|
|
|
|||||||||||
Problem Statement |
|
|||||||||||
|
|
Quibbler |
|
||||||||||
|
Back-EMF technology uses the voltage generated by a spinning motor to obtain the speed of the motor’s rotation (figure 1). A back-EMF package can essentially consists of an A/D converter and an FPGA to speed up the data acquisition process and a processor to do the processing of the gathered back-emf values. And it also includes a permanent magnet motor of course. The cost of a back-EMF package depends on the volume. For an ASIC (Application Specific Integrated Circuit), the cost may be only $5.
Figure 1 A single back-EMF speed measurement [1] With back-EMF, we can recorder an action and then manipulate the recorded action. This capability gives us many possibilities in human robot interaction design. We can recorder an action and then transfer it to another robot. It’s also possible to manipulate the data to speed up or slow down the action when we replay it or even replay it in reverse direction. For example, if we lift the left arm of the robot, its right arm automatically puts down like a mirror of the left arm’s action. With recorded data of multiple joints, it’s possible to consider the robot’s motions as a whole and therefore achieve more fluid robot actions. In robot control, we can use back-EMF to build a tiny robot as an ‘avator’ that works as a joystick to control the real robot. The limitation of back-EMF technology is that there is drift in the estimated position. In practice, it can be as low as 0.1% or as high as 1%. So if we move the motor 90 degrees, we could have a position error of as much as 0.9 degrees or as low as 0.09 degrees. There is also variance in the back-EMF constant between motors. This might complicate recording using one motor and playing back using another motor if we want a high degree of accuracy between the two motors, for example.
|
|
||||||||||
|
|
||||||||||||
|
|
||||||||||||
| PHRI • Fall 2005 • Carnegie Mellon University | ||||||||||||