Air data display device and calibration method thereof

Abstract

The present invention relates to an air data display device (1) and calibration method for providing information about the airspeed of a vertical take-off and landing (VTOL) aircraft, particularly for a helicopter (50), and information about the altitude of the VTOL aircraft. The VTOL aircraft includes a pitot tube system (51) for determining the stagnation air pressure at its location and a static pressure port system (52) for determining the static pressure at its location. The air data display device (1) includes an airspeed and altitude determination module (2) for determining the airspeed and altitude of the VTOL aircraft in real time from flight data by using a regressor (3) obtained by training an artificial neural network with training data.

Description

[Technical Field]

[0001] The present invention relates to an air data display device for a vertical take-off and landing (VTOL) aircraft, particularly a helicopter, for providing information regarding the airspeed of the VTOL aircraft and for providing information regarding the altitude of the VTOL aircraft, the VTOL aircraft including a pitot tube device for determining stagnation air pressure at a pitot tube device and providing pitot tube data including information regarding the stagnation air pressure at the pitot tube device, and a static port device for determining static pressure at a static port device and providing static port data including information regarding the static pressure at the static port device. The present invention also relates to a method for calibrating such an air data display device. [Background technology]

[0002] Air data displays belonging to the first-mentioned technical field are known. Air data displays are included in so-called pitot-static systems implemented on vertical take-off and landing (VTOL) aircraft. Such pitot-static systems use pitot tube devices located on the VTOL aircraft to measure stagnation air pressure at the pitot tube device locations and static pressure port devices located on the VTOL aircraft to measure static pressure at the static port device locations. Using the air data display, the airspeed and altitude of each VTOL aircraft are determined using pitot tube data containing information about stagnation air pressure measured by the pitot tube device at the pitot tube device locations on the VTOL aircraft and static port data containing information about static pressure measured by the static port device at the static port device locations on the VTOL aircraft. The airspeed and altitude of the VTOL aircraft are then displayed to the pilot of the VTOL aircraft, or data containing information about the airspeed and altitude of the VTOL aircraft is then transferred to the VTOL aircraft's autopilot, causing the VTOL aircraft to fly automatically in autopilot mode.

[0003] These pitot-static systems are essential to the safety of VTOL aircraft. If a pilot does not know the correct airspeed and correct altitude, or if they rely on incorrect values ​​for the VTOL aircraft's airspeed and altitude, the pilot may attempt a maneuver that leads to a crash of the VTOL aircraft. Similarly, if data containing information about the VTOL aircraft's incorrect airspeed and / or incorrect altitude is transmitted to the VTOL aircraft's autopilot, it could lead to a crash of the VTOL aircraft.

[0004] Therefore, the pitot-static system must reliably display the airspeed and altitude of the VTOL aircraft to the pilot with an accuracy equal to or greater than that required by regulatory specifications. To accomplish this, the air data display must reliably determine the airspeed and altitude of the VTOL aircraft with an accuracy equal to or greater than that required by regulatory specifications and provide correct information regarding the airspeed and altitude of the VTOL aircraft.

[0005] However, in a VTOL aircraft, the stagnation pressure at the pitot tube device and the static pressure at the static port device depend not only on the VTOL aircraft's actual airspeed and actual altitude, but also on the VTOL aircraft's current flight configuration and current flight state. More precisely, the VTOL aircraft is surrounded by air currents and turbulence, which in turn depend on where the pitot tube device and static port device are located on the VTOL aircraft. These air currents and turbulence at the pitot tube device and static port device change with changes in the VTOL aircraft's flight configuration and flight state. Therefore, at a particular VTOL aircraft airspeed and altitude, the stagnation pressure measured by the pitot tube device and the static pressure measured by the static port device depend on the VTOL aircraft's current flight configuration and current flight state. Therefore, to provide the VTOL aircraft's pilot with airspeed and altitude with acceptable accuracy, the VTOL aircraft's air data display must be calibrated.

[0006] Methods for calibrating air data displays are described, for example, in a paper by Denis Hamel and Alex Kolarich entitled "GPS-Based Airspeed Calibration for Rotorcraft: General Application to All Flight Regimes," presented at the Vertical Flight Society's 76th Annual Forum & Technology Display, held virtually from October 6 to 8, 2020. The paper concerns the calibration of pitot-static systems. Therefore, according to the paper, a GPS-based true airspeed method with adaptive implementation and analysis techniques is the most practical in terms of equipment and efficiency for providing a complete airspeed system calibration.

[0007] However, there are no known calibration methods that can correct the airspeed and altitude values ​​determined by an air data display on a VTOL aircraft for all flight configurations and conditions with the accuracy required by regulatory specifications.

[0008] In addition to these calibration issues discussed above, the stagnation air pressure at the pitot tube device and the static pressure at the static port device also vary depending on where the pitot tube device and the static port device are located on the VTOL aircraft. For example, a common observation is that the closer the pitot tube device and the static port device are located on the fuselage of the VTOL aircraft, the more the stagnation air pressure measured by the pitot tube device and the static pressure measured by the static port device deviate from the effective stagnation air pressure and static pressure. Nevertheless, when designing new VTOL aircraft that include pitot-static systems, the pitot tube device and the static port device must often be repeatedly relocated because the deviations in the measured stagnation air pressure and static pressure deviate too much from the effective stagnation air pressure and static pressure to be corrected by calibration with the required accuracy. This often results in multiple design iterations, each involving test flights using a prototype of that type of VTOL aircraft to verify the accuracy of the pitot-static system, followed by a new calibration of the employed pitot-static system's air data display, until a VTOL aircraft design and the placement of pitot tubes and static porting devices on the VTOL aircraft are found that meet the requirements of regulatory specifications for pitot-static systems. Because these iterations are time-consuming and costly, it is important to reduce the number of iterations required. One known method of reducing the number of iterations is to locate pitot tubes and static porting devices in non-critical pressure fields, particularly away from the VTOL aircraft's fuselage. However, this approach has the disadvantage of being costly and posing safety hazards to ground crews. Summary of the Invention [Problem to be solved by the invention]

[0009] The object of the present invention is to create an air data display device related to the technical field mentioned at the beginning, and a method for calibrating such an air data display device, so that the air data display device is able to provide correct values ​​of airspeed and altitude for any flight configuration and flight condition with the accuracy required by regulatory specifications, and which allows a cheaper and safer design procedure for designing VTOL aircraft. [Means for solving the problem]

[0010] The solution of the present invention is specified by the features of claim 1. According to the present invention, the air data display device is connectable to a pitot tube device for receiving pitot tube data provided by the pitot tube device, and the air data display device is connectable to a static port device for receiving static port data provided by the static port device, whereby the air data display device includes an airspeed and altitude determination module, the airspeed and altitude determination module being adapted to determine in real time the airspeed of the VTOL aircraft and the altitude of the VTOL aircraft from flight data using regressors obtained by training an artificial neural network with training data, the flight data including at least the pitot tube data, the static port data, vertical speed data including information regarding the vertical speed of the VTOL aircraft, and pitch attitude angle data including information regarding the pitch attitude angle of the VTOL aircraft.

[0011] According to the present invention, the air data display device is for a vertical take-off and landing (VTOL) aircraft, in particular a helicopter. This makes it irrelevant whether the VTOL aircraft or helicopter, respectively, is a manned or unmanned aircraft. If the VTOL aircraft or helicopter, respectively, is a manned aircraft, the pilot of the VTOL aircraft or helicopter, respectively, may be on board the VTOL aircraft or helicopter, respectively, or may be on the ground if the VTOL aircraft or helicopter, respectively, is remotely controllable by a remote control device. If the VTOL aircraft or helicopter, respectively, is an unmanned aircraft, it is advantageous for the VTOL aircraft or helicopter, respectively, to be remotely controllable by a remote control device. In all of these cases, the VTOL aircraft or helicopter, respectively, may also include an autopilot that relies on a pitot-static system or the like that includes the air data display device according to the present invention.

