A team of researchers from Stanford and Princeton Universities have developed TidyBot++, an open source AI-based domestic robot capable of performing various tasks such as loading the washing machine, watering plants, cleaning surfaces, etc. As shown in the article about the project published on the arXiv website, the researchers fine-tuned the robot using imitative learning algorithms for robotics.
“TidyBot++ is a mobile robot designed to perform household chores with precision and ease,” notes the study’s first author Jimmy Wu, of Princeton University. “Compared to other platforms,” adds researcher Jeannette Bohg of Stanford University, ”TidyBot++ is highly maneuverable, affordable and optimized for learning household chores. Its foundation ensures efficiency and smooth operation, even in messy environments.”
The robot consists of a modular aluminum frame, moves on 360-degree rotating motorized wheels, and is designed to be customized by the user by adding robotic arms or sensors. The user can control it via an interface on the smartphone and manage its activities according to his or her needs. The robot meanwhile “learns” from the instructions given to it to perform different tasks. In this sense as the researchers explain, the robot could help collect high-quality data to train artificial intelligence algorithms for robotics through learning by imitation.
“By making the design and software completely open source, we aim to democratize access to a robust mobile manipulator, enabling researchers around the world to advance mobile manipulation research,” Jimmi Wu emphasizes.
TidyBot++, is expected to cost around $6,000, but at the moment no official release date has been announced for the product, which for this reason is not yet on sale either online or in physical stores.
