It is currently Tue Dec 10, 2019 4:54 pm



Reply to topic  [ 154 posts ]  Go to page 1, 2, 3, 4  Next
 RF Tracking System 
Author Message
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
I've mentioned in another thread that I've been working on a position tracking system using RF, specifically ultra-wideband radio frequencies, and I've refrained from posting anything more about it until now. I was reluctant to make a topic until I had everything completely working 100%, but after working on this for 2 years straight, I'm running out of steam. Hopefully, making this thread will help motivate me to get that final 10% done.

...

1/21/2015 Edit: I've signed an agreement with a company, The Void, to give them exclusive license to my tracking technology and work with them to further develop it. They will be creating large scale virtual reality gaming centers, and they've been the only company to contact me to not only share the same vision but have the funding and business experience to make it a reality. As per our agreement, they've asked me to remove all of my technical posts and take down my Youtube videos.


Last edited by Krenzo on Wed Jan 21, 2015 4:47 pm, edited 2 times in total.



Wed Oct 17, 2012 9:15 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2220
Location: Menlo Park, CA
Reply with quote
Awesomely cool ! So glad you finally created a thread for this. I think you are right that there is probably no use in trying to get orientation working when IMU sensors can do this so easily and cheaply (and without scaling issues). Your accuracy of 1/10 of an inch...at what range does that extend? Just the 2 or 3 feet you are showing or further out than that? Also have you done any occlusion testing - say for example with a water balloon to see if there are any negative effects? Looks great though, and I am very curious to see what your final conclusions, specifications, and limitations for the system are. It could be a very interesting solution for global character tracking within a large occluded environment - say for example a paint-ball arena with a bunch of styrofoam props and obstacles. Oh and I love the LEGO idea :D


Wed Oct 17, 2012 9:30 pm
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:35 pm, edited 1 time in total.



Wed Oct 17, 2012 9:55 pm
Profile
Sharp Eyed Eagle!
User avatar

Joined: Tue Feb 21, 2012 11:57 pm
Posts: 429
Location: Irvine, CA
Reply with quote
Very cool, Krenzo! I'm really glad to finally see your system. Are you primarily targeting enthusiasts, or the general RTLS market? I imagine that your hardware is pretty high-end; I didn't even know there were receivers available that are capable of picosecond resolution.


Thu Oct 18, 2012 12:35 pm
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:35 pm, edited 2 times in total.



Thu Oct 18, 2012 1:00 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2220
Location: Menlo Park, CA
Reply with quote
Krenzo wrote:
All of my hardware is custom built using electronic parts that are readily available on sites like Digikey. I think the most expensive individual part I'm using is $10. All of those circuit boards are designed and hand soldered by me except for the FPGA board. I don't think there's actually a receiver with that much resolution you can just go out and buy that's cheaper than a thousand dollars. Designing it all myself allows me to keep costs way down compared to paying for hardware that others have developed and will keep the final system ore affordable than any other comparable system.


Mighty impressive! :shock:


Thu Oct 18, 2012 2:00 pm
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:36 pm, edited 1 time in total.



Thu Oct 18, 2012 2:40 pm
Profile
Sharp Eyed Eagle!
User avatar

Joined: Tue Feb 21, 2012 11:57 pm
Posts: 429
Location: Irvine, CA
Reply with quote
Wow Krenzo, color me impressed!


Thu Oct 18, 2012 3:05 pm
Profile
Binocular Vision CONFIRMED!
User avatar

Joined: Fri Jan 27, 2012 11:24 am
Posts: 228
Reply with quote
Krenzo, this is brilliant. I am going to try something similar, hopefully it will be of some help to you as well.

I presume when you add more transmitters, you are going to vary the frequencies that they emit at. Are you then going to have the receivers set at different receiving frequencies? How much do you have to vary the frequency before you can cut out interference between them?

This is really cheap - http://www.ebay.co.uk/itm/433Mhz-wirele ... 1100755229 - does it look any good to you? Do you need to get a special omni-directional aerial?

If you have any kit to recommend that would be awesome. Preferably cheap and Arduino compatible.

Keep up the good work.


Mon Oct 22, 2012 12:32 pm
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:36 pm, edited 3 times in total.



Mon Oct 22, 2012 1:06 pm
Profile
Binocular Vision CONFIRMED!
User avatar

Joined: Fri Jan 27, 2012 11:24 am
Posts: 228
Reply with quote
Oh thanks, I see now how your one works. I wasn't thinking of doing a time of flight thing I was just thinking of measuring the amplitude of the emitter and then using inverse square law or whatever to estimate the distance to it. I thought that if the wavelength was long enough to avoid any disturbances with objects, and the emitters was constant then you might be able to get a sensor accurate enough to make that work. What do you think? I haven't worked with RF before so I don't really know about how accurate different pieces of kit are.


