Is there anything worth being invented?

Thread Starter

praondevou

Joined Jul 9, 2011
2,942
Of course I'm talking about electronics (maybe mechanics) related stuff. Also I'm talking about someone living in the western world.

I just finished a (simple) RF remote light switch, the parts cost me about $35 I think. On ebay I get the same thing (finished unit) for $10, directly imported from China. (Not exactly, mine is expandable, theirs not). I know that mass production is an important part of the secret to low prices.

However, if one wanted to build something one cannot buy because it either doesn't exist or because it's too expensive, what would it be?:confused: And I mean something useful, not just for fun.

Where I lived 25 years ago we knew cool stuff existed, but we didn't have access to. At this time it WAS fun to build something.

Almost everything nowadays is either too cheap (you buy it) or to complicated (you can't build it).
 

thatoneguy

Joined Feb 19, 2009
6,359
Yes.

Look at the single wheel version of the Segway for example.

We now have cameras that trigger when the smile is detected, that's a cool invention.

There are a lot of opportunities, the problem being, you need to work very fast, and with FPGA for demos to get the speed needed.

Motion detection that is triggered only by a human shape, instead of a dog, for example.

Night Vision cameras that combine the "green screen" with a false color overlay of abnormal temperatures (may already be done w/.mil)

Smaller GoPro cameras with betterstorage, slap a GPS into them as well.

Yes, most of this is modification of existing tech. But once you get a leg in, you get to work for a company doing R&D and get all sorts of cool toys and a hefty budget to play with for ideas. The downside is they own all of your ideas then.

Memory augmentation for the human brain.

Auto-polarizing glass currently exists, but is far too expensive to use on a wide scale. Basically, single pixel B&W display, goes dark to light depending on polarizing current. Right now, it's still liquid crystal, form a film version of it.
 

MrChips

Joined Oct 2, 2009
30,706
Is there anything worth being invented? Sure there is. I have so many projects on the go right now they'll all never get done before I go.
 

Thread Starter

praondevou

Joined Jul 9, 2011
2,942
@thatoneguy

Ahem... :)

Thanks. I work in an R&D departement (not an R&D engineering company though) but I don't consider myself clever enough to develop any of what you said on my own.
Part of the problem nowadays I think is that you need advanced programming skills for most of the projects...
I liked the human shape idea. I also would like to invent a device that could under any circumstances reliably recognize a single voice command in a room.
 

Thread Starter

praondevou

Joined Jul 9, 2011
2,942
Is there anything worth being invented? Sure there is. I have so many projects on the go right now they'll all never get done before I go.
Do you mind telling me what projects you have in mind? Or are they going to be patented?

Yes sure there are simple projects we can develop, but sometimes it seems for every thing you can imagine there is already an answer, and if there isn't it's because a single human can't do it, for lack of time or money or both...
 

thatoneguy

Joined Feb 19, 2009
6,359
Part of the problem nowadays I think is that you need advanced programming skills for most of the projects...
I liked the human shape idea. I also would like to invent a device that could under any circumstances reliably recognize a single voice command in a room.
We have pretty much established that electronics are capable of sensory and signal matching at the level of a human, or better (sight and sound, anyway).

Is the single voice command something a human would recognize in that same room? (crowd, others talking, TV or music on, etc)

If so, break it down, this is where speedy DSP comes into play, but if a human can recognize a random word spoken in a room, a computer should be able to as well.

Voice menus can currently capture "Tech support" compared to "customer service". Google can match a 3 second snippet of a song to the artist and title.

There is no reason (other than time and money) that a single voice command in a room cannot be decoded, especially if the processor already has a sample of that command from that person to match.
 

MrChips

Joined Oct 2, 2009
30,706
Many of these are proprietary, so I cannot reveal them until they come out on the market. Yes, we are applying for patents on some. Amazingly, four of them came about from ideas or contacts through AAC within the last month. Opportunites abound on AAC! I've got more work than I can handle so we are on the look out for bright people on AAC. Wish to apply?
 