[0012] According to the present invention, the airdata display device provides information regarding the airspeed of the VTOL aircraft and provides information regarding the altitude of the VTOL aircraft. Thus, the airdata display device is advantageously adapted to provide information regarding the airspeed of the VTOL aircraft as determined by the airspeed and altitude determination module, and to provide information regarding the altitude of the VTOL aircraft as determined by the airspeed and altitude determination module. Thus, in one example, the airdata display device includes a display for displaying the airspeed and altitude of the VTOL aircraft. In this example, the information regarding the airspeed of the VTOL aircraft and the information regarding the altitude of the VTOL aircraft are provided by the airdata display device by displaying the airspeed and altitude on the display. In another example, the airdata display device includes an output port for outputting output data including the information regarding the airspeed of the VTOL aircraft and the information regarding the altitude of the VTOL aircraft. In this other example, the output port may be connectable to an autopilot of the VTOL aircraft. Alternatively, the output port may be connectable to another device, such as an on-board computer of a VTOL aircraft, so that output data from the air data display device is transmitted to the on-board computer and airspeed and altitude are displayed on a display controlled by the on-board computer. Alternatively, if the VTOL aircraft is remotely controllable by a remote control device, the output port may be connectable to the remote control device of the VTOL aircraft and output data from the air data display device may be transmitted to the remote control device and airspeed and altitude may be displayed on the remote control device. Thus, the air data display device may be located, for example, within the VTOL aircraft and transmit output data to the remote control device, or may be located, for example, within the remote control device and receive pitot tube data and static port data transmitted from the VTOL aircraft. In any of the variations with output ports, information regarding the airspeed of the VTOL aircraft and information regarding the altitude of the VTOL aircraft are provided by the air data display device in the form of output data.

[0013] In accordance with the present invention, an airdata display device includes an airspeed and altitude determination module adapted to determine the airspeed of the VTOL aircraft and the altitude of the VTOL aircraft in real time from the flight data using regressors obtained by training an artificial neural network with the training data. In one example, the airspeed and altitude determination module is a computer program product running on a computing unit, such as a control computer of the VTOL aircraft, or a separate computing unit distinct from the control computer of the VTOL aircraft. In another example, the airspeed and altitude determination module is a computing unit adapted to determine the airspeed of the VTOL aircraft and the altitude of the VTOL aircraft in real time from the flight data using regressors obtained by training an artificial neural network with the training data.

[0014] Regardless of whether the airspeed and altitude determination module is a computer program product or a computing unit, the airspeed and altitude determination module is adapted to determine the airspeed of the VTOL aircraft and the altitude of the VTOL aircraft in real time from the flight data using a regressor obtained by training an artificial neural network with the training data. Thus, during operation of the airspeed and altitude determination module, the flight data is advantageously provided to the regressor in real time to determine the airspeed of the VTOL aircraft and the altitude of the VTOL aircraft in real time from the flight data. Thus, in one example, the regressor outputs the airspeed and altitude of the VTOL aircraft. In another example, the airspeed and altitude of the VTOL aircraft can be calculated by the airspeed and altitude determination module from the output of the regressor. In the latter example, the output of the regressor can be, for example, a corrected stagnation air pressure and a corrected static pressure, from which the airspeed of the VTOL aircraft and the altitude of the VTOL aircraft can be calculated by the airspeed and altitude determination module.

[0015] According to the present invention, the flight data includes at least pitot tube data, static port data, vertical speed data including information regarding the vertical speed of the VTOL aircraft, and pitch attitude angle data including information regarding the pitch attitude angle of the VTOL aircraft.

[0016] Because the pitot tube data includes information about stagnation air pressure at the location of the pitot tube device, the pitot tube data may be, for example, the stagnation air pressure measured at the pitot tube device in any units output by the pitot tube device, an uncorrected airspeed calculated based on the measured stagnation air pressure and the measured static pressure, or a pre-corrected airspeed, which may be, for example, an uncorrected airspeed corrected with a known position correction from a look-up table.

[0017] Because the static port data contains information about the static pressure at the location of the static port, the static port data can be, for example, the static pressure measured at the static port in any units output by the static port, an uncorrected altitude calculated based on the measured static pressure, or a pre-corrected altitude, which can be, for example, an uncorrected altitude corrected with a known position correction from a look-up table.

[0018] Because the vertical velocity data contains information about the vertical velocity of the VTOL aircraft, the vertical velocity data can be obtained from the static pressure port data, for example, using the change in static pressure port data over time. This can be, for example, the change in static pressure measured at the static pressure port over time in any units output by the static pressure port, the change in uncorrected altitude calculated based on the measured static pressure, or the change in pre-corrected altitude over time. In either case, this can be the change per predefined time unit.

[0019] Since the pitch attitude angle data includes information about the pitch attitude angle of the VTOL aircraft, the pitch attitude angle data can be obtained, for example, from an Air Data Attitude Heading Reference System (ADAHRS) of the VTOL aircraft, in particular from a gyroscopic flight instrument or a Micro-Electro-Mechanical System (MEMS) gyroscope. To receive the pitch attitude angle data, the air data display device is advantageously connectable to a pitch attitude angle data providing unit that provides the pitch attitude angle data. This pitch attitude angle data providing unit can be, for example, a gyroscopic flight instrument or a Micro-Electro-Mechanical System (MEMS) gyroscope. However, the pitch attitude angle data providing unit can also be a separate unit or a computer that receives pitch attitude angle data from the aforementioned gyroscopic flight instrument or Micro-Electro-Mechanical System (MEMS) gyroscope, respectively. Thus, the pitch attitude angle data providing unit can be, for example, part of the Air Data Attitude Heading Reference System (ADAHRS) of the VTOL aircraft.

[0020] In an inventive method for calibrating an airdata display device, the airspeed and altitude determination module of the airdata display device is calibrated by obtaining regressors by training a neural network with training data comprising training data sets for calibrating the airdata display device, whereby each training data set is associated with a flight condition and includes flight data acquired during flight in the respective flight condition by a VTOL aircraft of the type for which the airdata display device is calibrated, particularly a helicopter. Each training data set includes, for a respective type of VTOL aircraft, reference output data corresponding to a desired output of the regressor while being flown by the VTOL aircraft at each flight condition, wherein the flight data includes at least: pitot tube data obtained from a pitot tube device of the VTOL aircraft for each type of VTOL aircraft while being flown by the VTOL aircraft at each flight condition, static pressure port data obtained from a static pressure port of the VTOL aircraft for each type of VTOL aircraft while being flown by the VTOL aircraft at each flight condition, vertical speed data including information regarding the vertical speed of the VTOL aircraft for each type of VTOL aircraft while being flown by the VTOL aircraft at each flight condition, and pitch attitude angle data including information regarding the vertical VTOL aircraft pitch attitude angle for each type of VTOL aircraft while being flown by the VTOL aircraft at each flight condition.

