This is a follow-up to my previous question:

How does this IMU work and how to convert its output into meaningful information?

for which many people asserted that I need a pitot tube and a static port in order to calculate airspeed and pressure altitude respectively, and errors will build up if I just integrate the accelerometer values from the IMU.

I'd like to know whether it is possible to measure airspeed without a pitot tube, accurately? Is there any electronic sensor that can be used for this purpose?

Also are there any airplanes around the world, Airliners, GA, Military, or even Experimental, that operate without any pitot tubes?

  • $\begingroup$ Not a duplicate, but essentially the same answers to this question. $\endgroup$ – kevin Oct 27 '17 at 2:34
  • 2
    $\begingroup$ @kevin: Does your answer cover the possibility of measuring airspeed other than a hand-held GPS? While your answer is correct, it is quite different from what I asked, as your answer was about discerning the airspeed of an airliner from the inside (with a hand-held GPS). I'm looking into other ways an avionics system could be designed, without a pitot tube - so this assumes that I have all access to the instruments. But I don't even have any instrument to measure the airspeed without a pitot tube. : ) $\endgroup$ – user18035 Oct 27 '17 at 2:41
  • $\begingroup$ In the accidents I referenced, the pilots did not have pitot tube; and given all other instruments they had in the cockpit, they could not deduce the airspeed and crashed. And the reasons why one cannot calculate airspeed with the remaining instruments it the same. $\endgroup$ – kevin Oct 27 '17 at 2:48
  • 1
    $\begingroup$ Tie knots in a long piece of dental floss... crack the window open... $\endgroup$ – J... Oct 27 '17 at 9:59
  • 2
    $\begingroup$ Just mentioning it for completeness sake, but you could duct tape/glue one of these lovely spinning wind speed meters onto the outside of your aircraft. Yes they would measure the airspeed but I think it goes without saying that there are probably good reasons pitot tubes are used instead. $\endgroup$ – JustWannaFly Oct 27 '17 at 14:59

Technically speaking, inertial measurements are not sufficient to derive airspeed. You need pressure measurement (or airspeed measurement, see LIDAR). Pitot-static probes are the most usual and conventional devices to measure pressure.

However, several researches have been conducted, to replace pitot-static tubes with different technologies. These technologies are mainly based on measuring surface pressure on a part of the fuselage. Of course the measurements (not a single value but several values from various locations) have to be calibrated via flight tests.

With the current popularity of machine learning I think the calibration algorithms will not be the main problem.

Some other technology being investigated

  1. there's also ongoing research and trials about using LIDAR to directly measure airspeed. Which won't need pressure measurement.

  2. ultrasonic devices are also able to measure airspeed directly, however their real flight usage is unknown (to the author).

  • 3
    $\begingroup$ I think it is problematic currently to certify a machine learning solution for a safety critical instrument, so it may be your main problem. $\endgroup$ – DeltaLima Oct 27 '17 at 7:39
  • 3
    $\begingroup$ @Clearer, sure that's called "star tracker" and is used for navigation purposes since the 60's. but it doesn't give you the airspeed. $\endgroup$ – Gürkan Çetin Oct 27 '17 at 9:48
  • 4
    $\begingroup$ @Clearer that will give you ground speed not airspeed. $\endgroup$ – Notts90 supports Monica Oct 27 '17 at 10:48
  • 1
    $\begingroup$ @Clearer, by the way, the sun is too close to earth and as far as I know, for star tracker they use distant stars. Using the sun makes sense somewhat but your heading would be too important probably, which is already not the best data you have with an IMU device. $\endgroup$ – Gürkan Çetin Oct 27 '17 at 13:56
  • 1
    $\begingroup$ @Harper Since incipient stall is dependent on AoA rather than airspeed, stall warnings are usually triggered by an AoA sensor of some sort. Either an AoA vane on the fuselage or a simple flapper switch on the wing leading edge or often an actual horn or whistle whose opening is placed on the wing at such an angle that it makes a noise at a certain AoA. $\endgroup$ – TomMcW Oct 27 '17 at 17:23

Airspeed - not reliably under all circumstances. You need to feel that wind somewhere in order to get a direct and accurate measure of the speed of it, and the pitot tube is an accurate and proven instrument to measure total pressure.