Wendy

Joined Mar 24, 2008
23,415
Why yes! Do I have to move?

There are a lot of people here who could use employment. If you are serious PM me, one name stands out in particular.
 

Thread Starter

praondevou

Joined Jul 9, 2011
2,942
Is the single voice command something a human would recognize in that same room? (crowd, others talking, TV or music on, etc)
....

There is no reason (other than time and money) that a single voice command in a room cannot be decoded, especially if the processor already has a sample of that command from that person to match.
I'm doing some home automation stuff. Let's say a single word would not be enough. Just as an example: Turning on and off a light.
1. the first word would be something like "Computer" to draw it's attention
2. Could be just "Lights", so everytime I say it it would toggle

Of course it would be able to recognize the command under all circumstances, loud music, people talking, babies crying ...
Thinking again a human is a multiple input sensory "machine", if one sensor fails we are able to adapt to circumvent the situation.

That is where gesture recognition comes in, still a field of research. And again, I'm able to get some LEDs flashing or program a simple PLL into a PIC but THIS is linked to advanced programming skills, unfortunately.

Then also the more sensors you put into the system and the more reliable it gets less likely it is that you ever can sell any of it. :D

Probably the major obstacle to implement and develop something like it IS time AND money. The one or two hours you have per day (provided you are working and have a family) are maybe not enough to educate yourself in one lifetime.
 

BillO

Joined Nov 24, 2008
999
-... you need to work very fast, and with FPGA for demos to get the speed needed.
While I agree with this, the latest cutting edge of FPGA/CPLD will put you back about $25,000 per chip.

I have to agree with the OP. The playing field is no longer level. Big industry wants it that way, so that's the way it's going to be. They scout universities for the next level of technology, buy it up, copyright it, and keep it for themselves.

Of course you can get a hold of it, maybe as you line up to buy the iPhone 5. Or the next 16 core processor from Intel. However, you just simply can't play that game without a couple of billion dollars. Sorry.

All that said, these big boys tend to develop tunnel vision. After a short while they drink so much of their own cool-aid that they can no longer put things into a reasonable world perspective. Instead, they try to mould the world to their view.

So, where do the opportunities exist?

Where they always have. A good way outside the mainstream. Don't be afraid to do something different. Will it succeed? Maybe, but most likely not. Just keep trying. Will you get support? Heck no! Since you are doing something no one has any experience with you'll be considered a nut. At real lunatic, and everyone will tell you you're out of your mind. Chances are, they'll be right but that will be a clue that you're on the right track. The truly creative mind is by definition not well grounded.

Anyone know who Chester Carlson was (no google!)?

He had an idea. Lost his wife and family, lost his friends and job. However, he didn't loose sight of his ridiculous idea and spent many years working on it until he got it right. Now I'll bet everyone reading this has used his invention. He ended up a multi, multi millionaire and landed a gorgeous new young wife (go figure).
 

Thread Starter

praondevou

Joined Jul 9, 2011
2,942

Anyone know who Chester Carlson was (no google!)?

He had an idea. Lost his wife and family, lost his friends and job. However, he didn't loose sight of his ridiculous idea and spent many years working on it until he got it right. Now I'll bet everyone reading this has used his invention. He ended up a multi, multi millionaire and landed a gorgeous new young wife (go figure).
Thanks for that.

However , I think I would like to stay with my family...

And be a multi-millionaire...:D
 

thatoneguy

Joined Feb 19, 2009
6,359
I'm doing some home automation stuff. Let's say a single word would not be enough. Just as an example: Turning on and off a light.
1. the first word would be something like "Computer" to draw it's attention
2. Could be just "Lights", so everytime I say it it would toggle

Of course it would be able to recognize the command under all circumstances, loud music, people talking, babies crying ...
Thinking again a human is a multiple input sensory "machine", if one sensor fails we are able to adapt to circumvent the situation.
Think about the word "Lights".