[0021] Thus, advantageously, for each training data set, the flight data of the respective training data set is acquired and advantageously recorded simultaneously. Thus, advantageously, for each training data set, the flight data of the respective training data set is acquired and advantageously recorded during the same test flight with a VTOL aircraft of the respective type, while flying one maneuver in each flight condition. This has the advantage that, for each training data set, different types of data of flight data belong together. Different training data sets of training data can be acquired during the same test flight or during different test flights. In the latter case of different test flights, the different training data sets of training data are advantageously acquired with the same VTOL aircraft of the respective type. However, different training data sets of training data can also be acquired during different test flights using different VTOL aircraft of the same type.

[0022] In all of these variations, for each training data set, reference output data corresponding to the desired output of the regressor during flight of each type of VTOL aircraft under each flight condition and the flight data of each training data set can be simultaneously acquired or recorded, respectively, although the reference output data can also be obtained by calculation based on the flight data of each training data set.

[0023] Thus, flight data and reference power data may be measured and recorded, for example, during one or more test flights by a VTOL aircraft of a particular type, after which the recorded flight data and reference power data may be used to calibrate the air data display devices of the respective VTOL aircraft and to calibrate air data display devices of further VTOL aircraft of the respective type. In another example, flight data may be measured and recorded during one or more test flights for certifying a particular VTOL aircraft, during which reference power data is calculated from the flight data, after which the recorded flight data and the recorded or calculated reference power data are used to calibrate the air data display devices of the respective VTOL aircraft.

[0024] An advantage of the inventive air data display device and the inventive method for calibrating an air data display device is that it can quickly provide a calibrated air data display device that provides accurate airspeed and altitude values ​​with the precision required by regulatory specifications for any flight configuration and flight condition. Therefore, the inventive solution enables improved safety in VTOL aircraft operations by providing the VTOL aircraft's pilot and autopilot with accurate airspeed and altitude values ​​for the VTOL aircraft in real time. Furthermore, the inventive air data display device and the inventive method for calibrating an air data display device enable a cheaper and safer design procedure for VTOL aircraft, with little or no repeated relocation of pitot tube devices and static ports. Furthermore, the inventive solution allows the pitot tube devices and static ports to be located closer to the fuselage of the VTOL aircraft than when employing known air data displays. Avoiding parts that protrude far from the fuselage reduces the risk of injury to personnel handling the aircraft. Furthermore, the solution according to the present invention allows for simple and easy implementation on existing VTOL aircraft by using existing pitot tube devices and existing static pressure ports on each VTOL aircraft, allowing for easy calibration of the air data display devices after test flights on each VTOL aircraft to obtain training data and after using the training data to train the artificial neural network.

[0025] Preferably, the flight data includes roll attitude angle data containing information about the roll attitude angle of the VTOL aircraft. This has the advantage that the method for calibrating the air data device can more accurately calibrate the air data device. Because the roll attitude angle data contains information about the roll attitude angle of the VTOL aircraft, the roll attitude angle data can be obtained, for example, from an Air Data, Attitude, Heading and Reference System (ADAHRS) of the VTOL aircraft, in particular a gyroscopic flight instrument or a microelectromechanical system (MEMS) gyroscope.

[0026] Alternatively, the flight data may not include roll attitude angle data, which has the advantage of simplifying the calibration of the air data equipment and requiring less computing power.

[0027] Advantageously, the flight data includes sideslip angle data containing information about the sideslip angle of the VTOL aircraft. This has the advantage that the method for calibrating the air data device can more accurately calibrate the air data device. In a variation that allows for more accurate calibration of the air data device, the flight data includes roll attitude angle data containing information about the roll attitude angle of the VTOL aircraft, and sideslip angle data containing information about the sideslip angle of the VTOL aircraft. Because the sideslip angle data contains information about the sideslip angle of the VTOL aircraft, the sideslip angle data can be obtained, for example, from a dedicated sensor. For example, the sideslip angle can be calculated using a dedicated sideslip sensor that utilizes the total pressure measured at two additional static pressure ports from the VTOL aircraft, one facing right and one facing left.

[0028] However, the flight data may alternatively not include sideslip angle data, which has the advantage of simplifying the calibration of the air data equipment and requiring less computing power.

[0029] Preferably, the flight data includes information regarding the lateral acceleration of the VTOL aircraft, in particular lateral acceleration data including an amount of lateral acceleration of the VTOL aircraft. This has the advantage that the method for calibrating the air data device can more accurately calibrate the air data device. Since the lateral acceleration data includes information regarding the lateral acceleration of the VTOL aircraft, the lateral acceleration data can be obtained, for example, from an Air Data Attitude Heading Reference System (ADAHRS) of the VTOL aircraft, in particular from one or more accelerometers.

[0030] In a first variation that allows for more accurate calibration of the air data device, the flight data includes roll attitude angle data containing information about the roll attitude angle of the VTOL aircraft, and lateral acceleration data containing information about the lateral acceleration of the VTOL aircraft, particularly the amount of lateral acceleration of the VTOL aircraft. In a second variation that allows for more accurate calibration of the air data device, the flight data includes information about the lateral acceleration of the VTOL aircraft, particularly the lateral acceleration data containing the amount of lateral acceleration of the VTOL aircraft, and sideslip angle data containing information about the sideslip angle of the VTOL aircraft. In a third variation that allows for more accurate calibration of the air data device, the flight data includes roll attitude angle data containing information about the roll attitude angle of the VTOL aircraft, sideslip angle data containing information about the sideslip angle of the VTOL aircraft, and lateral acceleration data containing information about the lateral acceleration of the VTOL aircraft, particularly the amount of lateral acceleration of the VTOL aircraft.

[0031] However, the flight data may alternatively not include lateral acceleration data, which has the advantage of simplifying the calibration of the air data equipment and requiring less computing power.

[0032] Advantageously, a pitot-static system includes an air data display device according to the present invention, a pitot tube device, and a static port.

[0033] Preferably, a vertical take-off and landing (VTOL) aircraft, in particular a helicopter, is equipped with an air data display device according to the present invention. The VTOL aircraft advantageously further comprises a pitot tube device and a static pressure port. In this case, the VTOL aircraft advantageously includes the aforementioned pitot-static system including an air data display device according to the present invention. However, alternatively, the VTOL aircraft may not include a pitot tube device and a static pressure port. In any case in which the VTOL aircraft includes an air data display device according to the present invention, the VTOL aircraft advantageously further includes a display for displaying the airspeed of the VTOL aircraft from information regarding the airspeed of the VTOL aircraft received from the air data display device, and for displaying the altitude of the VTOL aircraft from information regarding the altitude of the VTOL aircraft received from the air data display device.

[0034] However, in other configurations, an air data display device according to the present invention is separate from the aforementioned pitot-static system and separate from the aforementioned VTOL aircraft.

[0035] If the VTOL aircraft is an unmanned aerial vehicle, the air data display device is advantageously part of a combination of the VTOL aircraft and a remote control device for controlling the VTOL aircraft. In this case, the air data display device can be located, for example, within the VTOL aircraft or within the remote control device. In the latter case, the air data display device is therefore separate from the VTOL aircraft because it is located within the remote control device.

[0036] Alternatively, the air data display device may be manufactured and sold entirely separate from the VTOL aircraft and separate from the remote control device, such that the air data display device may be adapted to be located within the VTOL aircraft or may be adapted to be located within the remote control device, for example.

[0037] Preferably, in the method for calibrating an airborne data display device, for each training data set, reference output data corresponding to the desired output of the regressor by the VTOL aircraft of the type flying in each flight condition is obtained from one or more reference sensors of the VTOL aircraft of the type flying in each flight condition. This has the advantage that the reference output data can be obtained in a reliable and easy-to-understand manner. The one or more reference sensors are advantageously calibrated sensors disposed on the VTOL aircraft. Therefore, the one or more reference sensors may be sensors disposed on the VTOL aircraft solely for the purpose of obtaining the reference output data during test flights flying in each flight condition. After the test flights, the one or more reference sensors can be removed from the VTOL aircraft. However, in a variant, the one or more reference sensors are permanently installed on the VTOL aircraft.

