We agreed at our previous (9/28/09) meeting that basic motion states would be helpful, and for this an accelerometer would obviously be helpful:
- Upside down
In the Goat study the accelerometer was calibrated to observation. A problematic issue for monkey tracking is group fission and fusion. It was agreed for this reason that even knowing the motion type might be helpful without knowing where they are, as long as it had one other vector, such as time.
Which monkey are we designing this for?
Christina would prefer to design this for Spiders: they spend more time vertically, while Titi’s are more typically quadrupedal, so one can more easily make assumptions about them: their locomotion is typically quadrupedal and foraging is upright. As well, spiders have prehensile tails and forage upside down or with their heads down in the lower “quadrant” (from 0 to -90). In that case one would know generally that they’re feeding (more or less).
Stuffed monkey and/or human modeling?
The 3 states noted above.
- more complex:
- brachiation (modeling as humans would lack bursts of speed: it is hand over hand and energy-intensive
- quadrupedal on branches like dog (howlers & titi always, spiders & woollies both)