TidyBot++: An Open-Source Holonomic Mobile Manipulator for Robot Learning
Jimmy Wu, William Chong, Robert Holmberg, Aaditya Prasad, Yihuai Gao, Oussama Khatib, Shuran Song, Szymon Rusinkiewicz, Jeannette Bohg
2024-12-17
Summary
This paper discusses TidyBot++, an open-source mobile robot designed to help with household tasks by learning from human demonstrations, making it easier for people to collect data and improve robot performance.
What's the problem?
Creating robots that can effectively perform household tasks is challenging because they need to learn from many examples of how humans do things. Traditional robots often have limited movement capabilities and can be expensive, making it hard for researchers to gather the data needed for effective learning. Additionally, existing robots may struggle with complex tasks that require precise movements.
What's the solution?
TidyBot++ addresses these challenges by using a flexible design that allows it to move in any direction easily, thanks to its holonomic base. This makes the robot more maneuverable and capable of performing a variety of tasks without getting stuck or moving inefficiently. The robot is equipped with a user-friendly mobile phone interface that lets users control it easily and collect data for teaching the robot how to complete tasks. This way, researchers can gather the necessary information to train the robot effectively without needing specialized equipment.
Why it matters?
This research is important because it provides an affordable and accessible way for researchers and developers to create robots that can assist with everyday tasks. By making it easier to collect data and improve robot learning, TidyBot++ could lead to better home assistance robots that can adapt to individual preferences and help people in their daily lives.
Abstract
Exploiting the promise of recent advances in imitation learning for mobile manipulation will require the collection of large numbers of human-guided demonstrations. This paper proposes an open-source design for an inexpensive, robust, and flexible mobile manipulator that can support arbitrary arms, enabling a wide range of real-world household mobile manipulation tasks. Crucially, our design uses powered casters to enable the mobile base to be fully holonomic, able to control all planar degrees of freedom independently and simultaneously. This feature makes the base more maneuverable and simplifies many mobile manipulation tasks, eliminating the kinematic constraints that create complex and time-consuming motions in nonholonomic bases. We equip our robot with an intuitive mobile phone teleoperation interface to enable easy data acquisition for imitation learning. In our experiments, we use this interface to collect data and show that the resulting learned policies can successfully perform a variety of common household mobile manipulation tasks.