[0038] It does not matter whether the reference output data is identical to data output by one or more reference sensors, or whether the reference output data is calculated from data output by one or more reference sensors. In either case, to obtain the reference output data, one or more reference sensors can be placed, for example, on the nose boom of the VTOL aircraft or on a trailing bomb towed behind the VTOL aircraft during test flights. Both such a nose boom and such a trailing bomb allow the one or more reference sensors to be placed far from the fuselage of the VTOL aircraft, providing the most accurate data possible for obtaining the reference output data. This makes the obtained data particularly accurate if the one or more reference sensors are calibrated sensors.

[0039] Depending on the desired output of the regressor during flight of the VTOL aircraft of that type under the respective flight conditions, the reference output data may be obtained from, for example, an airspeed and pressure altitude sensor that already includes correction for the effects of wind and turbulence, a dynamic and static pressure altitude sensor that already includes correction for the effects of wind and turbulence, a total pressure and static pressure altitude sensor that already includes correction for the effects of wind and turbulence, or a ground speed and climb rate sensor such as a GPS or inertial unit.

[0040] However, instead, the reference power data are obtained by calculation based on the flight data of the respective training datasets or by other methods. For example, the reference power data can be obtained by calibrating the flight data using a GPS calibration method such as that described in the paper "GPS-Based Airspeed Calibration for Rotorcraft: General-Purpose Application to All Flight Regimes" by Denis Hamel and Alex Kolarich, presented at the Vertical Flight Society's 76th Annual Forum & Technology Display, held virtually October 6-8, 2020.

[0041] Advantageously, the training data includes at least one training data set associated with at least one flight condition in the list of flight conditions, the list of flight conditions including horizontal flight at a first horizontal flight speed, climb at a first rate of climb and a first rate of climb, and descent at a first rate of descent and a first rate of descent. Thus, in a vertical climb variant, the first rate of climb is 0 knots. Furthermore, in a vertical descent variant, the first rate of descent is 0 knots. More advantageously, the training data includes at least one training data set associated with each flight condition in at least two of the list of flight conditions. Even more advantageously, the training data includes at least one training data set associated with each flight condition in at least three of the list of flight conditions. Thus, the more different flight conditions in the list of flight conditions associated with the training data include at least one training data set, the more accurately the air data display device can be calibrated.

[0042] Advantageously, the list of flight conditions further includes horizontal flight at a second horizontal flight speed, the second horizontal flight speed being greater than the first horizontal flight speed. This results in the training data including at least two training data sets, one relating to the flight condition horizontal flight at the first horizontal flight speed and the other relating to the flight condition horizontal flight at the second horizontal flight speed. This has the advantage of allowing for more accurate calibration of the air data display. This accuracy can be further enhanced if the first and second horizontal flight speeds are essentially equally spaced, for example, 100 knots and 200 knots. However, the interval between the first and second horizontal flight speeds may also vary.

[0043] Alternatively, the list of flight conditions does not include horizontal flight at a second horizontal flight speed, the second horizontal flight speed being greater than the first horizontal flight speed.

[0044] Advantageously, the list of flight conditions further includes a climb at a second climb rate and a first climb speed, the second climb rate being greater than the first climb rate. Thus, the training data includes at least two training data sets, one relating to a climb flight condition at the first climb rate and the first climb speed, and the other relating to a climb flight condition at the second climb rate and the first climb speed. This has the advantage of enabling more accurate calibration of the air data display device. This accuracy can be further improved if the first climb rate and the second climb rate are essentially equally spaced, for example, 500 feet per minute and 1,000 feet per minute. However, the interval between the first climb rate and the second climb rate may also be different.

[0045] Advantageously, the list of flight conditions includes a climb at a first climb rate and a climb at a second climb speed, the second climb speed being greater than the first climb speed. Advantageously, the training data includes at least two training data sets, one of which relates to a climb flight condition at the first climb rate and the first climb speed, and another of which relates to a climb flight condition at the first climb rate and the second climb speed. This has the advantage of enabling more accurate calibration of the air data display. This accuracy can be further increased if the first climb speed and the second climb speed are essentially equally spaced, for example, 50 knots and 100 knots. However, the interval between the first climb speed and the second climb speed may also be different.

[0046] Advantageously, the list of flight conditions further includes a climb at a second climb rate and a second climb speed. This has the advantage of allowing for more accurate calibration of the air data display. This accuracy can be further increased if the first climb rate and the second climb speed are essentially equally spaced, for example, 50 knots and 100 knots. However, the intervals between the first climb rate and the second climb rate may differ.

[0047] Alternatively, the list of flight conditions does not include climbing at a second rate of climb, the second rate of climb being greater than the first rate of climb.

[0048] Advantageously, the list of flight conditions further includes a descent at a second descent rate and a first descent flight speed, the second descent rate being greater than the first descent rate. Accordingly, the training data advantageously includes at least two training data sets, one of which is associated with a descent flight condition at the first descent rate and the first descent flight speed, and the other of which is associated with a descent flight condition at the second descent rate and the first descent flight speed. This has the advantage of enabling more accurate calibration of the air data display. This accuracy is further improved if the first descent rate and the second descent rate are essentially equally spaced, for example, 700 feet per minute and 1,400 feet per minute. However, the first descent rate and the second descent rate may be spaced apart by different intervals.

[0049] Advantageously, the list of flight conditions further includes a descent at a third descent rate and a first descent flight speed, the third descent rate being greater than the second descent rate. Accordingly, the training data advantageously includes at least three training data sets, one of which is associated with a descent flight condition at the first descent rate and the first descent flight speed, one of which is associated with a descent flight condition at the second descent rate and the first descent flight speed, and one of which is associated with a descent flight condition at the third descent rate and the second descent flight speed. This advantageously allows for even more accurate calibration of the air data display device. This accuracy is further improved if the first descent rate, the second descent rate, and the third descent rate are essentially equally spaced, for example, 700 feet per minute, 1,400 feet per minute, and 2,100 feet per minute. However, the intervals between the first descent rate, the second descent rate, and the third descent rate may be different.

[0050] The list of flight conditions advantageously includes a descent at a first descent rate and a second descent speed, the second descent speed being greater than the first descent speed. Thus, the training data advantageously includes at least two training data sets, one of which is associated with a descent at the first descent rate and the first descent speed, and the other of which is associated with a descent at the first descent rate and the second descent speed. This advantageously allows for more accurate calibration of the air data display. This accuracy is further improved when the first descent speed and the second descent speed are essentially equally spaced, for example, 50 knots and 100 knots. However, the interval between the first descent speed and the second descent speed may also be different.

[0051] Advantageously, the list of flight conditions further includes a descent at a second descent rate and a second descent speed. This allows for more accurate calibration of the air data display. This accuracy is further improved if the first and second descent speeds are essentially equally spaced, for example, 50 knots and 100 knots. However, the first and second descent speeds may be spaced apart by different amounts.

[0052] Advantageously, the list of flight conditions further includes a descent at a third descent rate and a second descent flight speed, the third descent rate being greater than the second descent rate. Advantageously, the training data includes at least three training data sets, one of which relates to a descent flight condition at the first descent rate and the second descent flight speed, one of which relates to a descent flight condition at the second descent rate and the second descent flight speed, and one of which relates to a descent flight condition at the third descent rate and the second descent flight speed. Advantageously, this allows for even more accurate calibration of the air data display device.

