Hackaday

UNO Q EchoGlow Workshop - Part 2

The second part of the UNO Q EchoGlow workshop walks participants through preparing the hardware and deploying a keyword‑spotting AI model to turn the board into a smart lamp. It begins with updating Arduino boards, installing the pre‑packaged EchoGlow brick in App Lab, and cloning the Qualcomm Arduino EdgeWS GitHub repository before running a series of Linux commands on the device. Key steps include verifying the seven‑out‑of‑seven setup checks, performing a pseudo‑reboot, and selecting the Arduino Uno Q Qualcomm chip as the deployment target. After building the model, the workflow generates an EIM file, which must be saved and marked “in use” before flashing the board. The presenter emphasizes safety – warning that an untested model can be discarded – and highlights practical details such as copying example sketches, naming conventions, and the back‑and‑forth authentication between Edge Impulse and Arduino App Lab. Hardware assembly is covered in depth, with three custom PCBs (microphone, LEDs, shield) connected via ribbon and flat cables, and tips for correctly orienting connectors. Successfully completing these steps enables developers to prototype an AI‑enabled EchoGlow lamp, demonstrating a full end‑to‑end pipeline from model training to physical product integration, and illustrating best practices for edge‑AI deployment on Arduino’s Qualcomm‑based platform.