Stanford Students Build Adorable AI-Powered Robot Dogs from Scratch in Intro Robotics Course

Students in CS 123: A Hands-On Introduction to Building AI-Enabled Robots at Stanford University are diving into the world of robotics and AI by constructing and enhancing their own quadruped robots, affectionately named "Pupper." Now in its third year, the course has evolved from an independent study project led by Stanford's robotics club and is taught by professors Karen Liu, Jie Tan from Google DeepMind, and Stuart Bowers from Apple and Hands-On Robotics. The course combines foundational lessons in motor control, movement, and hardware construction with advanced AI concepts, allowing students to build a fully functional robot from a starter hardware kit. Over 10 weeks, participants learn how to program Pupper to walk, navigate, respond to human commands, and perform specialized tasks, culminating in a final presentation known as the "Dog and Pony Show." Instructor Karen Liu emphasizes that building a robot from scratch is the best way to inspire and educate students. The use of Pupper as a quadruped design provides an ideal platform for beginners due to its combination of accessibility and advanced capabilities. "We believe that the best way to help and inspire students to become robotics experts is to have them build a robot from scratch," Liu explained. "It's a powerful introductory platform that can support cutting-edge AI algorithms." The course starts by covering the basics of robotics, such as how motors function and how robots can move. In the subsequent phase, students integrate AI techniques, using neural networks to enhance the robot's walking, vision, and interaction with the environment. Final projects often involve creative tasks, such as navigating mazes or performing mock firefighting exercises, with guests from leading tech companies like NVIDIA and Google offering feedback. Pupper's design, originally based on "Doggo" created by the Stanford Student Robotics club, has been continually refined to make it more user-friendly and powerful. The club's initial aim was to make robotics approachable and fun, and this philosophy is deeply embedded in CS 123. Early adopters of the Pupper course, like Ankush Kundan Dhawan, who became a head teaching assistant, attest to the instructors' passion and dedication. The course's hands-on nature and low barrier to entry—requiring only basic programming skills—are key to its success. Weekly lectures and labs, with playful titles like "Wiggle Your Big Toe" and "Do What I Say," ensure that students are engaged while learning practical skills. By the end of the quarter, students combine their newfound knowledge of locomotion, computer vision, and language to develop sophisticated AI-driven behaviors in Pupper. Instructors Liu, Tan, and Bowers are committed to keeping the content current and relevant, integrating new lessons and technological advancements each quarter. The course has garnered immense popularity among students, and the instructors anticipate continued growth in interest in both robotics and AI. They see CS 123 as a stepping stone for students to become future innovators and leaders in the field. The rise of Pupper and CS 123 reflects a broader trend in the expansion of AI and robotics education. By equipping students with the tools and knowledge to tackle complex challenges, the course aims to prepare them for the burgeoning robotics industry. The instructors hope to expand the course and foster a community that extends beyond Stanford, contributing to the advancement of AI robotics on a global scale. Industry insiders laud the course for its innovative approach and practical focus, noting that it bridges the gap between theoretical knowledge and real-world applications. Companies like Google and NVIDIA, which participate in the final presentations, recognize the value of well-rounded robotics and AI education and see CS 123 as a critical component in cultivating the next generation of roboticists. Karen Liu's background as a professor of computer science at Stanford adds depth to the course's academic rigor, while Jie Tan's expertise from Google DeepMind and Stuart Bowers' experience from Apple bring a blend of cutting-edge research and industry insights. Together, they have created a unique learning environment that empowers students to explore and innovate in the rapidly evolving field of AI robotics.