[0053] Particularly advantageously, the training data includes at least six training data sets, one of which is for a descent flight condition at a first descent rate and a first descent flight speed, one of which is for a descent flight condition at a second descent rate and a first descent flight speed, one of which is for a descent flight condition at a third descent rate and a first descent flight speed, one of which is for a descent flight condition at a first descent rate and a second descent flight speed, one of which is for a descent flight condition at a second descent rate and a second descent flight speed, and one of which is for a descent flight condition at a third descent rate and a second descent flight speed, which advantageously allows for even more accurate calibration of the air data display device.

[0054] However, in a variant, the list of flight conditions does not include a descent at a third rate of descent where the third rate of descent is greater than the second rate of descent.

[0055] Regardless of the descent flight speed, if the list of flight conditions includes a descent at a second descent rate, or also includes a descent at a third descent rate, the flight condition that is a descent at the maximum descent rate is advantageously associated with an unpowered descent, sometimes referred to as an autorotation descent, while the other flight condition or conditions that are descent at a descent rate are associated with a powered descent.

[0056] Alternatively, the list of flight conditions may not include a descent at a second rate of descent, where the second rate of descent is greater than the first rate of descent. Furthermore, the list of flight conditions may not include a descent at a second rate of descent. Advantageously, the first horizontal flight speed, the second horizontal flight speed, the first rate of climb, the second rate of climb, the first rate of climb, the second rate of climb, the first rate of descent, the second rate of descent, and the third rate of descent, and the first rate of descent and the second rate of descent are constant. However, as outlined below, one or more of the first horizontal flight speed, the second horizontal flight speed, the first rate of climb, the second rate of climb, the first rate of climb, the second rate of climb, the first rate of descent, the second rate of descent, the third rate of descent, the first rate of descent, and the second rate of descent may vary over time when acceleration in either direction is applied to simulate one of the corresponding flight configurations, as outlined below.

[0057] Preferably, each training data set is associated with a flight configuration and includes flight data acquired during flight in the respective flight configuration and flight conditions with which the respective training data set is associated, the flight conditions for each type of VTOL aircraft including the respective type of VTOL aircraft for which the air data display is calibrated. The training data includes at least one training data set associated with at least one of the following list of flight configurations: intermediate center of gravity, fully forward longitudinal center of gravity, fully aft longitudinal center of gravity, slightly forward longitudinal center of gravity, and slightly aft longitudinal center of gravity.

[0058] Each training data set is therefore advantageously associated with one flight condition and one flight configuration, whereby the flight condition is advantageously one of the aforementioned list of flight conditions and the flight configuration is advantageously one of the aforementioned list of flight configurations.

[0059] Advantageously, the training data includes at least one training data set associated with each flight configuration for each of at least two of the lists of flight configurations. More advantageously, the training data includes at least one training data set associated with each flight configuration for each of at least three of the lists of flight configurations. Even more advantageously, the training data includes at least one training data set associated with each flight configuration for each of at least four of the lists of flight configurations. Most advantageously, the training data includes at least one training data set associated with each flight configuration for each of the lists of flight configurations.

[0060] In a more advantageous variant, the training data includes at least one training data set associated with each flight condition and associated with each flight configuration for each of at least two of the lists of flight configurations. Even more advantageously, the training data includes at least one training data set associated with each flight condition and associated with each configuration for each of at least three of the lists of flight configurations. Even more advantageously, the training data includes at least one training data set associated with each flight condition and associated with each flight condition for each of at least four of the lists of flight configurations. Most advantageously, the training data includes at least one training data set associated with each flight condition and associated with each configuration for each of the lists of flight configurations.

[0061] If the flight configuration is a neutral center of gravity, the VTOL aircraft may be physically configured in the neutral center of gravity configuration, or the VTOL aircraft may be physically configured in a different configuration, such as a slightly aft center of gravity configuration, and the neutral center of gravity configuration may be simulated by flying at a constant forward acceleration to provide a pitch attitude angle equal to the neutral center of gravity configuration. This simulation of the neutral center of gravity configuration may be achieved under all flight conditions listed above.

[0062] If the flight configuration is a fully forward longitudinal CG, the VTOL aircraft can be physically configured in the fully forward longitudinal CG configuration, or the VTOL aircraft can be physically configured in a different configuration, such as an intermediate CG configuration, to simulate a fully forward longitudinal CG while flying at a constant forward acceleration where the pitch attitude angle is equal to the fully forward longitudinal CG configuration. This simulation of a fully forward longitudinal CG configuration can be achieved under all flight conditions listed above.

[0063] If the flight configuration is a fully aft longitudinal center of gravity, the VTOL aircraft can be physically configured in the fully aft longitudinal center of gravity configuration, or the VTOL aircraft can be physically configured in a different configuration, such as an intermediate center of gravity configuration, to simulate a fully aft longitudinal center of gravity and fly at a constant aft acceleration such that the pitch attitude angle is equal to the fully aft longitudinal center of gravity configuration. This simulation of a fully aft longitudinal center of gravity configuration can be achieved in all flight conditions in the list of flight conditions above.

[0064] If the flight configuration is a slightly forward longitudinal center of gravity, the VTOL aircraft can be physically configured in the slightly forward longitudinal center of gravity configuration, or the VTOL aircraft can be physically configured in a different configuration, such as an intermediate center of gravity configuration, to simulate a slightly forward longitudinal center of gravity while flying at a constant forward acceleration where the pitch attitude angle is equal to the slightly forward center of gravity configuration. This simulation of a slightly forward longitudinal center of gravity configuration can be achieved in all flight configurations in the list of flight conditions above.

[0065] If the flight configuration is a slightly aft longitudinal CG, the VTOL aircraft can be physically configured in the slightly aft longitudinal CG configuration, or the VTOL aircraft can be physically configured in a different configuration, such as an intermediate CG configuration, to simulate the slightly aft longitudinal CG configuration while flying at a constant aft acceleration where the pitch attitude angle is equal to the slightly aft longitudinal CG configuration. This simulation of the slightly aft longitudinal CG configuration can be achieved in all flight conditions listed above.

[0066] Simulating one configuration in the list of configurations while the VTOL aircraft is physically configured in another configuration in the list of configurations has the advantage that training data sets associated with both configurations can be collected in one and the same test flight. Thus, it is particularly advantageous to acquire or record training data with the VTOL physically configured in one configuration in the list of configurations, such as a mid-center-of-gravity configuration, while simulating other configurations in the list of configurations to acquire or record training data sets associated with other configurations in the list of configurations other than the one in which the VTOL aircraft is physically configured.

[0067] However, in other forms, the training data may only include a training data set relating to one flight configuration, and in particular one flight configuration from a list of flight configurations.

[0068] Advantageously, the regressor is a neural network regressor, in particular a fully connected neural network regressor including an output layer.

[0069] Advantageously, the neural network regressor is a fully connected feedforward neural network regressor, which has the advantage that neural network regressors are easier to work with because information only moves forward in the network, although alternatively the neural network regressor may be a recurrent neural network regressor.

[0070] The output layer advantageously has a tanh transfer function. This has the advantage that the output of the transfer function provides values ​​between -1 and +1, which makes it possible to apply penalties to nodes rather than just preventing them from activating. This increases the range of the output and therefore allows for more accurate calibration of the air data display. Furthermore, the derivative of the tanh is simple to calculate.

[0071] However, the output layer may alternatively have a transfer function other than the hyperbolic tangent (tanh).

