[Patent] High Probability of Intercept Radar Detector (part 2)

DrHow

Going “Plaid” ASAP (Tesla S) RDT refugee
Advanced User
Lifetime Premium Member
Joined
May 18, 2018
Messages
2,548
Reaction score
4,043
Well one good thing about the current Patents, it shows thinking outside the current V1 box. Plus it shows components that have not been in current V1s IE LNAs for example. So I agree this isn't Radenso Theia type stuff, be it does look like a better V1 with additionally capabilities and if there is some type of TMF3 to combat BSM all the better.

Again, does a souped up V1 for $500 fit ones needs better then a $1200 Theia (guessing at a high price just as an example of why an improved V1 might be better choice for somebody then a maxed out Theia). In other words, whether an improved V1 comes out of not, or whether Theia sweeps the high end RD market: two things are still going to happen! Walmart is still going to be selling its $29 RDs and VR is still going to be selling $500 V1s.

What will be in my vehicle, most likely Theia if even only delivers half of what I think it can. I need the protection for how much and how far I drive with the limited time I have.
Fun alert, at limited expense to another: Well, it could be these patents will be applied on a surface mounted chip integrated into the new volume/on-off/mute/Full logic control. That explains the delay in the new volume control causing production to stop. :)


serious, agree with @cihkal on this one. “New and improved tide, new color box and packets with oxi cleaning”. In other words, incremental change to a dying design?
Post automatically merged:

Heh. I really do hope we see something soon from VR.

Personally I'll be sitting out the initial launch of Theia until it's had plenty of testing time in the real world by RDF members (and not just range!). On paper it looks good in many ways, but the devil is always in the details, and so far no one has managed to deliver the perfect RD yet. But I still keep returning to my V1...
V2 “two weeks“ is shorter than Radenso actual GA. People will buy V2 or whatever it is called. It will be real, and working Long before “Theory” comes out. Don’t get me wrong, love the Rai and the work to date. Mikey gets to taste blood first. No matter what prevent offense Radenso did by showing what is coming a ways off to freeze purchases.
 
Last edited:

CobawLT2010

Doesn’t time fly when you’re waiting two weeks?!
Advanced User
Premium Member
Joined
Oct 26, 2011
Messages
7,508
Reaction score
3,690
Location
Sick of here...
I actually know I am not wrong about this... you just need to follow the sweep of known false algorithms once the main signal is detected since most falses are wider/sloppier or have multiple front horns or patch receivers scanning at the same time looking at how wide the signal is. You don’t need AI per say... it would likely have more latency if you didn’t have multiple receivers.

there’s always more than one way to skin a cat...
 
Last edited:

woopies

Lurking
Intermediate User
Joined
Jul 25, 2014
Messages
473
Reaction score
704
Location
10280 Alliance Road, Cincinnati, OH
I actually know I am not wrong about this... you just need to follow the sweep of known false algorithms once the main signal is detected since most falses are wider/sloppier or have multiple front horns or patch receivers scanning at the same time looking at how wide the signal is. You don’t need AI per say... it would likely have more latency if you didn’t have multiple receivers.

there’s always more than one way to skin a cat...
You have the right idea; however, you don't need multiple antennas--just a wide bandwidth receiver (how much it can see at once)

How to implement
1. Sweep whole spectrum
2. Keep history (What VR calls virtual image)
3. Repeat
4. Maintain MxN matrix of virtual images
5. Feed MxN matrix into some AI
3. Profit?

Also, machine learning is really the process of finding weights (coefficients).

Ugh, I wish I had my matlab files.....Anyways, we implemented this algorithm to filter out line noise. The idea was that you feed some known signal into a noisy line, tell the system what it should be, then let the decision maker (error) determine what the actual values should be.

Note: The algorithm above is used in active noise canceling.
 

Tallyho

Advanced User
Premium Member
Joined
Mar 21, 2013
Messages
3,101
Reaction score
5,671
You have the right idea; however, you don't need multiple antennas--just a wide bandwidth receiver (how much it can see at once)

How to implement
1. Sweep whole spectrum
2. Keep history (What VR calls virtual image)
3. Repeat
4. Maintain MxN matrix of virtual images
5. Feed MxN matrix into some AI
3. Profit?

