are robots' motions staggered because of the limitation

>>8258489
Both.

If you really wanted smooth motion, I'd consider constructing a mimicry of muscles and joints.
Shouldn't be too difficult to construct an arm.

> theoretical physics/maths btw.

File: bostonrobot.jpg (866KB, 2056x1536px) Image search: [Google]

866KB, 2056x1536px

>>8258496
i think i am going to do something like this.

i'm in my last of 6 years of double electrical/mechanical engineering and comp sci minor

really wanted to go into robotics but i'm having a hard time finding a job with no experience, so i'd like to try my hand at building my own. probably will be difficult as fuck but at this point i have a good spectrum of experience; i think i could pull something off

Isn't power a limiting factor in robotics? Could you build a useful robot with a diesel engine?

>>8258587
That would work depeding on the the robot.

>>8258489
>>>>limitations
it's like it is because they aren't made better dumbass

>>8258489
It's mostly caused by the control methods used. There's been a lot of work done for industrial robots, which are very stiff and deliberate, and much less research done on building and controlling robots for natural, fluid motion.

>>8258500
Make a robot that is like a dog that you can kick but it never falls. Also make fishes

>>8258624
Most dogs fall when you kick them

>>8258489
Those are hobby servos, which aren't great, but are capable of smooth "organic" looking motion. They were used in a lot of Jim Henson's animatronics for example. This person has done a very simple control program:Go to a position at a set speed; stop; go to another position at a set speed; stop... They're not taking into account the inertia of the arm or making curved paths.

>>8258489
Robotic motion is staggeredby the limitation of human ability.

File: variable-impedance actuator.png (740KB, 1024x1021px) Image search: [Google]

740KB, 1024x1021px

>>8258489
For walking and other dynamic motions, it's limited by the actuators. When we jack up the actuator bandwidth and torque in simulation, we get robots capable of doing amazing things. Heck Boston Dynamics and Schaft have both demonstrated what can be done when you have high torque actuators. Schaft won the first DARPA robotics challenge because of their actuators.

For grasping and manipulation it's limited by computing. Honestly, we probably have all the computing power we need for grasping, we might not be using it correctly. Robots have a really fucking hard time dealing with flexible objects, it takes a PR2 a really long time to fold a towel because of computing power/algorithm limitations.

Right now actuators aren't developing as fast algorithms and computing power are.

Your pic related uses fucking hobby servos, which are just about the worst actuators ever for robots.
>>8258496
>>I'd consider constructing a mimicry of muscles and joints
Simply copying muscles and joints isn't a very good approach. We've seen this before. People tried to make heavier than air flying machines by copying birds, trying to make giant flapping machines. In the end, what won the day was understanding the principles behind flight rather than blindly copying what was already there.

Simply copying an arm exactly with artificial muscles is rather silly when we can embody the principles arms use, such as variable impedance actuation, with appropriately designed rotary actuators


>>8258587
Boston Dynamics has built robots with internal combustion engines. In fact, some of their robots must use internal combustion engines because of the inefficient hydraulic actuators they use and how they use them. But this still doesn't solve all the problems, heck even having a tether that can provide UNLIMITED POWER still doesn't solve all the problems. In fact it introduces more, Boston Dynamics LS3 was rejected by the military because it was too loud

>>8258701
good info mate

can anyone recommend some good robot companies for an engineer to try to get into?