Mon Oct 22, 2012 1:14 pm
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:36 pm, edited 1 time in total.



Mon Oct 22, 2012 1:26 pm
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
:o :o :o

This is insanely impressive! Working in those time scales sounds like a huge challenge. How do you keep the transmitter and receiver synchronised, or is it measuring the entire phase delay between transmission and reception?

Keep up the awesome work! I definitely want to see more of this! :D


Mon Oct 22, 2012 1:34 pm
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:37 pm, edited 1 time in total.



Mon Oct 22, 2012 1:48 pm
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
Ok. I think I get it now. I was neglecting to keep in mind that the transmitters were spaced apart from each other. Are all the transmitters operating at the same time, but at different frequencies?


Mon Oct 22, 2012 2:18 pm
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:37 pm, edited 1 time in total.



Mon Oct 22, 2012 2:47 pm
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
Krenzo wrote:
MSat wrote:
Ok. I think I get it now. I was neglecting to keep in mind that the transmitters were spaced apart from each other. Are all the transmitters operating at the same time, but at different frequencies?


I answered that in my reply to Chriky: "I am not varying the frequency. Each transmitter has a specific time window it is allowed to transmit, and that prevents transmitters from interfering with each other. Ultrawideband radio signals occupy a frequency range of at least 500 MHz. They occupy a lot of frequencies at once unlike narrowband signals which makes it difficult to prevent multiple signals from interfering."

It's advantageous to use as much bandwidth in a signal as possible which makes the signal more resistant to reflections. The UWB signal is very short, only a nanosecond or two. You can easily send a million samples in one second which gives plenty of time to divide up among lots of transmitters and have sufficient transmissions from each one.



Doh! Total brain fart. :oops:

Also, thanks for the additional explanation. It's really interesting that you're getting such high precision. I would have never imagined that such a system would be feasible, especially at such a low price.

Would you be able to determine position/orientation/rotation using one transmitter, but something like a 'spherical array' of unidirectional receivers?


Mon Oct 22, 2012 4:44 pm
Profile
Sharp Eyed Eagle!
User avatar

Joined: Tue Feb 21, 2012 11:57 pm
Posts: 429
Location: Irvine, CA
Reply with quote
MSat wrote:
Also, thanks for the additional explanation. It's really interesting that you're getting such high precision. I would have never imagined that such a system would be feasible, especially at such a low price.


It's only inexpensive if you don't factor in Krenzo's engineering time! But I agree, it's very cool.

MSat wrote:
Would you be able to determine position/orientation/rotation using one transmitter, but something like a 'spherical array' of unidirectional receivers?


I don't think this would work. If you have a single transmitter, then you can determine your angle of deflection to the transmitter, but that doesn't tell you which side of the transmitter you are on.


Mon Oct 22, 2012 9:02 pm
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
FingerFlinger wrote:

It's only inexpensive if you don't factor in Krenzo's engineering time! But I agree, it's very cool.


Of course. But still, a project like this would probably have a sizeable team behind it, so there would be a bigger investment to recoup.

FingerFlinger wrote:

I don't think this would work. If you have a single transmitter, then you can determine your angle of deflection to the transmitter, but that doesn't tell you which side of the transmitter you are on.


You're totally right. I was trying to imagine a different implementation, and tried to simplify it with only one transmitter, which I realize now won't work.

Well, perhaps it might work if you use two fixed receivers and a transmitter on the same plane, but spaced apart (the receivers should probably be placed at the far ends of the user space). The main transmitter sends a "ping", or a burst of some given frequency. Given that the space between the primary transmitter and fixed receivers is known, the exact moment of rf propagation can be determined. The user wears a receiver/transmitter that simply repeats the RF pulse train after a given, known time. Depending where the user is will dictate the time it takes for the repeated signal to reach each fixed receiver, and therefore giving the ability to calculate position. If you have multiple directional antennas on the user's receiver for orientation tracking, perhaps you can transmit related data through a different wireless protocol, or somehow mux it into the repeated signal perhaps modulating the frequency, or burst duration (not sure how feasible that is).


Wed Oct 24, 2012 6:16 pm
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:37 pm, edited 1 time in total.



Mon Nov 05, 2012 9:27 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2220
Location: Menlo Park, CA
Reply with quote
@Krenzo: Questions for the RF guru around here. In another thread a guy was talking about an RF triangulation system that could achieve sub-meter accuracy. Typically you hear about WiFi triangulation with an indoor accuracy of a few meters and outdoor accuracy of 10's of meters. What kind of equipment do you need to achieve high accuracy (sub-meter) triangulation? It doesn't seem like accurate timing would be important since it's all based on strength measurements. Why is difficult to accurately measure signal strength. And a related question - why is outdoor accuracy worse for triangulation? It seems like without any multipath issues it would be better.