Other possible methods:

  1. Laser based(LIDAR). A 20 year old NASA report can be found here. It mentions at the end that the method was not always accurate:

    An inherent source of error in the system is noise generated by ambient or background illumination. The most intense source during daylight hours is, of course, the sun. As a consequence, the smallest detect- able scattered light signal is a strong function of the angle between the optical axis and the direct line to the sun. On some occasions, velocity measurements with the sheet-pairs system were impossible when this angle was less than about 30°

  2. Pressure measurement from the skin of the aircraft. Problem is that the stagnation point moves as the aircraft angle of attack changes, the pitot tube front opening always hits the stagnation point. But yeah you could indeed stick a whole lot of static pressure ports on the aircraft and calibrate them with a proper towed flying pitot.

For ground speed and navigation it is a different matter, GPS makes nulling the integration errors for ground speed very simple.

  • $\begingroup$ Why do we need integration, as the GPS itself can directly calculate the ground speed? $\endgroup$ – user18035 Oct 27 '17 at 6:20
  • 2
    $\begingroup$ Yes it can. It does so with a low update rate and with a phase shift (time delay) since it computes the velocity from previous position signals. I've tried to not exceed max speed using my handphone GPS when I drove in a car without a speedometer, it is very hard to do since it always tells you how fast you were going 5 seconds ago. $\endgroup$ – Koyovis Oct 27 '17 at 6:56
  • $\begingroup$ Very few people will quibble with your claim that pitot tubes are accurate and reliable. Least of all those onboard Northwest 6321, Aeroperu 603, or Air France 447. Pitot failures are among the least survivable accidents in aviation. $\endgroup$ – Harper - Reinstate Monica Oct 27 '17 at 16:39
  • $\begingroup$ @Harper Well, AF447 was unsurvivable not so much do to pitot failure as due to pilot failure. The pitot system was working fine for the last several minutes of flight and all of the indications were accurate during that time. Still, your general point is accurate. Unreliable airspeed is a big problem. $\endgroup$ – reirab Oct 28 '17 at 8:08
  • $\begingroup$ Lidar is indeed very accurate, and can be used over a wide range of speeds, from extremely low velocities to hypersonic velocities. Study of the current literature will show that. $\endgroup$ – mongo Oct 28 '17 at 14:06

It is possible to think devices that could potentially replace Pitot tubes, while it is not very obvious how good they would operate under conditions of the real flight, for instance:

  • Compare temperature of the heated wire cooled by the air flow with the temperature of the similar wire that is in the same air but shielded from the flow.
  • Measure the time sound (or ultrasound) takes to travel between two points within the air flow. The travel speed should be the speed of sound in the air plus the speed the air itself is moving, carrying the propagating sound wave.
  • If there are some particles (snow, hail, etc) in the air, it may be possible to measure the speed of these particles in the air flow.

The first two types of devices seem exist, but I found no information on using them as Pitot tube replacements.

  • $\begingroup$ The first device you mention is usually called a hot-wire anemometer. I wouldn't like to have to climb out onto the aircraft nose at 35,000 feet altitude to replace the delicate little wire if it broke(!) $\endgroup$ – Andrew Morton Oct 27 '17 at 8:11
  • $\begingroup$ It is also possible to have more than one wire in the single device. $\endgroup$ – h22 Oct 27 '17 at 8:44
  • $\begingroup$ Perhaps I should have emphasised the word "delicate" more. Although hot wires survive well where they are used in automobiles in air mass-flow meters, there are usually warnings in, e.g., Haynes manuals to avoid touching the wire. So it couldn't be out in the open exposed to things like hail. If it were found to be superior to using Pitot tubes, I'm sure a mechanism would be devised to automatically replace the wire. $\endgroup$ – Andrew Morton Oct 27 '17 at 8:50
  • $\begingroup$ The issue with the hot wire anemometer is you still need to know air pressure, or rather air density. Thinner air does not cool as well as denser air. $\endgroup$ – Trevor_G Oct 27 '17 at 18:22
  • $\begingroup$ Seems no problem just to use barometer for that. $\endgroup$ – h22 Oct 28 '17 at 6:15

OP question answer: Yes, it is possible to much more accurately measure airspeed with Doppler lidar, than it is with a pitot /static system.

