posted by Bryon Moyer
Pico-projectors sound like an absolute dream for anyone who has to travel with a standard projector in tow. Being able to attach a small device to your phone has the potential to eliminate one heavy bag and to give your shoulder a badly-needed rest. In reality, of course, they don’t have the power that a “real” projector has, so you’re not going to thrill an audience in a big room using your phone. But when the need arises, they can allow for ad hoc display of anything on a modest patch of wall.
Microvision is in this business; they make a pico-projector that uses a single laser for each color and a movable mirror to create a workable image for a small presentation or a heads-up display.
Much of their future work will focus on the same thing everyone else is focusing on: reducing power. Right now they do that by blanking any lasers that aren’t in use at a given instant; future gains will be had by improved green and blue LED efficiency, working an efficient modulation scheme, and taking advantage of Moore’s Law improvements in the digital controller. All to get under 1 W. I know; sounds high, but this thing is all about projecting enough light to overcome the ambient. Ever wonder how much power those monster projectors in auditoriums suck down? Lots, judging by the cooling units… And that’s for use in a darkened room…
But power reduction is an evolutionary improvement. This year they’re looking to change the game by measuring reflections when the scanning laser beam is blocked by something like a finger. Since the system knows where it was shooting the beam when the unusual reflection occurred, it can provide coordinates for the obstruction. In other words, if you point, the system can figure out what you’re pointing to.
This really changes the nature of the presentation from simply a passive transmission of visual images to a – must I use the phrase? – collaborative engagement. The idea is to turn the wall where this is being projected into a temporary tablet on the scale of 2-3’ (5-8’ in a darkened room).
I’m assuming that this isn’t trivial to do. The background against which these “reflective anomalies” (I made the phrase up for lack of a better one) are detected isn’t predictable. One day the projector might be in a bright room shining on a yellow wall; another day, another café, perhaps darker, with a light blue wall. So they will presumably need to normalize the current environment as a calibration step so that they can figure out when a reflection is anomalous.
But this seems typical of a lot of the work being done in the realm of smart systems: extract useful signals out of extraordinarily noisy environments. We actually seem to be getting pretty good at it.
This capability isn’t available yet, but they said to watch for it in 2014. And watch we will.
posted by Bryon Moyer
Our bodies have evolved on practical, not principled, grounds. So we have one sensor for photons in the visible spectrum, and we have a completely different sense for photons at wavelengths just longer those in the visible range. One we perceive as light; the other as heat.
But in fact, we now know, intellectually, that they’re just different frequencies of the same thing. It just doesn’t feel that way.
Well, Bosch has taken a page out of the physiology book with its recent infrared detector array. Instead of detecting photons, it also detects heat. They use a porous silicon membrane plus diodes to generate an image that’s admittedly not high-res. And not intended to be.
This isn’t about badass pinpoint-perfect night vision; it’s about low-cost industrial or other applications where resolution per se isn’t the goal. Perhaps you’re trying to distinguish human from animal. Or monitoring train bearings to make sure they’re not overheating.
In fact, this sensor didn’t originate in the consumer Sensortec division; it’s marketed out of the automotive division, which makes me wonder whether they had (or have) a specific auto application in mind. Or perhaps “locomotive” is close enough to “automotive”… (or does that sound crazy?)
I’d send you to a release for more information, but… they didn’t issue one for this. I’d send you to a web page… but I can’t find one (I haven’t found a way to get to these kinds of products on Bosch’s website… it’s all about higher-level modules and systems… even Googling doesn’t help). It was just a conversation I had at the MEMS Executive Congress. I’d say, “You saw it here first,” except perhaps, “You saw it here only” is more appropriate. And has me wondering, “Did I really have that conversation??”
So perhaps your friendly neighborhood Bosch Automotive salesperson would be the next step if you want more info…
posted by Bryon Moyer
We’ve taken an occasional look at physically-unclonable functions (PUFs), and, in particular, IntrinsicID’s implementation of them, as they seem to have gone further in productizing their technologies than others have. As we’ve noted before, PUFs rely on intrinsic physical variation from chip to chip. While this variation may drive IC designers and EDA guys nuts, it’s leveraged in PUFs so that a unique key can be created for each individual machine or USB dongle. The key is never stored, so it’s much much harder for some ne’er-do-well to crack.
As we’ve also noted before, IntrinsicID launched a security facility called Saturnus for protecting cloud contents. They have now integrated their PUF technology with InsideSecure’s secure microcontrollers to create USB keys. Combined with Saturnus, they now work with DropBox: Selected files, when uploaded, will be encrypted first. They cannot be decrypted without the key that’s on the USB stick. Wait, no, it’s not literally on the stick – it’s generated on the stick. That’s what makes this work; the key is generated each time it’s used.
As far as I can tell, this is the furthest PUF technology has gone, at least in the white-project commercial world. (Who knows how it’s being used in black projects…) By integrating itself into an accepted consumer application like DropBox, it becomes relevant to a broad range of user.