run SLAM on the Raspberry Pi host, connect to the ESP32 micro-ROS agent, start the LIDAR publisher, and launch Cartographer or Nav2:

Bash

# Start micro-ROS agent on Raspberry Pi
ros2 run micro_ros_agent micro_ros_agent serial --dev /dev/ttyUSB0

# Start RPLIDAR node
ros2 launch rplidar_ros2 rplidar_launch.py

# Launch Cartographer SLAM Node
ros2 launch openduck_nav cartographer_slam.launch.py

Troubleshooting (FAQ)

Odometry drift usually originates from wheel slippage or uneven surfaces. Calibrate the MPU6050 IMU and implement a Kalman Filter (using the robot_localization package) to combine encoder counts and inertial data.

Verify the baud rate settings match between the ESP32 firmware and the micro-ROS serial node launch command (default is usually 115200 or 921600).

Back to Projects