San Diego-based robotics company Robolink just took the top honors for Best Innovation Award in the Robotics and Drone category at this year’s CES for their Zümi AI robotic car platform, which aims to teach kids fundamental concepts on AI and self-driving vehicle technology. The Zümi robotic cars are capable of identifying objects, recognizing hand gestures, and learning how to navigate the optimal route to a destination on an interactive track.
The Zümi robotic car sports all the hardware needed for an AI vehicle, including a Raspberry Pi Zero, Pi Camera, a pair of DC motors, a 128 x 64 OLED display, and a host of sensors — a gyroscope, an accelerometer, and six IR sensors. An interactive track is used with the robotic car and can be populated with numerous objects, such as people, trees, buildings, etc., that can be arranged any number of ways that the vehicle learns to maneuver through.
Robolink outfitted the Zümi platform with ‘industry-relevant’ tools utilized in most self-driving vehicles in development — like the open source TensorFlow machine learning framework and OpenCV computer vision programming library, which is used to train Zümi to drive autonomously around the test track.
Users can program Zümi via Robolink’s drag-and-drop web-based IDE or Python, and more advanced developers can write their own code using TensorFlow or Keras for self-driving vehicle experiments. The AI platform is also equipped with interactive tutorials to help those new to programming, robotics, and AI.
“Zümi is the first educational self-driving car kit that will teach you about the concept of artificial intelligence and self-driving car technology in a fun and engaging way. With our friendly step-by-step tutorials, you will be able to train her on how to navigate through a miniature map in less than an hour. The more she learns, the better she’ll get.”
Robolink is currently crowdfunding Zümi on Kickstarter with pledges starting at $99, which is a steal, as the platform will retail at $180 when it hits the market.