Hiwonder PuppyPi ROS Quadruped Robot with Raspberry Pi, Integrated with Large AI Model (ChatGPT), Supports AI Vision, Voice Interaction, LiDAR, and Robotic Arm Attachment

The Hiwonder PuppyPi is an advanced, AI-powered quadruped robot dog built on the Raspberry Pi 5 platform and running ROS1 and ROS2. It is an exceptional tool for South African students, educators, and makers looking to explore Python programming, computer vision, and embodied AI. With its 8-degree-of-freedom coreless servos and integrated camera, this robot offers a highly capable platform for hands-on robotics development.

Advanced ROS-based quadruped platform

Powered by the Raspberry Pi 5, this quadruped robot runs on both ROS1 and ROS2 operating systems to support complex Python programming. Its multimodal AI model, camera, and voice interaction module enable the robot to understand its environment and perform agile tasks.

Eight degrees of freedom

The robot features eight degrees of freedom (DOF) powered by high-performance coreless servos. A specialised link structure design in its legs ensures precise, flexible, and dynamic motion control for complex actions.

Raspberry Pi and HD camera integration

Equipped with a 2DOF HD wide-angle camera, the control system leverages Raspberry Pi to run Python scripts and OpenCV image processing. This hardware combination provides a robust foundation for building diverse AI vision projects.

Real-time computer vision capabilities

Using the OpenCV vision library, the head-mounted HD camera detects and extracts the colour and position of target objects. It applies a PID control algorithm for real-time target locking, enabling functions like visual line following and autonomous ball kicking.

Precision target and object tracking

The robot detects and locates specific coloured objects in real-time using OpenCV algorithms. Guided by PID control, its head actively and precisely tracks moving targets within its field of view.

Autonomous stair climbing and navigation

The robot uses OpenCV to detect the spatial position and geometric contours of stairs. This visual data allows it to make autonomous decisions to steadily climb up and down stairs on its own.

Visual tag and code recognition

By processing tag codes with OpenCV, the robot identifies their position and orientation in real-time. Users can program customised interactive movements based on these coordinates.