One final question, is doppler shift measurement possible at these ranges and speeds?


Tue Nov 06, 2012 8:48 pm
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:37 pm, edited 1 time in total.



Tue Nov 06, 2012 10:34 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2220
Location: Menlo Park, CA
Reply with quote
I read it somewhere that outdoor accuracy was less. It didn't make sense to me and I'm glad it doesn't make sense to you either. I'm not thinking in terms of highly accurate positioning. More of a companion / correction signal for a dead reckoning system. So meter accuracy would be adequate. However I also just found a stat that said that WiFi triangulation error (with affordable equipment) is something like 15% of the distance between transmitters so that could get pretty bad.


Tue Nov 06, 2012 11:19 pm
Profile
One Eyed Hopeful

Joined: Sat May 21, 2011 4:12 pm
Posts: 24
Location: Minnesota
Reply with quote
Great Job! I've worked on something similar using ultrasonic ranging but nothing in the RF spectrum. How are you comparing the phase difference between the Tx and Rx signals? comparator tied to a timer? Do all your Rx waveforms look as good as the Tx?

I want to show you this: www.mdpi.com/1424-8220/9/12/10190/pdf

They are using two rf transreceivers to measure distance. The idea is that one of the two is a slave repeater with a "constant" repeater delay. The transmitter is programmed to send a packet as soon as it hears from the previous one sent. It is essentially a "ping - pong" range finder. each pong will increment a counter and you can check your counter at a much slower pace and thus back calculate the distance.

The main advantage I see is that it doesn't need synchronous clocks and might be feasible on cheaper hardware. You might be able to use a similar scheme with your UWB ToF system.


Sun Dec 02, 2012 11:03 pm
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
mars3554 wrote:
I want to show you this: http://www.mdpi.com/1424-8220/9/12/10190/pdf

They are using two rf transreceivers to measure distance. The idea is that one of the two is a slave repeater with a "constant" repeater delay. The transmitter is programmed to send a packet as soon as it hears from the previous one sent. It is essentially a "ping - pong" range finder. each pong will increment a counter and you can check your counter at a much slower pace and thus back calculate the distance.

The main advantage I see is that it doesn't need synchronous clocks and might be feasible on cheaper hardware. You might be able to use a similar scheme with your UWB ToF system.



That is in a sense what I proposed, although instead using one transmitter, the trackable repeater with fixed response time, and two fixed receivers. Such a setup should allow for absolute positioning, and also negate the need for a common clock. The transmitter is used to both instigate a response ping from the repeater, and also as a reference for the two receivers to measure the delay of the repeater. With the delay measurement from each receiver, the location of the repeater could be determined in 2D.


Mon Dec 03, 2012 12:02 am
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:38 pm, edited 1 time in total.



Mon Dec 03, 2012 1:43 am
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:38 pm, edited 1 time in total.



Mon Dec 03, 2012 2:11 am
Profile
Sharp Eyed Eagle!

Joined: Sat Dec 22, 2007 3:38 am
Posts: 425
Reply with quote
I've seen some system that instead of measuring input power and translating that to distance, instead measure output power, and use closed-loop feedback to maintain input signal intensity (i.e. as the Tx and Rx separation increases, the output power must be increased to produce the same input amplitude). Downside: implementing the closed loop and maybe a latency of one or two times the path length if you're moving exceedingly fast, and your accuracy is partially limited by the control you have over your amplifier (though you can measure both the output power and input power and use one as an offset of the other to take advantage of receiver sensitivity), but the advantage is you get a really excellent range as the receiver never saturates and only loses signal when your amplifier can no longer output enough power, and consistent accuracy over that range.


Mon Dec 03, 2012 10:07 am
Profile
Petrif-Eyed
User avatar

Joined: Sat Jan 09, 2010 2:06 pm
Posts: 2255
Location: Perpignan, France
Reply with quote
No update since December, are you still working on this Krenzo ? It looks very interesting, just happened to read some things about UWB and found this thread, nice. :)

I wondered what would be the overall cost of such a setup and if a developer with some basic soldering skills could build it. What do you think ?