[0072] Advantageously, the regressor includes at least two hidden layers, which has the advantage that the air data display device can be calibrated more accurately than if the regressor included only one hidden layer, or no hidden layers at all.

[0073] However, alternatively, the regressor may include only one hidden layer or no hidden layers.

[0074] Preferably, each of the at least two hidden layers contains at least 32 neurons, more preferably at least 60 neurons, and most preferably at least 100 neurons. Because each of the at least two hidden layers contains at least 32 neurons, good calibration of the air data display device can be achieved. However, the greater the number of neurons contained in each of the at least two hidden layers, the more accurate the achievable calibration of the air data display device. However, the greater the number of neurons contained in each of the at least two hidden layers, the greater the computational power required to calibrate the air data display device. Therefore, it is advantageous for each of the at least two hidden layers to contain fewer than 300 neurons. Nevertheless, each of the at least two hidden layers may contain fewer than 32 neurons, 300 neurons, or more than 300 neurons.

[0075] Preferably, the first hidden layer of the at least two hidden layers has a relu transfer function, which has the advantage of allowing reliable calibration of the air data display device with relatively little computational power.

[0076] Advantageously, all but the last of the at least two hidden layers have a relu transfer function, which has the advantage that a reliable calibration of the air data display device is possible with a relatively small amount of computational power, which is particularly noticeable when the regressor includes three or more hidden layers.

[0077] Advantageously, all hidden layers of the at least two hidden layers have a relu transfer function, which has the advantage that a reliable calibration of the air data display device can be achieved with a relatively low computational effort. However, in a variant, the last hidden layer of the at least two hidden layers has a continuously differentiable transfer function, in particular a tanh transfer function, which has the advantage that a more accurate calibration of the air data display device can be achieved.

[0078] In all of the above variants using a relu transfer function, a transfer function different from the relu transfer function can be employed. For example, in the above variants using a relu transfer function, the transfer function is not a relu transfer function but a gelu transfer function or an SiLu transfer function.

[0079] Preferably, in each training data set, each type of flight data (e.g., pitot tube data, static port data, etc.) represents values ​​obtained by filtering a length of at least 0.5 seconds, particularly preferably at least 1 second, and even more preferably at least 2 seconds, of the data stream of each type of flight data measured at each flight condition and, if applicable, in each flight configuration. This has the advantage that each training data set contains reliable and representative values ​​for each type of flight data that are representative of the flight conditions and flight configuration to which each training data set pertains. In these variations, in each training data set, each type of flight data (pitot tube data, static port data, etc.) represents values ​​obtained by filtering no more than 10 seconds of the data stream of each type of flight data measured at each flight condition and, if applicable, in each flight configuration.

[0080] Thus, when filtering the data stream to obtain each value of each type of flight data, the filter applied may, in one example, be an average value of the data stream of each type of flight data calculated over the filtered length of the data stream, or, in another example, a low pass filter applied to the data stream of each type of flight data over the filtered length of the data stream, with the output value of the low pass filter being, for example, a value of 2 Hz.

[0081] In another aspect, in each training data set, the data for each type of flight data (e.g., pitot tube data, static port data, etc.) is a value obtained by filtering a data stream of the corresponding type of flight data measured at each flight condition and, if applicable, in each flight configuration for a length less than 0.5 seconds or more than 10 seconds. In yet another aspect, the training data set does not include, for each type of flight data, a value obtained by filtering the data stream of each type of flight data.

[0082] Advantageously, the calibration is validated with further training data, which has the advantage of allowing the safety of the calibration to be controlled. To achieve this, advantageously, about 80% to 90% of the training data is used for calibration, and the remaining about 20% to 10% of the training data is used for validation.

[0083] However, alternatively, the method is performed without such verification.

[0084] Other advantageous embodiments and feature combinations will become apparent from the following detailed description and the claims as a whole. [Brief explanation of the drawings]

[0085] The drawings used to explain the embodiments are shown below. [Figure 1] 1 is a simplified schematic diagram of a pitot-static system for a vertical take-off and landing (VTOL) aircraft, particularly a helicopter, for indicating the airspeed of the VTOL aircraft and the altitude of the VTOL aircraft to a pilot of the VTOL aircraft, the pitot-static system including an air data display device according to the present invention. [Figure 2] 1 is a simplified schematic side view of the front of a helicopter as an example of a VTOL aircraft having an air data display device according to the present invention; [Figure 3a] FIG. 1 shows data recorded during two test flights with a helicopter from which training data was obtained, used to calibrate an air data display device according to the invention with a method according to the invention. [Figure 3b] FIG. 1 shows data recorded during two test flights with a helicopter from which training data was obtained, used to calibrate an air data display device according to the invention with a method according to the invention. [Figure 4] FIG. 10 shows data recorded during a third test flight verifying the calibration of an air data display device according to the present invention. [Figure 5] 10A-10D are detailed views of data from three maneuvers flown during the third test flight to show in more detail how well the method according to the invention and the air data display device according to the invention work. DETAILED DESCRIPTION OF THE INVENTION

[0086] In the drawings, like elements are designated by like reference numerals. Figure 1 shows a simplified schematic diagram of a pitot-static system 100 for a vertical take-off and landing (VTOL) aircraft, particularly a helicopter, that indicates the airspeed and altitude of the VTOL aircraft to a pilot of the VTOL aircraft. The pitot-static system 100 includes a pitot tube device 51 that determines stagnation air pressure at the location of the pitot tube device 51 and provides pitot tube data that includes information about the stagnation air pressure at the location of the pitot tube device 51. Additionally, the pitot-static system 100 includes a static port device 52 that determines static pressure at the location of the static port device 52 and provides static port data that includes information about the static pressure at the location of the static port device 52. Additionally, the pitot-static system 100 includes an air data display device 1 according to the present invention for a VTOL aircraft, particularly a helicopter, that provides information about the airspeed of the VTOL aircraft and provides information about the altitude of the VTOL aircraft. Pitot-static system 100 further includes a display 101 that displays the airspeed of the VTOL aircraft and the altitude of the VTOL aircraft, and thus indicates the airspeed and altitude of the VTOL aircraft to the pilot of the VTOL aircraft.

[0087] When the pitot-static system 100 is implemented in a VTOL aircraft, the pitot-static system is included in the VTOL aircraft's Air Data Attitude Heading Reference System (ADAHRS). Thus, the pitot tube device 51 and static port device 52 of the pitot-static system 100 are mounted on the VTOL aircraft and therefore belong to the VTOL aircraft. Furthermore, the display 101 is arranged to display the airspeed and altitude of the VTOL aircraft to the pilot of the VTOL aircraft. Thus, if the pilot is to sit in the cockpit of the VTOL aircraft, the display 101 is arranged in the cockpit of the VTOL aircraft and therefore belongs to the VTOL aircraft. However, if the VTOL aircraft is remotely controllable, the display 101 may be arranged in a remote control device that controls the VTOL aircraft. In this case, the display belongs to the remote control device.

[0088] When pitot-static system 100 is implemented in a VTOL aircraft, air data display device 1 is connected to pitot tube device 51 to receive pitot tube data provided therefrom, and is connected to static port device 52 to receive static port data provided therefrom. Furthermore, air data display device 1 is connected to display 101 to provide information about the airspeed of the VTOL aircraft and information about the altitude of the VTOL aircraft to display on display 101 to the pilot. Thus, for example, air data display device 1 can be disposed within the VTOL aircraft and thus be part of the VTOL aircraft, or, if the VTOL aircraft is remotely controlled, air data display device 1 can be disposed within the remote control device. In either case, if the VTOL aircraft is equipped with an autopilot for automatically flying the VTOL aircraft, air data display device 1 is connected to the autopilot to provide information about the airspeed of the VTOL aircraft and information about the altitude of the VTOL aircraft to the autopilot of the VTOL aircraft.

[0089] An air data display device 1 according to the present invention can be manufactured and sold separately from the VTOL aircraft and the pitot-static system 100. However, because the air data display device 1 can be part of the VTOL aircraft and part of the pitot-static system 100 as described above, the air data display device 1 can be connected to the pitot tube device 51 to receive pitot tube data provided by the pitot tube device 51, and can be connected to the static port device 52 to receive static port data provided by the static port device 52.

[0090] According to the present invention, the airdata display device 1 includes an airspeed and altitude determination module 2. The airspeed and altitude determination module 2 is adapted to determine the airspeed of the VTOL aircraft and the altitude of the VTOL aircraft in real time from the flight data using regressors 3 obtained by training an artificial neural network with training data. In one example, the airspeed and altitude determination module 2 is a computer program product running on a computing unit, such as a control computer of the VTOL aircraft, or on a computing unit separate from the control computer of the VTOL aircraft. In another example, the airspeed and altitude determination module 2 is a computing unit, such as a computer onboard the VTOL aircraft or integrated into a remote control device, adapted to determine the airspeed of the VTOL aircraft and the altitude of the VTOL aircraft in real time from the flight data using regressors 3 obtained by training an artificial neural network with training data.

[0091] The flight data includes at least pitot tube data, static port data, vertical speed data including information regarding the vertical speed of the VTOL aircraft, and pitch attitude angle data including information regarding the pitch attitude angle of the VTOL aircraft.

[0092] Because the pitot tube data includes information about stagnation air pressure at the location of the pitot tube device 51, the pitot tube data may be, for example, the stagnation air pressure measured by the pitot tube device 51 in any units output by the pitot tube device 51, an uncorrected airspeed calculated based on the measured stagnation air pressure and the measured static pressure, or a pre-corrected airspeed. This pre-corrected airspeed may be an uncorrected airspeed corrected with a known position correction, for example, from a look-up table.

[0093] Because the static port data includes information about the static pressure at the location of the static port device 52, the static port data may be, for example, the static pressure measured at the static port device 52 in any units output by the static port device 52, an uncorrected altitude calculated based on the measured static pressure, or a pre-corrected altitude. The pre-corrected altitude may be, for example, an uncorrected altitude corrected with a known position correction from a look-up table.

[0094] Because the vertical velocity data contains information about the vertical velocity of the VTOL aircraft, the vertical velocity data can be obtained from the static pressure port data using, for example, the change in the value of the static pressure port data over time. This can be, for example, the change in static pressure measured by static pressure port device 52 over time in any units output by static pressure port device 52, the change in uncorrected altitude calculated based on the measured static pressure, or the change in pre-corrected altitude over time. In either case, this can be the change per predefined time unit.

[0095] Since the pitch attitude angle data includes information about the pitch attitude angle of the VTOL aircraft, the pitch attitude angle data can be obtained, for example, from an Air Data Attitude Heading Reference System (ADAHRS) of the VTOL aircraft, in particular from a gyroscopic flight instrument or a Micro-Electro-Mechanical System (MEMS) gyroscope. To receive the pitch attitude angle data, the air data display device 1 is advantageously connectable to a pitch attitude angle data providing unit that provides the pitch attitude angle data. This pitch attitude angle data providing unit can be, for example, a gyroscopic flight instrument or a Micro-Electro-Mechanical System (MEMS) gyroscope. However, the pitch attitude angle data providing unit can also be a separate unit or a computer that receives pitch attitude angle data from the aforementioned gyroscopic flight instrument or Micro-Electro-Mechanical System (MEMS) gyroscope, respectively. Thus, the pitch attitude angle data providing unit can be, for example, part of the Air Data Attitude Heading Reference System (ADAHRS) of the VTOL aircraft.

[0096] In a variant, the flight data further includes roll attitude angle data containing information about the roll attitude angle of the VTOL aircraft. Since the roll attitude angle data includes information about the roll attitude angle of the VTOL aircraft, the roll attitude angle data can be obtained, for example, from an Air Data, Attitude, Heading and Reference System (ADAHRS) of the VTOL aircraft, in particular from a gyroscopic flight instrument or a micro-electromechanical system (MEMS) gyroscope. To receive the roll attitude angle data, the air data display device is advantageously connectable to a roll attitude angle data providing unit that provides the roll attitude angle data. Thereby, the roll attitude angle data providing unit may be the same unit as the aforementioned pitch attitude angle data providing unit or may be separate from the aforementioned pitch attitude angle data providing unit. The roll attitude angle data providing unit may be, for example, a gyroscopic flight instrument or a micro-electromechanical system (MEMS) gyroscope. However, the roll attitude angle data providing unit may also be yet another unit or a computer that receives roll attitude angle data from the aforementioned gyroscopic flight instrument or the micro-electromechanical system (MEMS) gyroscope, respectively. Thus, the roll attitude angle data providing unit may be part of an Air Data Attitude Heading Reference System (ADAHRS) of a VTOL aircraft, for example.

[0097] In a further variation, the flight data further includes sideslip angle data containing information regarding the sideslip angle of the VTOL aircraft. In a variation that allows for more accurate calibration of the air data device, the flight data includes roll attitude angle data containing information regarding the roll attitude angle of the VTOL aircraft, and sideslip angle data containing information regarding the sideslip angle of the VTOL aircraft. Because the sideslip angle data contains information regarding the sideslip angle of the VTOL aircraft, the sideslip angle data can be obtained, for example, from a dedicated sensor. For example, the sideslip angle can be calculated using a dedicated sideslip sensor that utilizes the total pressure measured at two additional static pressure ports from the VTOL aircraft, one facing right and one facing left.

[0098] In yet another variant, the flight data further includes information regarding the lateral acceleration of the VTOL aircraft, in particular lateral acceleration data including an amount of lateral acceleration of the VTOL aircraft. Since the lateral acceleration data includes information regarding the lateral acceleration of the VTOL aircraft, the lateral acceleration data can be obtained, for example, from an Air Data Attitude Heading Reference System (ADAHRS) of the VTOL aircraft, in particular one or more accelerometers. In order to receive the lateral acceleration data, the air data display device is advantageously connectable to a lateral acceleration data providing unit that provides the lateral acceleration data.

[0099] In a variation that allows for more accurate calibration of the air data device, the flight data includes roll attitude angle data containing information regarding the roll attitude angle of the VTOL aircraft, and lateral acceleration data containing information regarding the lateral acceleration of the VTOL aircraft, particularly the amount of lateral acceleration of the VTOL aircraft. In a further variation that also allows for more accurate calibration of the air data device, the flight data includes information regarding the lateral acceleration of the VTOL aircraft, particularly the amount of lateral acceleration of the VTOL aircraft, and sideslip angle data containing information regarding the sideslip angle of the VTOL aircraft. In a third variation that allows for more accurate calibration of the air data device, the flight data includes roll attitude angle data containing information regarding the roll attitude angle of the VTOL aircraft, sideslip angle data containing information regarding the sideslip angle of the VTOL aircraft, and lateral acceleration data containing information regarding the lateral acceleration of the VTOL aircraft, particularly the amount of lateral acceleration of the VTOL aircraft.