Discussion below:

I have used particle scattering and Geiger mode lidar for velocity and flow measurements, particularly (no pun intended) when a tempo/spatial map of the flow is desired. In atmospheric air, there are always some particles!

Here is an example of recent work in creating a lidar based sensor for aircraft velocity measurements: https://hal.archives-ouvertes.fr/hal-01111306/document This device provides TAS, angle of slip and AOA.

Here is a device which utilizes Doppler velocimetry. While the article suggests that the technique used is not lidar, there are several lidar techniques which perform analogous functions. http://optics.org/news/5/12/35

Here is a BAE concept which has been demoed at airshows, and utilizes Doppler lidar processing, utilizing a UV laser. http://www.baesystems.com/en/article/bae-systems-develops-laser-airspeed-sensor-for-aircraft

Here is a Doppler sensor for air data patent, which is 6 years old. https://www.google.com/patents/US8434358

Quoting this press release:

Airbus Group has completed successful flight tests on a fiber-optic, eye-safe, laser-based sensor system that delivers accurate airspeed information in the three axis at low and even negative airspeed. This range of capability is not possible with pitot tubes, the longstanding industry standard for airspeed sensors.

A pitot tube / static port is a rather simple and inexpensive device.

IMUs and GPS proposed solutions are misplaced and will have a phenomenal error budget and do not accurately measure airflow.

NASA promoted a Rayleigh scattering lidar device a couple of years ago (2015?), so there are new developments.

A couple of decades ago the Navy was promoting an ultrasound doppler device.

While there are many ways of replacing the functionality of a pitot / static airspeed sensor, all are more expensive than a pitot system. However, the most likely technology to emerge with a reasonable sensor cost will be a Doppler lidar variant.

  • $\begingroup$ The Airbus Laser Sensor seems promising! How long do you think before it hits the market? And how expensive might it be? $\endgroup$ – user18035 Oct 29 '17 at 1:53
  • $\begingroup$ The 6 yr. old Japanese Patent looks quite convincing too... Let's hope all these new tech. are equal to, if not better than the pitot / static system! $\endgroup$ – user18035 Oct 29 '17 at 2:01
  • 1
    $\begingroup$ Doppler lidar is more accurate, and has two orders of magnitude more precision when measuring airflow than a conventional pitot static system. Looking at other technologies, such as MEMS devices for projectors, vibration sensors and IMU type applications, costs have precipitously dropped once there is volume. $\endgroup$ – mongo Oct 29 '17 at 13:11

If my understanding is correct, all the solutions mentioned in the answers so far would provide true airspeed, not indicated airspeed. True airspeed is useful for navigation, but not for flight to determine critical speeds.

A pitot static system measures indicated airspeed which is the best indicator of things like when a stall will occur (at the slow end) and when the tail will rip off (at the fast end).

  • $\begingroup$ Very true indeed! TAS can even be obtained from a GPS - it needn't be accurate, but IAS on the other hand is very crucial to the sustenance of the flight itself! I agree with you. $\endgroup$ – user18035 Oct 28 '17 at 11:02
  • $\begingroup$ @AnandS, please explain to me how TAS can be obtained from a GPS. $\endgroup$ – mongo Oct 28 '17 at 13:56
  • $\begingroup$ @mongo: The GPS automatically provides the TAS, like the Garmin 530 (I know that from FSX!), also TAS can be calculated from the distance traveled in a second (automatically done by GPS, or we can do it manually). Am I right? $\endgroup$ – user18035 Oct 28 '17 at 15:58
  • 2
    $\begingroup$ @AnandS, I thought that TAS on the 530 was "Traffic Advisory System." But there is a DENALT/TAS/WINDS page. which will do calculations. But the catch is that to use the TAS calculator, you enter the CAS. So it relies on the manual input of CAS into the calculator. TAS is not ground covered, rather it is the speed of the aircraft in the air mass, adjusted for atmospherics (like DENALT). $\endgroup$ – mongo Oct 28 '17 at 16:40
  • $\begingroup$ And for CAS, we need the Pitot / Static system... It's a vicious cycle, huh? $\endgroup$ – user18035 Oct 29 '17 at 1:51

A wind anemometer could provide an indication of relative airflow:

Source: https://en.wikipedia.org/wiki/Anemometer#/media/File:Wea00920.jpg

Or a metal plate with a spring

Source: http://www.williammaloney.com/Aviation/VintageWingsOfCanada/DeHavillandTigerMoth/images/09TigerMothAirspeedIndicator.jpg

The front of the plate is receiving a dynamic pressure, while the rear of the plate is an approximation of static pressure.