A FFT or spectrogram of the word to show the frequencies of that word in a certain order and timing. Think of it as a digital filter that blocks anything that doesn't match the "fingerprint" of lights it has, within a certain tolerance.

dsPICs can do this, and there is open source code around for this application. No, it isn't easy, but it can be done at the hobby level now on relatively cheap hardware.
 

Thread Starter

praondevou

Joined Jul 9, 2011
2,942
Will it be able to distinguish between other words ending on "ights" ? Fights, nights, rights...
Maybe you are right but the differences seem so subtile to me.

On the other hand, the probability that someone says "computer" then "knights" is quite low...:D

The ambient noise level maybe another issue. The distance from the microphone too. Actually I would need to make sure I always speak louder than a possible noise source, like music, but that's also true for a human listener...
Then there is the fact that when speaking louder the FFT spectrum would also change, wouldn't it? So what I said wouldn't match the stored pattern anymore...

Well, I did not actually tried it, so how can I say it's TOO complicated? I'll have a look into it. One day.:)
 

ErnieM

Joined Apr 24, 2011
8,377
I'm doing some home automation stuff. Let's say a single word would not be enough. Just as an example: Turning on and off a light.
1. the first word would be something like "Computer" to draw it's attention
2. Could be just "Lights", so everytime I say it it would toggle

Of course it would be able to recognize the command under all circumstances, loud music, people talking, babies crying ...
Thinking again a human is a multiple input sensory "machine", if one sensor fails we are able to adapt to circumvent the situation.

That is where gesture recognition comes in, still a field of research. And again, I'm able to get some LEDs flashing or program a simple PLL into a PIC but THIS is linked to advanced programming skills, unfortunately.

Then also the more sensors you put into the system and the more reliable it gets less likely it is that you ever can sell any of it. :D

Probably the major obstacle to implement and develop something like it IS time AND money. The one or two hours you have per day (provided you are working and have a family) are maybe not enough to educate yourself in one lifetime.
I already have one of those in my living room in the form of my XBOX Kinects. It responde to voice commands so you don't need a manual controller to make adjustments.

Works rather well when watching Netflix.
 

Thread Starter

praondevou

Joined Jul 9, 2011
2,942
I already have one of those in my living room in the form of my XBOX Kinects.
Yeah this technology is incredible.... Since I don't have a XBox or PS3 or Wii I cannot test it however. I always wanted to know how good the interaction actually is (gesture recognition...)
 

strantor

Joined Oct 3, 2010
6,782
The world needs high power, high efficiency brushless tractions motors & controllers. Try to find a brushless motor & controller big enough for an EV; not there. You already have experience in the field.
 

Thread Starter

praondevou

Joined Jul 9, 2011
2,942
The world needs high power, high efficiency brushless tractions motors & controllers. Try to find a brushless motor & controller big enough for an EV; not there. You already have experience in the field.
MOTORs ! Seems to be your favourite field of research :)

I just watched a video of an inventor that build a high power generator supposedly able to lower your energy consumption by 90%...

Looked like a thing we do not mention (here) ;)
 

strantor

Joined Oct 3, 2010
6,782
MOTORs ! Seems to be your favourite field of research :)
it is! And I am "working" on such a brushless controller myself. The problem is balance of time/money. Funding your own R&D costs a lot of money, which requires you to work a lot, which takes away your time for R&D. so basically I have found the perfect balance, which gives me no time OR money for it. I have not made a single step in any direction for > 1 month. I'm actively forgetting all that I researched, and I have a few hundred dollars worth parts sitting around that aren't any good without other parts that I can't afford, and you see where this is going.
I just watched a video of an inventor that build a high power generator supposedly able to lower your energy consumption by 90%...

Looked like a thing we do not mention (here) ;)
Oh yeah, I have one of those. built it from plans on the internet. I don't pay any money to "the man".:p
 
Top