Starting and Open Source Project?
Posted: Tue Oct 14, 2014 12:14 am
Hello,
Where does it all start? Is it simple idea we've all seen before, and just never really approached as a group, where the group would be the best idea?
Okay, I want to start an Open Source Humanoid Robot. Well, it still takes parts, this is the kind of place to start, in the middle of an Open Source Lounge. I know a bit about writing AI software, if you wanted to hold a conversation with a robot, you could. It could be done with a Banana Pi, and Solid State Hard Drive. But, the way sensors and motors all work, I figure the USB ports on the things would be the internal network of the Droid. The idea would be to work out a long list of bugs, and I see them in ASIMO, and other Humanoid Robots. You would think, you only need one accelerometer and gyro, wrong. You need those in the arms and legs as well to pre calculate, the RISK architecture of being human. You don't really revolve all around your inner ear, you feel g forces throughout your body, how you coordinate all of that doesn't ever stop you from using you're body's state of motion to calculate your own center of gravity.
If I strictly worked on the conversational AI, it could be taught or learn to talk to it's arms legs and hands, as well as listen to the conversation of sensors. I would use just one Banana Pi, to calculate the center of gravity and just report that in a cycle. Another will pre-calculate based upon pre planned paths, where error correction just waits for a slip or something. It's like saying a 3D version of everything has to happen first, it happens faster than we can perceive it, and the result is that it attempts to follow the planned path, has path error correction and obstacle avoidance. Just literally, have a goal of getting out housework, or maybe getting you're bots multiple jobs.
Where does it all start? Is it simple idea we've all seen before, and just never really approached as a group, where the group would be the best idea?
Okay, I want to start an Open Source Humanoid Robot. Well, it still takes parts, this is the kind of place to start, in the middle of an Open Source Lounge. I know a bit about writing AI software, if you wanted to hold a conversation with a robot, you could. It could be done with a Banana Pi, and Solid State Hard Drive. But, the way sensors and motors all work, I figure the USB ports on the things would be the internal network of the Droid. The idea would be to work out a long list of bugs, and I see them in ASIMO, and other Humanoid Robots. You would think, you only need one accelerometer and gyro, wrong. You need those in the arms and legs as well to pre calculate, the RISK architecture of being human. You don't really revolve all around your inner ear, you feel g forces throughout your body, how you coordinate all of that doesn't ever stop you from using you're body's state of motion to calculate your own center of gravity.
If I strictly worked on the conversational AI, it could be taught or learn to talk to it's arms legs and hands, as well as listen to the conversation of sensors. I would use just one Banana Pi, to calculate the center of gravity and just report that in a cycle. Another will pre-calculate based upon pre planned paths, where error correction just waits for a slip or something. It's like saying a 3D version of everything has to happen first, it happens faster than we can perceive it, and the result is that it attempts to follow the planned path, has path error correction and obstacle avoidance. Just literally, have a goal of getting out housework, or maybe getting you're bots multiple jobs.
