Gadgetory

when I was seven years old I watched the movie star wars for the prey first time at all my way back home in the car I decided to become a robot scientist and it never changed my mind and I'm here today my name is Dennis ha I'm a professor of mechanical engineering at Virginia Tech and the founding director of Ramallah robotics is a really a really wide field from one end of the spectrum from where the mechanical sciences mechanical engineering kinematics design to the other end of the spectrum where the computer science side of artificial intelligence and anything in between you really need to know everything to develop a pool system we first try to generate walking motion just by you know almost like a stop-motion animation just creating the emotional legs and hopefully it's gonna walk and it turns out that that's not the case both Charlie at Darwin use a method called GMP zero moment point control method CMP walking is more like a robot walking trying to keep my GMP right under my foot so as long as you can create a motion that the CIP is under your foot or the foot support polygon then theoretically you've got to be stable humans walk in a very different way first of all when you walk you swing your legs swing your leg touch the fall stand up again and then you fall forward and you swing your leg and your body also swings forward now zmp is probably the most stable way to implement bipedal walking the problem is inherently zmp method can only work on smooth hard flat strain so the next step in biker walking is a good example is our new robot sapphire you probably know the big dog from Boston Dynamics they're very impressive for Lake robot think of sapphire as a two leg version of the big dog people when they see our robots because they watch sci-fi movies like the iRobot and they say wow this robot Charlie walking this is rainy and cool but why cat why doesn't it jump why doesn't it run power source is a big problem the batteries that we use releasing polymer batteries can for example power up Charlie for only about fifteen or twenty minutes better actuators that mimic biological muscles better we need more compliance and different way of doing force control so all of these component technologies need to catch up be catched up to develop the next generation robotics in about ten years hopefully we'll start to see a more of domestic robots that can actually help people in the home one of these robots to use tools to go for he asked a robot hey fetch me a beer from the fridge you do not want to have a special refrigerator just for the robot you want to go but to open up the door of the fridge or pick up the phone use the scissors all the tools designed for humans thus I believe developing humanoid robots is very important