Deepnote
IntegrationsPricingFor TeamsFor EducationJoin usDocs
Sign in
Robotics / United States University and Toyota logo

Deepnote Supports World-Class Robotics Education

Dr. Russ Tedrake, renowned professor and vice president of Robotics Research at Toyota, uses Deepnote for teaching robotics. To say that his students' projects are inspirational would be an understatement. Have a look for yourself at some of their achievements. Such projects fill the imagination with wonder and anticipation for the future of robotics.

Image sources: Playing Piano with a Robotic Hand by Seong Ho Yeon & Exploring Physics-Based Approaches to Robotic Table Tennis by Dylan Zhou and Chaitanya Ravuri

Use cases

Optimization, simulation, interactive data visualization, reinforcement learning

Data stack

Drake, MeshCat, Nevergrad, OpenAI Gym, Gradescope

United States University and Toyota about 1
United States University and Toyota about 2
01

Benefits of peer programming

Prior to using Deepnote to teach both Robotic Manipulation and Underactuated Robotics, it was difficult to debug student code since interfaces were not collaborative. With Deepnote, Dr. Tedrake now has the ability to peer program with students in real time. This has not only made debugging more efficient but it has also provided new learning opportunities as students benefit from direct collaboration with Dr. Tedrake.

02

Computational documentation

Publically available tutorials for Dr. Tedrake's industry-standard robotics package, Drake, are also hosted on Deepnote. Since Deepnote allows provisioning via Docker, the tutorials become more stable and maintainable when compared to using other notebook platforms.

03

Integration with domain-specific tools

Dr. Tedrake relies on a 3D, remotely-controllable, viewer called MeshCat, which is built on top of three.js. Since Deepnote allows incoming connections, MeshCat visualizations can be served over HTTPS without needing any additional infrastructure in place. Students can simply open up their browsers and interact collaboratively with their visualizations.

Interact with MeshCat now

Click open controls on the image below, then, hit play to watch MeshCat in action!

Increasing quality of student projects

Deepnote was a huge success for the Robotic Manipulation course in the fall. Student projects have been getting better each year but this year the improvement was dramatic.

Dr. Russ TedrakeProfessor and VP of Robotics Research at Toyota
Dr. Russ Tedrake's avatar

A platform for teaching and computing

Being able to peer program, interact with students' visualizations, and provision via Docker makes Deepnote the most effective notebook platform for my teaching.

Dr. Russ TedrakeProfessor and VP of Robotics Research at Toyota
Dr. Russ Tedrake's avatar

Support from the team

When I needed tunneling support and practical ideas for setting up my courses, I was able to work directly with Daniel Zvara, a software engineer at Deepnote. This kind of close collaboration is a key part of delivering effective course material to my students.

Dr. Russ TedrakeProfessor and VP of Robotics Research at Toyota
Dr. Russ Tedrake's avatar

Outcome

Allowing students to connect to a customized and fully collaborative compute environment, as well as integrate with domain-specific tools such as MeshCat, serves to unlocks new learning opportunities. Deepnote is proud to play a role in robotics education and research. Dr. Tedrake and his students continue to inspire us with their innovative applications of robotics.

Learn more

Check out more amazing projects from Russ' team here. If you would like to learn more about the Drake robotics package, please read this excellent Medium post.

For more information on how Deepnote supports teaching, please see our recent workshop on Collaborative Data Science Education.

Outcome for United States University and Toyota

Get started with Deepnote today

Deepnote is a new kind of data science notebook. Jupyter-compatible with real-time collaboration and running in the cloud.

Deepnote
Use cases