Also, machine learning is really the process of finding weights (coefficients).

Ugh, I wish I had my matlab files.....Anyways, we implemented this algorithm to filter out line noise. The idea was that you feed some known signal into a noisy line, tell the system what it should be, then let the decision maker (error) determine what the actual values should be.

Note: The algorithm above is used in active noise canceling.
But isn't this still fundamentally a completely different process which makes it a frequency detector rather than a signal identifier as offered by AI?

As I understand it they are two completely different, mutually exclusive approaches to solving the same problem. Is this not the case?
 

woopies

Lurking
Intermediate User
Joined
Jul 25, 2014
Messages
473
Reaction score
704
Location
10280 Alliance Road, Cincinnati, OH
But isn't this still fundamentally a completely different process which makes it a frequency detector rather than a signal identifier as offered by AI?

As I understand it they are two completely different, mutually exclusive approaches to solving the same problem. Is this not the case?
Hm, I don't know if I'm reading this right but the signal classifier that Radenso brings still uses the same system/architecture as every other receiver platform. It's just that Radenso applied an AI to figure out what that signal is.
 

Heywood

Learning Something New, Still Dying Stupid
Advanced User
Lifetime Premium Member
Joined
May 12, 2016
Messages
3,753
Reaction score
6,868
Location
Alberta Canada
But isn't this still fundamentally a completely different process which makes it a frequency detector rather than a signal identifier as offered by AI?

As I understand it they are two completely different, mutually exclusive approaches to solving the same problem. Is this not the case?
And if it was easy to do, why hasn’t someone done it?
I’m sure there’s ..... sorry, I know there’s smarter people who’ve probably thought of this and decided not too.

I’m sure that if you took the current detectors to anyone in the that deals with signal/frequency identification for the military, they’d probably laugh their asses off how conventional detectors work. Oh wait, I think that’s been done and patented now

@Tallyho
Rai looks at the signal like we look at dogs and cats once trained. They have similar characteristics like paws, fur, 2 eyes and a tail, but when you look at any dog .... you know it’s a dog.

That’s really what AI is doing. It’s seen lots of dogs and cats in training. It’s not really learning anything new. It’s using its onboard intelligence to know it’s a dog.

When it comes to dogs, it goes one step further. Radar guns are purebreds. Just like we can tell a German Shepard from a French poodle or a Doberman vs a Rotty.

Why?
Because it, like us, have seen lots of German Shepard’s. Doesn’t matter the color of the coat. The length of the hair. It’s the stance, the ears, the snout.

That’s what AI is doing on Theia. Looking for dogs in a field with cats, mice, gophers, and rabbits. All share common traits, but easily identified once taught what they are. It figures it out from there.
 

Tallyho

Advanced User
Premium Member
Joined
Mar 21, 2013
Messages
3,101
Reaction score
5,671
Hm, I don't know if I'm reading this right but the signal classifier that Radenso brings still uses the same system/architecture as every other receiver platform. It's just that Radenso applied an AI to figure out what that signal is.
I'm trying to understand how that can be the case when we are discussing systems that rely on different approaches. For example, mikedotd discusses how conventional RD's are based on the "fail-open" approach, whereas the AI/NN approach uses the exact opposite fail-closed approach:

All conventional RD's have detected a signal inside the scanned frequency range, processed it to filter out false alerts(time, other detected frequency harmonics), then alerted. This design allows for false positives, but will always alert to a true positive (fail open if you will). Radeno's new design appears to be signature based, which means it will only alert to true positives that match a known signature, but at the expense of being blind to unknown true positives (fail closed). That could eventually be a drawback if new police radar signatures exist that Radenso has not encountered, if customers haven't kept their RD up to date, or if there are real world conditions that influence the signature that the RD sees, preventing the AI from reaching a match to a true positive.
Then fogboundturtle identifies this approach and again illustrates the downside of this form of whitelisting by a neural network where it would not alert to an unknown threat:

Let’s talk about potential pitfall using this technology: if a signal signature has not been learned by the neural network, it will not alert therefore could potentially put at risk the speeding drivers So it becomes paramount for Radenso to update their neural network deconvolution frequently as new equipment get released.
How can they take completely different approaches that require different processes and technologies and be considered tantamount to one another?
 

woopies

Lurking
Intermediate User
Joined
Jul 25, 2014
Messages
473
Reaction score
704
Location
10280 Alliance Road, Cincinnati, OH
I'm trying to understand how that can be the case when we are discussing systems that rely on different approaches. For example, mikedotd discusses how conventional RD's are based on the "fail-open" approach, whereas the AI/NN approach uses the exact opposite fail-closed approach:



Then fogboundturtle identifies this approach and again illustrates the downside of this form of whitelisting by a neural network where it would not alert to an unknown threat:



How can they take completely different approaches that require different processes and technologies and be considered tantamount to one another?
The only thing that differs is the processing of the information. Both still have an analog input, some band pass, some amplifier. SDR just pushes everything to software. Therefore, everything you don't do in hardware is done is software.
 

Jon at Radenso

Manufacturer
Advanced User
Premium Member
Manufacturer
Joined
Oct 9, 2012
Messages
1,407
Reaction score
9,700
Age
31
Hm, I don't know if I'm reading this right but the signal classifier that Radenso brings still uses the same system/architecture as every other receiver platform. It's just that Radenso applied an AI to figure out what that signal is.
Nope, that is not the case.

1. Theia is the only radar detector in the world that uses a complex quadrature front end (it can see phase)
2. It does not use a traditional dual or triple mixing stage superheterodyne design
3. The instantaneous analysis bandwidth is changeable on the fly, and at minimum will be 400% more than any other detector
4. It can sweep multiple times through roughly 3.5ghz of bandwidth before 16ms pop is finished shooting

Theia's system architecture is basically completely different than any other radar detector, even without AI.
 

GTO_04

Advanced User
Premium Member
Joined
May 22, 2011
Messages
4,552
Reaction score
5,269
Nope, that is not the case.

1. Theia is the only radar detector in the world that uses a complex quadrature front end (it can see phase)
2. It does not use a traditional dual or triple mixing stage superheterodyne design
3. The instantaneous analysis bandwidth is changeable on the fly, and at minimum will be 400% more than any other detector
4. It can sweep multiple times through roughly 3.5ghz of bandwidth before 16ms pop is finished shooting

Theia's system architecture is basically completely different than any other radar detector, even without AI.
That is impressive to say the least. Do other RDs with bogey counters (identify multiple signals), use any phase detection at all?

GTO_04
 

woopies

Lurking
Intermediate User
Joined
Jul 25, 2014
Messages
473
Reaction score
704
Location
10280 Alliance Road, Cincinnati, OH
Nope, that is not the case.

1. Theia is the only radar detector in the world that uses a complex quadrature front end (it can see phase)
2. It does not use a traditional dual or triple mixing stage superheterodyne design
3. The instantaneous analysis bandwidth is changeable on the fly, and at minimum will be 400% more than any other detector
4. It can sweep multiple times through roughly 3.5ghz of bandwidth before 16ms pop is finished shooting

Theia's system architecture is basically completely different than any other radar detector, even without AI.
I'm a little confused. You are saying sweep but not heterodyning? Also, you don't need a quadature architecture to see phase.
 

Jon at Radenso

Manufacturer
Advanced User
Premium Member
Manufacturer
Joined
Oct 9, 2012
Messages
1,407
Reaction score
9,700
Age
31
I'm a little confused. You are saying sweep but not heterodyning? Also, you don't need a quadature architecture to see phase.
I was trying to be specific - we still have one mixing stage, but this is not a double or triple conversion superhet like other radar detectors. There are no other radar detectors on the market that can get I/Q data; Theia is a true I/Q receiver (and we've filed a patent on multiple methods of obtaining I/Q data in a radar detector).
 

DrHow

Going “Plaid” ASAP (Tesla S) RDT refugee
Advanced User
Lifetime Premium Member
Joined
May 18, 2018
Messages
2,548
Reaction score
4,043
I'm a little confused. You are saying sweep but not heterodyning? Also, you don't need a quadature architecture to see phase.
love your avatar, description, when combining reading what you claim your address is. Priceless.
 

LouG

PSL +5
Intermediate User
Joined
Jan 2, 2014
Messages
2,777
Reaction score
3,007
And if it was easy to do, why hasn’t someone done it?
I’m sure there’s ..... sorry, I know there’s smarter people who’ve probably thought of this and decided not too.
It's about money, and the potential market for a $1000 plus detector.
It's the bubble syndrome again, we think because we obsessives will buy it, anybody would.
They won't, at least not in enough quantity to make it worth doing.
In addition, there's the societal change that's going on. Kids don't champ at the bit to get a licence, some aren't bothering at all. Down here real speeding is rare compared to 30 years ago. Detector use is also less common. You also see it in the fact that boy racers are more interested in burnouts than setting records from city to city. It was common conversation when I was much younger about how quickly you did point A to point B. Now people look at you like you're a potential murderer.
PS. I'd love to know the profit margin in Radenso's "game changer".
 

nycarpenter

Learning to Fly
Beginner User
Joined
Feb 17, 2014
Messages
264
Reaction score
558
It's about money, and the potential market for a $1000 plus detector.
It's the bubble syndrome again, we think because we obsessives will buy it, anybody would.
They won't, at least not in enough quantity to make it worth doing.
In addition, there's the societal change that's going on. Kids don't champ at the bit to get a licence, some aren't bothering at all. Down here real speeding is rare compared to 30 years ago. Detector use is also less common. You also see it in the fact that boy racers are more interested in burnouts than setting records from city to city. It was common conversation when I was much younger about how quickly you did point A to point B. Now people look at you like you're a potential murderer.
PS. I'd love to know the profit margin in Radenso's "game changer".
The younger generation call an Uber, rent an apartment, go on vacation, spend 2x the money they make then bitch they can't pay the bills. I would not want to try to market a 1000 radar detector to the under 35 crowd anytime soon. Better odds of selling sand to the Arabs. I bet the R7 at $600 will drop in price soon.
 

cihkal

Pure Energy
Advanced User
Premium Member
Joined
Apr 21, 2014
Messages
3,578
Reaction score
6,351
.
 
Last edited:

nycarpenter

Learning to Fly
Beginner User
Joined
Feb 17, 2014
Messages
264
Reaction score
558
Personal Attacks (Mandatory 1-Point Infraction)
I'm 30 and just gotta say, not impressed.

Let's not puke ignorance all over the place.

The whole shit on other generations has been going on forever, and it just makes you and anyone who perpetuates this shit look really stupid.

Grow up dude, this is a RD forum. Go drink your koolaid elsewhere.

"Selling sand to the Arabs"... I see how you play. You ever been to an Arab country? I have, many are very nice people.

Before you get your panties in a bunch, why don't you take a seat and listen for a while. Some of us, despite our age, actually make a difference. Ive designed and managed the construction of systems as the GC/lead engineer, systems that have created American jobs. I actually try to make a difference and many people my age try their best as well. Much better than acting like an ignorant racist.

Hateful fool.
Your an idiot! Sell sand to the Arabs is not a hateful saying it's like sell snow to the snowman fool! Go get a life Jack ass! I've been a in construction for 30 years never have I seen a lazier generation then the youth today! Always broke and looking for a free ride.
Post automatically merged:

 
Last edited:

XDA

PSL +5
Intermediate User
Joined
Oct 22, 2015
Messages
438
Reaction score
706
Location
Texas
I don't know what my upper price limit would be for a next generation super radar detector. If it didn't false, had arrows & bogie counter, outstanding range, and provided data like Theia is supposed to, it would be highly desirable to me. Within reason, of course, money would be no object.
 

Discord Server

Latest threads

Latest posts

Forum statistics

Threads
79,123
Messages
1,204,723
Members
20,088
Latest member
SVRftw
Top