Augmented reality sucks without being able to augment reality. Usually, this is done by looking at a screen with an image captured by a camera. This means that you’re limiting your entire vision to that of the sensor of the camera. This is fine in the future when we have high tech cameras that surpass our eyes in sensitivity, but sucks now when all we really have to work with are webcams.
The problem with getting light onto the retina is that, because of those pesky laws of refraction, the ray of light has to come from in front of the eye. With not so great augmented reality like this pair of ski goggles, you get a screen that you can look at, acting more as a meter. With fancy systems like retinal laser projection, a glass plate is placed at an angle in front of the eye to reflect some of the light from a projector placed above, facing down. This is nice but not so stylish and fairly bulky.
An ideal system would have:
- Clear view port - Need to be able to actually see reality and project information on top of it. Optically clear, just clear to the user (think shade glasses…those have regions that are completely opaque).
- High resolution - At least a few megapixels.
- Fast update - Otherwise, any augmentation wouldn’t be able to track movement (relative to eye).
- Thin - If you fall on your face, you don’t want to lose an eye.
Computer vision software to point out messes in a room. Uses a laser beam to highlight messes in the room to annoy the mess makers so they’ll clean it up!
The CV system would probably use background subtraction to find the mess. Maybe with some statistical feature detection to ignore things such as a bag of bread moving around on a counter top. A time delay between detecting a mess and pointing out the mess.
Not sure if I posted this already…
Aerodynamic rubber bands!!!! Come on!!! Cross section cut would probably look like a teardrop, with the fat (possibly flat) side inward to make it possible to shoot and give it some mass compared to a flat band…even though this is aerodynamically backwards. Ideally, I would think the fat part would face forward, like a wing, giving the front of the rubber band less drag and the tail greater drag, increasing stability. You would shoot it by forcing it inside out, where it would restore its shape after being released.
Realistically, a very flat band, like the Aerobie, would probably be best. To shoot, you would just shoot it like a regular rubber band, where it would flatten on its own after being shot.
Could make it out of some nice stretchy rubber compound, maybe latex…has to be fancy for someone to buy it. I’d love to have one when I was a kid…it’d be great to shoot rubbor bands across the backyard! POW POW! MY LIP IS BLEEDING!!!!
So, people with crappy audio/video sources and receivers that use PLL’s for SPDIF clock recovery complain could get a nice clean signal.
Basically, would decode the SPDIF signal and re-encode with a more stable PLL, maybe hiding timing errors in data bits or preamble bits.
That’s all. Would be very sensitive of course. Aperture and overall size could/would be large for high bandwidth. Point of focus would be attached to the mass.
Auto echo cancel app for audio/video recording for smart phones since audio recording kind of sucks right now. Would attempt to find and cancel worst of the room reverberation. Would be a deconvolution problem that would probably be handled on a server rather than the actual device, like voice recognition is these days.
An LCD aperture for cameras. Would be a gimmicky feature for inexpensive cameras. Would use an LCD rather than leaf based aperture. Would allow programmable shapes to be used for aperture, making pretty pictures in blurred images.
Could use it for an ND filter, too.
But mostly, could use it for coded aperture deconvolution for focal blur removal and depth extraction!
Obvious. Use a cheap single photomultiplier (can find them for $25) and a DLP chip. A lens system would focus and image onto the DLP. Activating single pixels on the DLP would direct the light into the photomultiplier, giving a detector resolution equal to that of the DLP. Gain could be increased, at the expense of resolution, by activating multiple pixels on the DLP. Similar to a DLP television, a color wheel could be used for color images.
Benefit over a CCD? Probably none. One frame per second would require each pixel to switch on for 4.8us at 1920x1080 image. So, 10ns for a shutter speed of 1/400.
Rotating mirrors would probably be better for night vision.
Found someone doing this: http://www.engadget.com/2006/10/03/rice-university-scientists-create-a-revolutionary-single-pixel-c/, although, not with photomultiplier. They’re doing it in an interesting way though…by randomly selecting pixels, then constructing the image after…maybe for noise suppression, or to get the light levels up above the detector threshold?
Come on, disk space is practically free these days. Server where scientists can upload research papers so anyone can download them for free. Come on, $30 to read if orally administered baking soda increases race horse performance, with not a cent of that going to the author…wha!?
For a useful reference manual, you need to be able to quickly get to the topic you’re trying to review or equation you’re trying to look up. This is PERFECT for a touch screen with a slick collapsible interface. Allowing an engineer to have all of his references, with all of his notes in digital form, able to upload/back up would be great.
Maybe the business would be some software to automatically create and populate the structures from scanned books. A book interface in a sense.
Just that. A mini mechanical piano that uses keys to move levers to excite the teeth on a music box comb. Ideally, it would allow recording and playback with a few solenoids tied to the keys. That way, you could record and play back your own music.
For an electric/hybrid vehicle:
Use an electromagnet placed in the suspension to recharge the battery, or apply power to the motor directly (whichever is the most efficient). The current from the electromagnet would be the current going to the electric motors.
An additional benefit would be electronic control of the suspension “dampening” forces. Since the maximum force would be at maximum power to the electric motor, and minimum force at no power, you would still need independent dampeners…maybe like ferromagnetic dampeners like used in the corvette.
Looked around and most of the ideas were to use permanent magnets…which are weak. The main advantage would be for long straights, like on a bumpy freeway.
Could always get an accelerometer and calculate the amount of power per mile using the closest freeway…need an accelerometer.
Man…we really need room temperature superconductors…would make things so much easier.
Maybe useful for MEM’s.
I saw a TED show about a man with a broom handle and a weight. The weight had bristles so, when the handle was vibrated, the bristles would “ratchet” against the handle and move the weight up.
I think it would be better use rotations for this concept since you wouldn’t have to add tension waves.
Bristles would be angles and make contact with the handle. When the handle was slightly rotated, the bristles would rotate and push the weight upwards. The bristles in the opposite orientation would slip and hold.
Having a rotational oscillating, back and forth, would cause the weight to move upwards.
Something like a digital picture frame that listens in on your conversation using fairly crappy (crappier the funnier the results) voice recognition to try to get a gist of what you’re talking about. Whatever the topic, it would show relevant pictures and video (probably muted) about whatever it heard.
It could use youtube, flickr, google image search, etc.
Ideally, it wouldn’t listen very well and get the topic wrong…resulting in lol’s rather than serious’s. You can imagine some pretty awkward and funny pictures that would show up. Maybe it could show a little label of what it heard and what it searched for.
Update: Decided sleep was overrated, so put it together tonight. Just uses random pictures from the first google images result page. I think videos would be funniest, especially when people would start talking about what you just said. Might be funny to use collections of news clips so it looked like the news was related to whatever you were talking about (most word hits, something like that).
Little RC car that drifts by itself. Would need a gyro or accelerometer and electronic compass. Would end up being a stability control system.
It would be intended for kids that couldn’t actually drift an RC car, but could let them have fun before they got bored (more likely) or help them learn by having different levels of assistance.
And yes, people already drift RC cars…although, I was thinking plastic wheels that would have little grip.
