Canon PowerShot S90
ƒ/2
6 mm
1/30
640
Puzzle Series: What is this, or what do you want it to be?
Something I played with today… If it’s too easy, then explain how it works…. =)
25 responses to “What’s That? (101)”
-
I agree but, won’t they be required to have unique finger prints? Oh, and it is capacitive.
-
Fingerprints….
Why do we have them…?
Better grip ?(I thought it was a solar garden light…..)
-
Climbing to something like a gecko ? Like Robert Full ‘s work ?
Or so that robots can know who they are shaking hands to… =)
Or something else…but too Frenchy to say here in public haha =P -
The ridges create more surface area on the finger. Perhaps the ridges can better translate pressure to the sensors underneath the clear materials. Essentially making them more sensitive just like our own fingertips with the many little ridges.
-
I’m going to guess the prints serve both for traction and localizing touch.
-
To mimic real touch sensation when the pressure data is relayed at the controllers fingertips?
-
The finger of robot.
Why fingerprints? Because robots need a way to be identified… too!
-
To simulate the way humans perceive the texture of objects.
When you draw your finger across a surface the texture creates vibrations in your fingertip. The vibrations tell you about the surface. There are frequencies which you’re most sensitive to, and you’ll vary your rate of movement to the most sensitive frequencies.
This can be tested by tracking the rate people draw their fingers across different grit sandpaper.
You can see the curves for different receptors here:
http://www.neurobiography.info/teaching/images/somatosensory/vib... -
Silver things = transducers
The hole in the middle is to let the laser light through.
The ‘fingerprint’ is to code the diffraction pattern of the laser to allow measurement of the deformation caused by external pressure i.e. what it’s touching. -
Bingo bfinney; it is a robo finger and Bingo Jim Rees; the fingerprints give a 30x improvement in texture sensing, but they don’t fully know why. I asked USC Prof. Gerald Loeb about this and whether the vibration modes of ridges would convey more information than a smooth surface: “You would think so, but no. We don’t know how they work.” Sounds like they should talk to Todd Huffman. And solerena clearly wins the "what do you want it to be" part… =)
For a sense of the limitations of robots of today, Loeb asked us to imagine our fingers were frostbit and numb. Our ability to grip and manipulate objects is shot because of the lack of sensory feedback, not because of a hindered ability to move our fingers, as the muscles are up in the forearm.
Tactile feedback is essential for robots to be able to replace the delicate repetitive tasks that humans still do in factory settings, with injury and $20 billion of annual cost.
This robo-finger is a detailed copy of the human finger. Even the fingernail is essential to pin the elastic skin for the slip sensors to pick up the shifting of fluid under the skin.
The pressure sensor is in the blue box; the fluid under the skin couples to it hydraulically. Microvibrations of the skin caused by sliding over textured surfaces are sensed as sound waves by the pressure sensor.
The silver sensors are an array of electrodes that sense location and force of touch vectors as the conductive fluid is displaced, changing the measured impedance.
I asked if there was any area where they have improved the design from biomimicry, but not yet. For example, creating molds for novel fingerprint patterns is too expensive for broad experimentation (I suggested a femtosecond laser as a fix for that).
The flat tip of the finger with the fleshy nub allows for curve sensation and contour tracking.
The thermistor provides heat flux measurements. Here is a short video snippet from the talk that addresses the way we sense a metal table as “cold” and tablecloth as a “warm” material even though both are at room temperature. The same signal processing technique is needed for the robo-finger. And both have a time delay built into the “neural path” since the mere pressing of the skin inward lowers the temperature sensor (since the outer tissues are generally colder than the core).
-
I would guess the texture detail sensitivity is on the order of the ridge spacing of the finger print. Also I would wonder if textures finer then the print detail would produce high frequency vibrations in the realm of viscous damping by the ‘skin’ and conductive fluid. This all seems like something you could model with some SC time.
-
I am not sure that the effect of fingerprints on touch/perception is that clear…
But recalling my Neurology lectures from 1978…..(!) I think tactile sensory input has a lot more to do with skin sensor type and sensor density…
(think retina as an analogy…)
One need only think of certain skin areas that are so endowed…(blush).So in this case above, (robotic design) they serve a (technical ) purpose….
Seems different than how humans perceive…
We need a neurobiologists opinion… -
I keep ending up at your stream, Steve, by various circuitous routes.
You post the coolest stuff. This is a prime example. Keep up the great work!
Leave a Reply to ukweli Cancel reply