New algorithm enables MIT cheetah robot to run and jump, untethered, across grass.  

 

Jennifer Chu | MIT News Office
September 15, 2014  

 Speed and agility are hallmarks of the cheetah: The big predator is  the fastest land animal on Earth, able to accelerate to 60 mph in just a  few seconds. As it ramps up to top speed, a cheetah pumps its legs in  tandem, bounding until it reaches a full gallop. 

 Now MIT researchers have developed an algorithm for bounding that  they’ve successfully implemented in a robotic cheetah — a sleek,  four-legged assemblage of gears, batteries, and electric motors that  weighs about as much as its feline counterpart. The team recently took  the robot for a test run on MIT’s Killian Court, where it bounded across  the grass at a steady clip. 

 In experiments on an indoor track, the robot sprinted up to 10 mph,  even continuing to run after clearing a hurdle. The MIT researchers  estimate that the current version of the robot may eventually reach  speeds of up to 30 mph. 

 The key to the bounding algorithm is in programming each of the  robot’s legs to exert a certain amount of force in the split second  during which it hits the ground, in order to maintain a given speed: In  general, the faster the desired speed, the more force must be applied to  propel the robot forward. Sangbae Kim, an associate professor of  mechanical engineering at MIT, hypothesizes that this force-control  approach to robotic running is similar, in principle, to the way  world-class sprinters race. 

 “Many sprinters, like Usain Bolt, don’t cycle their legs really  fast,” Kim says. “They actually increase their stride length by pushing  downward harder and increasing their ground force, so they can fly more  while keeping the same frequency.” 

 Kim says that by adapting a force-based approach, the cheetah-bot is  able to handle rougher terrain, such as bounding across a grassy field.  In treadmill experiments, the team found that the robot handled slight  bumps in its path, maintaining its speed even as it ran over a foam  obstacle. 

 “Most robots are sluggish and heavy, and thus they  cannot control force in high-speed situations,” Kim says. “That’s what  makes the MIT cheetah so special: You can actually control the force  profile for a very short period of time, followed by a hefty impact with  the ground, which makes it more stable, agile, and dynamic.” 

 Kim says what makes the robot so dynamic is a custom-designed,  high-torque-density electric motor, designed by Jeffrey Lang, the  Vitesse Professor of Electrical Engineering at MIT. These motors are  controlled by amplifiers designed by David Otten, a principal research  engineer in MIT’s Research Laboratory of Electronics. The combination of  such special electric motors and custom-designed, bio-inspired legs  allow force control on the ground without relying on delicate force  sensors on the feet.   

 Kim and his colleagues — research scientist Hae-Won Park and graduate  student Meng Yee Chuah — will present details of the bounding algorithm  this month at the IEEE/RSJ International Conference on Intelligent  Robots and Systems in Chicago. 

 Toward the ultimate gait 

 The act of running can be parsed into a number of biomechanically  distinct gaits, from trotting and cantering to more dynamic bounding and  galloping. In bounding, an animal’s front legs hit the ground together,  followed by its hind legs, similar to the way that rabbits hop — a  relatively simple gait that the researchers chose to model first. 

 “Bounding is like an entry-level high-speed gait, and galloping is  the ultimate gait,” Kim says. “Once you get bounding, you can easily  split the two legs and get galloping.” 

 As an animal bounds, its legs touch the ground for a fraction of a  second before cycling through the air again. The percentage of time a  leg spends on the ground rather than in the air is referred to in  biomechanics as a “duty cycle”; the faster an animal runs, the shorter  its duty cycle. 

 Kim and his colleagues developed an algorithm that determines the  amount of force a leg should exert in the short period of each cycle  that it spends on the ground. That force, they reasoned, should be  enough for the robot to push up against the downward force of gravity,  in order to maintain forward momentum. 

 “Once I know how long my leg is on the ground and  how long my body is in the air, I know how much force I need to apply to  compensate for the gravitational force,” Kim says. “Now we’re able to  control bounding at many speeds. And to jump, we can, say, triple the  force, and it jumps over obstacles.”   

 In experiments, the team ran the robot at progressively smaller duty  cycles, finding that, following the algorithm’s force prescriptions, the  robot was able to run at higher speeds without falling. Kim says the  team’s algorithm enables precise control over the forces a robot can  exert while running. 

 By contrast, he says, similar quadruped robots may exert high force,  but with poor efficiency. What’s more, such robots run on gasoline and  are powered by a gasoline engine, in order to generate high forces. 

 “As a result, they’re way louder,” Kim says. “Our robot can be silent  and as efficient as animals. The only things you hear are the feet  hitting the ground. This is kind of a new paradigm where we’re  controlling force in a highly dynamic situation. Any legged robot should  be able to do this in the future.” 

 This work was supported by the Defense Advanced Research Projects Agency. 

•  but why cheetahs in steemit very danger , Enjoy