Ever wanted to get things done with just waving your hands? In this episode, we make this possible by training Google’s Teachable Machine to recognize three basic poses. Let's build some basic tools such a hammer, a lift and a grip with LINKKI and operate them with your hand gestures.
Construction & Assembly
Step 1. Connect the Micro:bit to LINKKI. - Connect Micro:bit(1) to the computer. - Attach Micro:bit(2), (3), and (4) to their respective motors. Step 2. Create a pose model in Teachable Machine. - Set up the webcam to capture your full body. - Capture and upload pose samples in Google Teachable Machine. - Train the poses and copy the model link. Step 3. Link the Micro:bit to the pose model. - In the micro:bit of AI website, connect the Micro:bit and paste the model link. - When you perform the trained pose in front of the camera, Micro:bit(1) receives the signal and relays it to Micro(2), (3), and (4). Precautions - Use the micro:bit of AI website in Chrome or Edge browsers. - Every time you download new code to the Micro:bit, reconnect it to the web.
Further reading
What is Google Teachable Machine? Google’s Teachable Machine is a program that helps computers learn to recognize different types of information like images, sounds, or movements. By showing the computer various data samples, it can learn to identify and categorize similar items later. For example, if you show it multiple pictures of cats and dogs, it can eventually tell the difference between a cat and a dog on its own. How does Teachable Machine work? First, you collect data samples—like photos or videos of different poses—and input them into the computer. The computer then goes through a “training” phase, learning to distinguish between each pose or image. Once training is complete, the computer can recognize new images or movements and decide which category they belong to. This process results in what’s called a “model.” How do I use a trained model with Micro:bit+LINKKI Motor Kit? The completed model isn’t directly uploaded to Micro:bit. Instead, the micro:bit of AI website imports the Teachable Machine model. The website uses a camera to detect new poses and classifies them (e.g., Pose 1 or Pose 2). The classified results are then sent to the Micro:bit, which responds by performing specific actions (e.g., rotating a motor). LINK Micro:bit of AI Google Teachable Machine
