The Barkour scoring system is based on how long it typically takes a small dog to complete a course in a novice agility competition. Robots must navigate the course, set up in a 5-meter-square area, and successfully complete four obstacles: walking over an A frame, weaving in and out of poles, executing a 0.5-meter broad jump and stepping onto an end table. Dogs of Google employees, known as “dooglers,” took on the course first, getting through in about 10 seconds; the robots usually take about 20.
As robots become more prevalent in the real world — the New York Police Department, for example, is using a “robot dog” to investigate high-risk situations from a safe distance — creating a standard for mobility that can be adapted for various applications could make the bots even more useful.
“We believe that developing a benchmark for legged robotics is an important first step in quantifying progress toward animal-level agility,” the Barkour team wrote in a blog post, adding that their benchmark can be “easily customized” for different situations.
Source: Google Research
This story originally appeared in the July/August issue of Government Technology magazine. Click here to view the full digital edition online.
