Bipedal humanoids may indeed come, but quadrupeds are already here. They are in laboratories, doing inspections at power plants and refineries, playing soccer and even, much to the concern of many, becoming police officers.
Boston Dynamics' Spot is easily the most instantly recognizable of the group, but many startups and research institutions have put their own spin on the category. Heck, even Xiaomi made one for some reason. While bipedal providers are looking to prove their work, quadrupeds are doing the work.
Google's DeepMind team (which recently absorbed a large chunk of Alphabet's beleaguered Everyday Robots team) just published a research document which describes a possible benchmarking system to quantify the performance of these machines. With a name like “Barkour”, one has to wonder if the department worked backwards from the title.
Google Research points to several impressive feats performed by quadrupeds over the years, from climbing mountains to running and jumping ("flipping is a lot easier than walking," an MIT professor once told me), but there haven't actually been any. a baseline. to determine the effectiveness of the system.
Since these machines are inspired by animals, the research team determined that real animals would provide the best analog performance for their robotic counterparts. That meant setting up an obstacle course in the lab and having a dog run it – he looks at the tenacious little dachshund above. The course was made up of four obstacles in a 5x5 meter area, which is more dense than the dog shows that inspired it.
Performance is rated on a scale of 0 to 1, a simple binary to determine whether the robot can successfully cross the gap in the roughly 10 seconds it takes a dog of similar size to do so. Several penalties are for slow speeds and jumping or missing obstacles on the course. Google concludes:
We believe that developing a benchmark for legged robotics is an important first step in quantifying progress toward agility at the animal level. […] Our findings demonstrate that Barkour is a challenging benchmark that can be easily customized, and that our learning-based method for solving the benchmark provides a quadruped robot with a single low-level policy that can perform a variety of low-level agile skills.
The organization says Barkour has proven to be an effective reference point even in the face of inevitable unexpected events and hardware problems. The robot dog used in the test was able to get up and return to the starting line on its own in case of failure.
