Passive IR traffic advisory system

zaitcev

zaitcev

En-Route
Joined
Sep 30, 2010
Messages
3,257
Display Name
Display name:
Pete Zaitcev
I'm considering taking on a little side project: an air traffic information system that is based on a passive IR interferometer. By locating IR sensors at, say, wingtips and the tail, a distance can be measured and closer objects reported to a display in the cockpit.

Such system is different from TIS-B because it is passive and thus can detect UAVs and aircraft that do not participate in ADS-B for some reason (broken, malicious, legally underequipped).

I am not sure the idea is entirely sane in the sense that I'm trying to detect a fairly small motion against moving background. Worse, for an object on collision course the relative motion really is very small.

Also, there's a question of expense.

Still, I don't see a fatal flaw in the idea, technically.
 
How would your system detect aircraft which lack operating wing/tail-tip IR emitters? Are you going to make those devices mandatory for flight for all aircraft as part of 91.205(b)? Otherwise how are you going to get people to install and maintain them?
 
The envisioned system is passive, as I mentioned. I'm thinking about close to mid IR band, around 3um. A sensor cooled with ambient air ought to be able to discern heat of engine of just about any kind at useful distances. Although, granted, it may be challenging to detect a journalist's drone in time to avoid a collision.

I haven't thought about the business and marketing side of things.
 
OK, I've got it -- you're not trying to determine distance to target by distance between IR emitters on the target, but rather to get 360 degree sensor coverage for the detecting aircraft. It's an interesting idea, although I wonder about cost of such a system as well as the typical issues with IR systems related to atmospheric interference (moisture, etc) and false targets.

And despite the President making privacy issues related to UAV's a Department of Commerce issue, the operation of those things is still DoT/FAA's to regulate, and they are taking the issue of separation from aircraft very, very seriously. I foresee very strict regulations to keep such devices separated from normal aircraft operations, probably with restrictions about operating near an airport or above some altitude like 200-400 feet AGL without coordination with ATC or the like.
 
I probably should mention that I'm superficially familar with military IR systems used for missile warning and targeting. Even excluding those mounted in warheads, such systems are typically built around a single expensive sensor, either air-cooled or cooled with liquid nitrogen or other expendable coolant. They never use stereo vision to detect objects of interest for some reason (with possible exception of built-in SA system of F-35, which may be the closest equivalent of what I want, if I knew anything about its characteristics). The question is, how difficult is the problem of stereoscopic detection? I don't think even Page and Brin's private 767 is going to carry a supercomputer onboard if it turns out necessary for the task.

As it happens, my local university has a nationally recignized expert in machine vision, so I'm going to ask his opinion on this.
 
zaitcev said:
I probably should mention that I'm superficially familar with military IR systems used for missile warning and targeting. Even excluding those mounted in warheads, such systems are typically built around a single expensive sensor, either air-cooled or cooled with liquid nitrogen or other expendable coolant. They never use stereo vision to detect objects of interest for some reason (with possible exception of built-in SA system of F-35, which may be the closest equivalent of what I want, if I knew anything about its characteristics). The question is, how difficult is the problem of stereoscopic detection? I don't think even Page and Brin's private 767 is going to carry a supercomputer onboard if it turns out necessary for the task.

As it happens, my local university has a nationally recignized expert in machine vision, so I'm going to ask his opinion on this.
Click to expand...

Stereoscopic vision is not a difficult math problem when you are just measuring an intensity differential and offset. I'd think that issues with sensor resolution would be the biggest issue.

But, I've only measured things a foot away and down to the hundredth of an inch, so your problem is a little different =)
 
One beer in...

Does it have to be totally passive? Could you combine an IR emitter (LED based, probably), that would "ping" the environment like a flashbulb. Flash every second or so, modulate it so you can detect your own pulse, look only during the expected return interval. You might even be able to get a crude distance from it.
 
I flew in an airplane that could "see" other aircraft, and where they were. And the other aircraft just needed a mode C transponder, not a mode S. It was a Diamond DA-40 I believe.

Isn't this sort of what the ATC system is headed to? Everyone knows where everyone else is and everyone is put somewhere where they wont run into anyone? And ATC sees it all, but you don't have to talk to ATC, you get your clearance in text and fly that.

The UAV's participate. They get their clearance and their equipment flies their route automatically, or with their "pilot's" assistance if necessary.

Sort of like one big video game.
 
