Methodology/Algorithm for Filtering False Alarms

[kovach]

My first post, I've done some searching, so I apologize if this has been addressed, but I couldn't find it.

Is there some common methodology or algorithm used as a filter to minimize alarms that weren't truly radar/laser guns?

I'm primarily interested in the Escort high-end brands in comparison to the V1. I've read many reviews of them here and elsewhere, and these topics are certainly discussed, but not in great detail.

I know the V1 has limited filtering (Logic, and Advanced-logic), wondering how that compares to Escort's filtering (obviously the GPS capabilities is obviously related but a side issue I'm not really interested in here).

Do these filters only work on X band, or K and/or Ka band too?

[takeshi]

I'm sure it's all highly proprietary and closely guarded information.

[jimbonzzz]

We know the methods they use (see the "Filtering" page in the Articles section of guysoflidar.com). As for the exact algorithms, we can speculate and figure out bits and pieces for different detectors, but will probably never have a 100% picture of how it is done for any certain detector.

Since the signals from motion sensors are, for all practical purposes, electronically identical to police radar, ANY method of filtering will have a negative impact on performance. When selecting a radar detector, one must evaluate the impact of such methods, and then decide how much protection they are willing to trade for a quieter drive.

[kovach]

Thanks, that's a good article!

Can we make any general assumptions of which of those methods mentioned are used by Escort and V1?

[jdong]

Thanks, that's a good article!

Can we make any general assumptions of which of those methods mentioned are used by Escort and V1?

Both for sure use

(1) Pulse guarding: Radar blips that do not last a given amount of time (200ms to 800ms depending on frequency, brand of radar detector, revision number, and POP setting) are rejected

(2) Harmonic searching: Ka band is checked for a known oscillator frequency at 1/3 or 2/3 the present Ka frequency.

(3) Obvious on-the-box/in-the-manual features such as reduced X/K band sensitivity modes ("City"), GPS, etc.

In addition the V1 for sure uses, but I'm unsure if Escorts use:

(1) More aggressive harmonic searching: The J-out takes harmonic searching to the next level, constantly checking and remembering locations of harmonics and even retracting previously issued alerts if they are later determined to be a false. The V1 has been demonstrated here in videos to reject more Cobras in POP mode compared to the Belscort M4 platforms (the STI apparently does great in this regard too)

(2) FM-CW analysis: Signals that constantly vary frequency even slightly are rejected.

(3) Statistical analysis in Logic modes: Muting is dependent on the number and direction of alerts.


In addition, Belscorts are theorized to use:

(1) Statistical analysis: AutoScan mode seems to take into account frequency-clustering and relative strengths of X/K band alerts to cancel out door openers.

(2) Patent shotgun: Ooooh check out all the long numbers on the box..... ;-)

(3) Off-axis attenuation: The M4 horn has been optimized to respond primarily to on-axis threats in an attempt to ignore door openers which tend to be off-axis encounters.


Whistlers are confirmed to use:

(1) A unique fast+slow sweep plus triple-checking mechanism with changing sensitivity levels for accurate POP and quick-trigger alerts while doing an acceptable job at eliminating Cobra falses.

(2) Possibly on the Pro-78 platform, on-axis preference too, based on user comparisons.

[jdong]

Since the signals from motion sensors are, for all practical purposes, electronically identical to police radar, ANY method of filtering will have a negative impact on performance. When selecting a radar detector, one must evaluate the impact of such methods, and then decide how much protection they are willing to trade for a quieter drive.

This is such a great point that a few of us here have been trying to make, and it can't be stressed enough. Motion sensor signals look identical to police radar. If your detector doesn't alert to motion detectors at a certain location, there is a non-negligible risk it will NOT alert to police radar at the same location on that band! This risk is almost 100% in non-GPS detectors and is still pretty high for GPS-based detectors -- based on the # of frequencies that need to be locked out due to the number of door openers in your average shopping complex, I'd guess this risk is in the order of 10-25% for typical false locations!

[CJR238]

I may point out the filtering only delays an alert for a second or so. Not a big deal on a high end RD.

[jdong]

I may point out the filtering only delays an alert for a second or so. Not a big deal on a high end RD.

It could be a big deal, though, if the signal is blippy to begin with. But I've been told by MEM-TEK that recent Escort products are responding much better to quick-triggered threats.

[CJR238]

It could be a big deal, though, if the signal is blippy to begin with. But I've been told by MEM-TEK that recent Escort products are responding much better to quick-triggered threats.

Very true. But for example my 9500I got a quick blip and alerted me, than almost 2 miles later there the LEO was I/O random cars. This may not work as well with curvy roads but it it has enough detection range to notify me of a LEO actively monitoring speeds.
A LEO is going to want to get a small history of the speeder he's going to pull over, so a quick trigger pull isn't going to happen very often if at all IMO. My assumption is the LEO is going to monitor the speeders speed for at least 2 seconds and any high end RD isn't going to filter that long, or that strong of a signal.