Here is a video that could interest Brantlew for outdoor positioning if he didn't see it already :



And here is an interesting article and PDF about a DIY UWB radar (is it similar to your setup ?) :
http://hackaday.com/2007/10/29/diy-ultr ... and-radar/


Fri Mar 15, 2013 7:13 pm
Profile WWW
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2220
Location: Menlo Park, CA
Reply with quote
That is a cool video Fredz. The accuracy looks very good for character position but the range is pretty short. I wonder how much the range could be boosted? It would be great to have other technologies to do an indoor version of Red Rovr with. Radio is one that I have thought about but not sure of all the issues with reflection and range and is one of the reasons I was so curious about Krenzo's work. Another one that I have thought about is magnetic like this company does.

http://www.technologyreview.com/news/428494/startup-uses-a-smartphone-compass-to-track-people-indoors/


Fri Mar 15, 2013 7:35 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Jan 09, 2010 2:06 pm
Posts: 2255
Location: Perpignan, France
Reply with quote
Interesting idea, I just tested with my phone and the results seem predictable, I didn't really expect that it would work. :shock:

I've put the phone in 6 distinct locations on a grid on the floor with a 30cm distance between the points and I've looked at the x, y, z and the magnetic field (in µT). Shockingly the values are quite repeatable for the 6 positions, I'm impressed. I guess the fact that there is a lot of metal in my house must somewhat help.

Now I need to implement something on the phone to verify that it does really work, by showing the predicted position depending on the sensor measure. Damn, 30th project in my list, I really need more free time. Or I should stop sleeping altogether...


Last edited by Fredz on Fri Mar 15, 2013 8:23 pm, edited 2 times in total.



Fri Mar 15, 2013 8:06 pm
Profile WWW
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2220
Location: Menlo Park, CA
Reply with quote
Yeah, I think it could be a really interesting solution and you could even artificially "salt" the indoor environment by embedding random metal pieces in the concrete foundation. I consider it a strong candidate for a drop-in GPS replacement for low accuracy indoor positional tracking.


Fri Mar 15, 2013 8:13 pm
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
Wow! That seems like it would be insanely tricky to implement for so many different reasons. If it could work consistently (including across devices) then that's pretty amazing, but I have to admit I have my doubts.


Fri Mar 15, 2013 9:00 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2220
Location: Menlo Park, CA
Reply with quote
MSat wrote:
Wow! That seems like it would be insanely tricky to implement for so many different reasons. If it could work consistently (including across devices) then that's pretty amazing, but I have to admit I have my doubts.


But not too tricky if your goal is something like 1 meter accuracy with a 250ms latency tolerance which is fine for the application I had in mind.


Fri Mar 15, 2013 9:32 pm
Profile
3D Angel Eyes (Moderator)
User avatar

Joined: Sat Apr 12, 2008 8:18 pm
Posts: 11394
Reply with quote
This sounds promising, but I'm not sure it's accurate enough for my needs (for example, to track a glove).

Though I do wonder if it could be combined with other sensors to get something usable.

_________________
check my blog - cybereality.com


Fri Mar 15, 2013 9:38 pm
Profile
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:38 pm, edited 1 time in total.



Fri Mar 15, 2013 9:48 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Jan 09, 2010 2:06 pm
Posts: 2255
Location: Perpignan, France
Reply with quote
Krenzo wrote:
Yes, I'm still working on it. I made significant progress on it and have it working in real-time. I'm working on getting a video made of it in operation in Unity and UDK.
Great ! Do you intend to make a product of this in the end or are you just tinkering with it for fun ?


Sat Mar 16, 2013 6:43 am
Profile WWW
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:38 pm, edited 1 time in total.



Sat Mar 16, 2013 1:38 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Jan 09, 2010 2:06 pm
Posts: 2255
Location: Perpignan, France
Reply with quote
Krenzo wrote:
Yes, I hope to make a product of it. I don't know if I could target consumers with it though.
Nice, do you have an idea of the price point yet ?

Krenzo wrote:
I could target what I think is an attractive price point for my system alone, but then you have to consider you would need a Rift + my system + a backtop or wireless video system and the costs start adding up.
You're talking about a large-scale outdoor version it seems. I guess for a seated or standing use in a small room it wouldn't need the backtop or wireless video system and the cost could be cut down, right ?

Krenzo wrote:
Plus, do people have enough space in their homes to really make use of a large scale tracking system?
Probably not, but I guess a small-scale version for seated/standing use could be very appealing for the average consumer. If you can build it at this scale and make it precise enough tough, with millimeter accuracy if it's even possible.

I'd be personnaly very interested in such a thing, but at a sensible price since it would only add positional 3dof to the already existing rotational 3dof.


Sat Mar 16, 2013 6:17 pm
Profile WWW
Binocular Vision CONFIRMED!

Joined: Tue Sep 07, 2010 10:46 pm
Posts: 265
Reply with quote
...


Last edited by Krenzo on Wed Jan 21, 2015 4:38 pm, edited 1 time in total.



Sat Mar 16, 2013 6:49 pm
Profile
Display posts from previous:  Sort by  
Reply to topic   [ 154 posts ]  Go to page 1, 2, 3, 4  Next

Who is online

Users browsing this forum: Google [Bot] and 3 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
Powered by phpBB® Forum Software © phpBB Group
Designed by STSoftware.