Face Expression Recognition Robot with Arduino UNO Q

Learn to build a face expression recognition robot using the Arduino UNO Q's hybrid architecture that combines a Qualcomm QRB2210 application processor with an STM32U585 microcontroller. Explore how this unique board enables both AI/ML inference and real-time control in a single platform, making it ideal for headless edge-AI applications including computer vision, local web servers, ONNX/TensorFlow Lite inference, and USB-camera processing. Discover the mechanical design of a pan/tilt system with USB webcam and NeoPixel LED ring that automatically tracks faces and responds to detected emotions. Master the process of retraining a YOLOv11-nano model using transfer learning, starting from pretrained COCO weights and fine-tuning on a labeled facial-expression dataset from Roboflow. Understand how to export the model to ONNX format, optimize it, and deploy it directly onto the UNO Q for on-device processing without requiring cloud services or external GPU. Examine the Python code that handles image capture, preprocessing, and inference on the Linux side, while learning how the microcontroller manages servo control using PID loops for smooth motion tracking. See how Arduino's Bridge RPC system enables communication between the CPU handling computer vision tasks and the microcontroller managing real-time movement control. Witness the complete application in action as the robot detects faces, tracks the largest bounding box, and changes LED colors or patterns based on predicted emotions like "happy," "surprised," or "neutral." Gain insights into expanding this project for gesture-reactive art, smart home sensing, human-robot interaction, or mobile robotics applications using the UNO Q's powerful edge-AI capabilities.