I had thoughts regarding use of a binocular system based on cheap cameras using visible spectrum for use in VFR conditions. I'm not at my home computer where I saved the PDFs of some of the research papers I found that described the possible software solutions that take the two video inputs and back-derive the 3D coordinates of the objects that generated the images on the two cameras. I think this is a solved problem in robotic vision systems, but I could be wrong. The research papers I skimmed weren't promising, but I vaguely recall they weren't very current either.

If the UAV or other aircraft that you want to be aware of make enough noise, one might be able to use microphones instead. Range would be reduced but might be easier to perform the interferometry (or other complex processing) in real time using software on something like a RaspberryPi.
 
You are still going to miss balloons, sailplanes, hang and paragliders, skydivers in free fall and under canopy, toy RC stuff, birds, and the worst offender mountains. Instead of trying to get everything out there on the TV inside the plane, look away from the TV and outside at the pretty world.
 
coloradobluesky said:
I flew in an airplane that could "see" other aircraft, and where they were. And the other aircraft just needed a mode C transponder, not a mode S. It was a Diamond DA-40 I believe.

Isn't this sort of what the ATC system is headed to? Everyone knows where everyone else is and everyone is put somewhere where they wont run into anyone? And ATC sees it all, but you don't have to talk to ATC, you get your clearance in text and fly that.

The UAV's participate. They get their clearance and their equipment flies their route automatically, or with their "pilot's" assistance if necessary.

Sort of like one big video game.
Click to expand...


Sounds like a TCAS system. It interrogates surrounding transponders directly and receives a response on 1090Mhz. Which is why eventually TCAS is reported to be able to receive ADS-B traffic with upgrade.

It is range limited as compared to ADS-B because it is direct point to point.
 
How do you get it to filter out ground clutter?

Of course, it could always be look-up only, but what good is that?
 
This has been considered before. In fact it was an Aprils fools joke on avweb maybe 2000/2001; I remember because I fell for it and they had a little fun with me in the letters to the editors column.

One thing that may throw you a curve is the use of superbright leds in lights. I've heard some of the IR gear doesn't detect LEDs the same as incandescents. The explanation I got on that was related to enhanced vision systems and runway lighting.
For slow aircraft at close range, I think it might be OK. It would be hard to beat TCAS ll for detection and tracking of longer range aircraft at same azimuth and altitude.
 
The company I work for put together a HUD/Enhanced Vision forward looking IR package for our fleet. I wasn't directly involved, but we do support it now. The exterior lens is heated, I suppose for anti ice, the camera is cyrogenetically cooled. Of course, thats an whole other level of complexity. But, maybe some of that is necessary for it to work in a cloud at near freezing temps.
 
Just like UV detectors on military aircraft, I think you'll be seeing just as many false indications as you would real aircraft. You'd have to narrow the FOV so it isn't picking up vehicles on the ground. Also, I don't see any of these electric UAVs emitting enough IR energy to be detected on a passive system.
 
Another issue with IR is severe attenuation with humidity. You won't be seeing anything in IR in IMC for example, or at best you'll get a marginal improvement from visual wavelengths, particularly in high-humidity environments.
 
I can understand how you'd get azimuth, but not range or relative altitude. Also, would it easily distinguish an aircraft amid warm ground clutter?

Also, I don't think it would work well in clouds. I have a handheld thermal camera and it cannot see nearly as far through fog as it can in the clear. And in fog we're talking about a few hundred feet. Maybe it would see something really hot like an engine (cowled engine?) but not sure how far that would be effective. Certainly it wouldn't be measured in miles.

There are aviation thermal IR cams on the market but they are high dollar and besides your description sounds different - more like azimuth and range rather than IR cam?

Interesting idea.
 
Last edited:
kontiki said:
The exterior lens is heated, I suppose for anti ice, the camera is cyrogenetically cooled.
Click to expand...

The cooling is used in systems working at longer wavelengths. This is necessary because any physical body emits IR radiation. The peak emission changes with tempearture, but there's a spread. Thus, parts of the receiver that are absolutely transparent, like the lens, emit it too. At certain temperature such emissions drown out the heat signature the device has to capture. Subcooling reduces this noise.

Since I'm most interested in software, I'm going to go with receivers cooled with air, meaning that I'm not going to see a few things. For example a skydiver in free fall, or a sweating skydiver under a canopy when it's +35C outside. However, at wavelengths I'm considering, exhaust stubs on GA airplanes are going to shine better than their landing lights.
 
I think its an excellent challenge though I really believe the future of collision avoidance lies in a GPS. There is already an RTCA Minimum Operational Performance Standard published for a GPS based TCAS hybrid, DO-300.

I was probably not feasible in time for universal ADS-B implementation, but I think it will happen one day.
 
You must log in or register to reply here.
Back
Top