I dont think the RCB-1 controler regurgitates the previous position. The servos really do have a signaling protocol that allows the controler to retrieve the current position.
Check out the PDF called "ICS Servo Special Protocol" here: http://robosavvy.com/Support/2
This will be validated experimentally soon.
The robot can last up to 25 minutes with a NiCD 600MaH battery.
What makes this robot unique is the use of the Kondo ICS-KRS series of servo motors that provide feedback not just PWM-based position control. Many hobby roboteers would be glad to learn about dozens of servos on the market that give you position feedback without having to hack the heck out of the servos and possibly insert a tiny15 into the gearbox in order to get some position feedback.
There are a couple of Korean companies that make servos that are communicable via serial, these are Robotis and Megarobotics. Probably all other servos on the market are one-way controled by PWM.
intelligence is all about low-level feedback control. when you take a million small such entities working together other things emerge like stragegy and and soul..:)
Congratulations on posting the first non-troll comment!
The AI-enabling feature is the position feedback that you can get from the servos. In order to create AI you need a feedback loop involving getting sensor data and updating servo positions repeatedly.
Normal RC-servos were designed to be embedded in a RC-car or RC-plane and controled by a human. In this case the feedback loop is acheived by the human seeing what the car is doing and updating the direction with his joystick.
For an autonomous robot to work, the cpu needs to be fed with lots of sensor data at high refresh rates. Hence this humanoid has the potential of becoming autonomous.
Personal robotics today doesnt exist outside of the hobbist techie and academic realms. But M*ft wont ignore an opportunity once the turnover is worth more then a few hundred million dollars and growing.
As they did with the Internet, Java, handhelds, games, phones, settop-boxes etc. they may eventually take over the personal robotics market too.:-(
LS
http://robosavvy.com/
Slashdot Mirror
Check out the PDF called "ICS Servo Special Protocol" here: http://robosavvy.com/Support/2
This will be validated experimentally soon.
I know someone who has done a similar project developing interesting gaits that utilize current AND position feedback form the servos. http://www.mundobot.com/projects/melanie/v2/enmela nie2.htm
What makes this robot unique is the use of the Kondo ICS-KRS series of servo motors that provide feedback not just PWM-based position control. Many hobby roboteers would be glad to learn about dozens of servos on the market that give you position feedback without having to hack the heck out of the servos and possibly insert a tiny15 into the gearbox in order to get some position feedback.
There are a couple of Korean companies that make servos that are communicable via serial, these are Robotis and Megarobotics. Probably all other servos on the market are one-way controled by PWM.
try to google : feedback control robot
for example:3 .htm
m
http://mundobot.com/projects/melanie/v3/enmelanie
http://www.ynl.t.u-tokyo.ac.jp/~fabio/bipedloc.ht
Sony Quorio, Honda Asimo etc.
The AI-enabling feature is the position feedback that you can get from the servos. In order to create AI you need a feedback loop involving getting sensor data and updating servo positions repeatedly.
Normal RC-servos were designed to be embedded in a RC-car or RC-plane and controled by a human. In this case the feedback loop is acheived by the human seeing what the car is doing and updating the direction with his joystick.
For an autonomous robot to work, the cpu needs to be fed with lots of sensor data at high refresh rates. Hence this humanoid has the potential of becoming autonomous.
Personal robotics today doesnt exist outside of the hobbist techie and academic realms. But M*ft wont ignore an opportunity once the turnover is worth more then a few hundred million dollars and growing. As they did with the Internet, Java, handhelds, games, phones, settop-boxes etc. they may eventually take over the personal robotics market too. :-(
LS http://robosavvy.com/