Neither of them would handle icing well without heating.

Also are there any airplanes around the world, Airliners, GA, Military, or even Experimental, that operate without any pitot tubes?

An aircraft at very high speed relative to the wind, was well above stall speed, and well below any never exceed speed wouldn't need air data until it slowed down (e.g. Space shuttle's deploy-able probes).

  • 1
    $\begingroup$ I can imagine how the takeoff roll goes on the flight deck... "80 knots" "check" "V1" "Rotate" "Anemometer is broken again" "I agree" $\endgroup$ – reirab Oct 28 '17 at 8:29
  • $\begingroup$ Are you sure you want to hang metal plates and spinning cups outside an airplane that is travelling 1000s of feet up in the sky at speeds of 100+ knots? What if they broke off and fell down? : ) But seriously - your points are very valid for grounded and fixed purposes, not really for use in (or out?) airplanes! $\endgroup$ – user18035 Oct 28 '17 at 11:13
  • 1
    $\begingroup$ An anemometer as you have pictured is agnostic as to the direction, and would fail to provide airspeed across the airfoil, at least in an conventional aircraft. $\endgroup$ – mongo Oct 28 '17 at 14:00
  • $\begingroup$ @mongo: Exactly! It will also fail when there are crosswinds existing, as it cannot calculate wind speed from a particular direction. It will sum it all up - Lovely concept for wind turbines and wind-mills! : ) $\endgroup$ – user18035 Oct 29 '17 at 4:16
  • 2
    $\begingroup$ @AnandS, well, the OP didn't insist on practical methods ;) I can throw in a few more: shut down one engine and calibrate airspeed from the windmilling prop RPM or N1, or deploy RAT (ram air turbine) and measure its raw voltage. Or on aircraft with reversible (direct) control, measure control force vs deflection (this is how pilots of light airplanes can feel the indicated airspeed). $\endgroup$ – Zeus Nov 2 '17 at 1:36

Yes, absolutely.

A decade ago Bill Premerlani made an ingenious recognition that a change of attitude relative to a change of groundspeed reveals the hidden airspeed state.

Very briefly, here is the driving equation. (I refer the reader to the below documents for a full explanation of the variables).

enter image description here

The DIYDrones link where B. Premerlani announces his work: https://diydrones.com/forum/topics/wind-estimation-without-an

The white paper: https://st3.ning.com/topology/rest/1.0/file/get/3690830434?profile=original

Here is the code section written by Premerlani for his autopilot: https://github.com/MatrixPilot/MatrixPilot/blob/5bacc66190b65fecc1f01f92326b22df4c237215/libDCM/estWind.c

Here is the same algorithm written by me, for another autopilot project: https://github.com/TauLabs/TauLabs/blob/d0512149b454894ae0cf98c177afc1fb6a68f806/flight/Modules/Airspeed/gps_airspeed.c#L99

There are a couple nuances:

  1. This only works when the plane is changing attitude. So for straight and level flight the airspeed estimation will diverge. This is a limitation for long distance cruise flights, since they presumably have few attitude changes. In those situations, you'd have to depend on knowing other factors, such as the airplane's configuration, if you wanted to continue to estimate airspeed.
  2. It requires coordinated flight. If the aircraft is slipping through the air sideways then the model premise is violated and the algorithm results will suffer.

I'm always surprised the B. Premerlani's work doesn't get more traction. Unlike all other GPS-based airspeed estimators that I know off, it does not require a level 360 degree turn. It is sufficient to have a series of small (~5 degree) pitch and/or heading changes. I can attest that it works well in practice.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy