| |
34
Chapter 3. Pilot Studies: Lessons Learned from Early Projects
Table 3.1: Assembly actions and possible sensor configurations
provides some sensor alternatives which may influence the precision of the perception
and the total sensor cost.
Taking a look at the simple example of preparing a horizontal board (inserting four
dowels), we see that this can be recognized using only one accelerometer attached
to the board itself (accelerometers can detect di erent movement directions). In
Figure 3.3(a) the output of the accelerometers is plotted: the four vertical patterns
indicate the hammering of the four dowels.
Alternatives would be to enhance the hammer with an accelerometer or to use an
electric contact that reacts as soon as the dowel has been fully inserted. What makes
this example interesting is that one can insert the four dowels in many di erent ways.
As the dowel insert points lie on opposite sides of the board, one usually has to turn
the board during its preparation. This action of turning the board can also be easily
recognized using the accelerometer attached to the board. How often and when the
board is turned however can be varied by the user.
Accelerometers are also used to detect the orientation of the boards, in relation to
each other. Furthermore, force pressure sensors mounted at the boards edges de-
tect the force between boards: The force sensors can distinguish (see Figure 3.3(b))
between joining boards (lower force) and tightening screws (stronger force). Addi-
tionally, IR-sensors detect the location and orientation of the boards with respect
|  |
|
| |
|
|
