ROS 2 Pick and Place System with Edge Impulse
Originally published on https://www.hackster.io by Naveen on 8 November 20924
This project focuses on creating an advanced robotic system for efficient object sorting and placement in dynamic environments using computer vision and robotics. A robotic arm integrated with a 3D depth camera performs pick-and-place tasks with precision. The system leverages computer vision to identify and locate objects in a cluttered environment, plan grasping and placing strategies, and execute actions seamlessly. Key hardware includes the Arduino Braccio++, Luxonis OAK-D for depth imaging, and Raspberry Pi 5 as the central controller running ROS 2 and associated tools for motion planning and object recognition. Therefore, ROS 2 is showcased sorting colorful plastic toys, such as penguins and pigs, making it an engaging demonstration of robotics in action.
The Plan
The development involved multiple stages, including setting up the Raspberry Pi 5 with ROS 2, installing software like MoveIt 2 for robotic motion planning, and training a YOLOv5 object detection model using Edge Impulse Studio. Data collection utilized the OAK-D camera to capture and label images of objects, followed by training and testing the model with high accuracy. The model was optimized and deployed to the OAK-D device using OpenVINO and DepthAI, enabling real-time object detection. Integration with the robotic arm required custom firmware and precise configuration of the URDF model and motion planning settings, ensuring smooth manipulation.
To operationalize the system, micro-ROS enabled seamless communication between the Arduino Nano RP2040 microcontroller and ROS 2, allowing the robotic arm to execute commands. DepthAI ROS nodes facilitated object detection and localization, publishing results to guide the robotic arm’s movements. The setup effectively combines robotics, computer vision, and embedded systems to perform flexible and accurate sorting tasks. This modular system can be adapted for various industrial or domestic applications, demonstrating the potential of robotics in automation.
Read more: ROS 2 Pick and Place System with Edge Impulse
Disclaimer: The content in this post includes excerpts from other blogs and websites. Full credit for the original work goes to the respective authors and publications. This content is shared here for informational purposes only, to give our customers direct insight on where to purchase the mentioned items, and all rights remain with the original creators.
