The Boston Dynamics video from October 2013 begins with a wide pan across a mountainside in Southern California’s Mojave Desert. Slowly, the viewer’s eye picks out two shapes inching their way up the uneven terrain: a man and a robot.
The robot is Boston Dynamics’ LS3, a horse-sized machine that picks its way daintily up rocky slopes and snow-laden hills, despite the 400 pounds of baggage it is capable of carrying on its back. Like the company’s more famous BigDog and Atlas, LS3 is made for locomotion, even over the toughest terrain.
When Google bought Boston Dynamics last week, that’s the technology that it acquired. Along with taking on rough ground, Boston Dynamics’ robots are fast and able to overcome the unpredictable. Its CHEETAH bot holds the world record for the fastest legged robot, logging a top speed of 29 miles per hour. BigDog can recover from a kick, even when standing on an icy surface. Its humanoid robots do just fine with obstacles too.
The robots come packaged with related patents, a sore spot for Google in its push into mobile technology: a robot leg suited to walking, running and climbing; methods for controlling a jumping robot; the hydraulic system behind BigDog’s jerky movement. They also come packaged with an experienced team of researchers, who founder Marc Raibert told me will all continue working from the company’s lab in Waltham, Mass.
The seven other robotics companies Google acquired form a diverse group; humanoid robots, arms and artificial intelligence are among their focuses. While Holomni specializes in rolling robots and SCHAFT and Meka’s bipedal machines can walk, none of them are focused on creating a robot that could walk through obstacles as ably as a human.
As a result, Boston Dynamics is a necessary piece of the puzzle if Google wants to draw upon existing robotic achievements to build a well-rounded machine. While open-source tools like the Robot Operating System have begun to make it easier to build a robot and every year brings more capable machines, robots remain fairly specialized.
While the hardware Google acquired is certainly impressive, I’m more interested in seeing what Google can accomplish with sensing and artificial intelligence; two fronts where innovation is badly needed to complement already impressive hardware. Between its experience building self-driving cars and past work with AI, not to mention a practically unlimited budget, Google is a fine candidate to take on the challenges that have long faced roboticists.
“We have had a great time at Boston Dynamics, building our unusual robots and bringing them this far along,” Raibert said in an interview with Gigaom. “Now we are excited to take this next step, to see how much further ahead we can take robotics, working with Google’s gangbuster team.”
Open Source Robotics Foundation CEO Brian Gerkey noted that the most exciting result to come out of the acquisition is how Google taking an interesting in robotics will push the industry ahead as a whole.
“There’s been a renaissance in personal robotics for a few years now, particularly here in Silicon Valley,” Gerkey said. “Google’s acquisitions are bringing a great deal of attention to robotics, but those of us working in the field have seen the momentum building to the point that something of this scope was bound to happen. With respect to Google and others, more investment and activity in the field will ultimately benefit all of us, as it will spur more innovation and hopefully break open new markets for the robot technology that we all love to work on.”