Development and improvement of articifial intelligence (AI) is one of the premier challenges for our society. Today, AI assists users through smart device assistants, autopilot programs, and even with recommendations for streaming media. In coming years, AI is expected to become even more common in day-to-day life as robotic assistants, delivery drones, and even self-driving vehicles transform society.
Given the vast range of hardware and components now available, a reliable, repeatable basis is needed for development of AI programs that will interact with physical hardware. Knowing this, Google and Cornell University have released the study-- "ROBEL: Robotic Benchmarks for Learning with Low-Cost Robots".
Influenced by the high-cost of current robotic hardware used to develop AI resources, the ROBEL project introduces an open-source platform of low-cost, highly-effective robots and resources used to set benchmarks and empower research and development on phyiscal hardware.
ROBEL includes plans and reference material for three unique robots- D'Kitty, D'Claw, and D'Lantern. Check them out in the Google AI blog, as well as their research paper linked directly below!