[0100] FIG. 2 shows a simplified schematic side view of the front of a helicopter 50 as an example of a VTOL aircraft equipped with an air data display device 1 according to the present invention. The helicopter 50 is equipped with a pitot-static system 100, thus including an air data display device 1 according to the present invention. Accordingly, a pitot tube device 51 is installed on the right side of the front of the helicopter 50, while a static port device 52 is installed on each side of the front of the helicopter. These static port devices 52 are interconnected. Both the pitot tube device 51 and the static port device 52 are arranged and positioned as known in the art. The pitot tube device 51 determines stagnation air pressure at the location of the pitot tube device 51 and provides pitot tube data that includes information regarding the stagnation air pressure at the location of the pitot tube device 51. The interconnected static port device 52 determines static pressure at the location of the static port device 52 and provides static port data that includes information regarding the static pressure at the location of the static port device 52. The air data display device 1 is connected to the pitot tube device 51 to receive pitot tube data provided by the pitot tube device 51, and is connected to the static port device 52 to receive static port data provided by the static port device 52. Furthermore, the air data display device 1 is adapted to obtain vertical speed data containing information regarding the vertical speed of the helicopter from the static port data by determining changes in the static port data over time.

[0101] The helicopter 50 further includes an Air Data Attitude Heading Reference System (ADAHRS) 56, as is known in the art. The Air Data Display Device 1 is connected to the ADAHRS 56 and receives pitch attitude angle data determined and provided by the ADAHRS 56.

[0102] In the helicopter 50, the air data display device 1 is connected to a display 101 (not shown in FIG. 2) of the pitot-static system 100, and information regarding the airspeed and altitude of the helicopter 50 is provided to the display 101, which displays the airspeed and altitude to the pilot of the helicopter 50. Furthermore, the air data display device 1 is connected to an autopilot of the helicopter 50, and information regarding the airspeed and altitude of the helicopter 50 is provided to the autopilot. Thus, the helicopter 50 shown in FIG. 2 can be flown by both the pilot and the autopilot using the air data display device 1 calibrated by the method of the present invention.

[0103] To obtain reference power data for training data during test flights, the helicopter 50 was fitted with a nose boom 53, as shown in FIG. 2 . A reference pitot tube assembly 54 and a reference static pressure port assembly 55 were attached to the tip of the nose boom 53. Due to the nose boom 53, the reference pitot tube assembly 54 and the reference static pressure port assembly 55 were located far from the fuselage of the helicopter 50 in a non-critical pressure field to minimize the effects of wind currents and turbulence around the helicopter 50, which could corrupt the stagnation air pressure and static pressure measured relative to the reference pitot tube assembly 54 and the reference static pressure port assembly 55. From the data obtained from the reference pitot tube assembly 54 and the reference static pressure port assembly 55, a corrected reference airspeed was calculated as known in the art, and from the reference static pressure port assembly 55, a corrected reference static pressure was calculated as known in the art. Thus, both the corrected reference airspeed and the corrected reference static pressure were calibrated as known in the art. The reference pitot tube assembly 54 and the reference static port assembly 55 were located at the tip of the nose boom 53 and were calibrated as known in the art so that they were known to provide sufficient values ​​for all flight configurations and flight conditions within the accuracy required by regulatory specifications. Instead of locating the reference pitot tube assembly 54 and the reference static port assembly 55 at the tip of the nose boom 53, they could have been mounted on a trailing bomb to obtain reference power data, which would also have provided reference power data within the accuracy required by regulatory specifications.

[0104] Training data for training the neural network that derives the regressor 3 for calibrating the air data display device 1 was acquired from the helicopter 50 by flying the helicopter 50 on two test flights. Data recorded during these two test flights is shown in FIGS. 3a and 3b. The flight data for the training data includes pitot tube data, static port data, vertical speed data containing information about the vertical speed of the helicopter 50, and pitch attitude angle data containing information about the pitch attitude angle of the helicopter 50. These flight data were obtained from the pitot tube device 51, static port device 52, and ADAHRS 56, as previously described. In the embodiment shown here, the flight data does not include roll attitude angle data, sideslip data, or lateral acceleration data. If one, two, or all three of the roll attitude angle data, sideslip data, and lateral acceleration data were added to the flight data, the calibration of the air data display device 1 would be even more accurate than shown in the embodiment of the present invention.

[0105] FIG. 3a shows data recorded during the first of two test flights by helicopter 50, and FIG. 3b shows data recorded during the second of two test flights by helicopter 50, from which the training data was derived. In both FIGS. 3a and 3b, six charts show the data recorded during each test flight as a function of the duration of each test flight in seconds.

[0106] In both Figures 3a and 3b, the top chart shows the uncorrected airspeed in knots calculated from the Pitot tube data and the static port data obtained from the Pitot tube device 51 and the static port device 52 using the following equation:

[0107]

number

[0108] where IAS is the uncorrected airspeed shown on the chart, pSSL is the standard sea level pressure (=101325 Pa), ρ SSL is the standard sea level air density (=1.225 kg / m 3 ), γ is the specific heat ratio of air (=1.4), p t_e is the uncorrected stagnation air pressure obtained from the Pitot tube device 51, p s_e is the uncorrected static pressure obtained from the static pressure port device 52.

[0109] In both Figures 3a and 3b, the second chart shows vertical speed in feet per minute calculated from uncorrected static pressure. Additionally, in both figures, the third chart shows pitch attitude angle data obtained from the ADAHRS 56, and the fourth chart shows uncorrected altitude in feet calculated from static pressure obtained from the static port device 52. Thus, uncorrected altitude is calculated from the static pressure p obtained from the static port device 52. s_e It was calculated using the following equation 2.

[0110]

number

[0111] where p SSL is the standard sea level pressure (=101325 Pa), a is the standard sea level lap rate (=0.001982 K / ft), g c = 32.17 lbm / slug, g is the acceleration due to gravity (= 31.174049 ft / sec 2 )R=96.0340(ft·lbf) / (lbm·K).

[0112] In both Figures 3a and 3b, the two lower charts show the corrected reference airspeed in knots and the corrected reference altitude in feet, obtained from the reference pitot tube device 54 and the reference static port device 55, respectively, on the nose boom 53 of the helicopter 50. The corrected reference airspeed and corrected reference altitude are then verified using known GPS calibration methods. Ideally, this GPS calibration method uses only GPS data recorded from flights in windless conditions, corrected once for air density, and used as calibration data along with vertical speed. However, in this case, i.e., when windy, the GPS quasi-static head and tailwind method (quasi-static GPS method) was applied to remove the effects of wind. Methods for calibrating air data displays are described, for example, in the paper "GPS-Based Airspeed Calibration for Rotorcraft: General Application to All Flight Regimes" by Denis Hamel and Alex Kolarich, presented at the Vertical Flight Society's 76th Annual Forum & Technology Display, held virtually from October 6th to 8th, 2020, as mentioned at the beginning.

[0113] During both test flights, the helicopter 50 was physically in an intermediate center of gravity configuration. Various maneuvers were performed during the test flights, ranging from level flight at different constant airspeeds, climbs at different constant rates of climb, and descents at different rates of descent, to unpowered descents with autorotation. Furthermore, these maneuvers were flown with different forward and aft accelerations to simulate various flight configurations with different centers of gravity, ranging from a fully forward longitudinal center of gravity configuration to a fully aft longitudinal center of gravity configuration. In this way, a full range of flight conditions and configurations was flown. Care was taken to cover at least two evenly spaced rates of climb and at least three evenly spaced rates of descent, with the fastest rate of descent being an autorotation, while the slower rate of descent was flown with powered descent.

[0114] A total of 6,400 training datasets were automatically extracted from the data shown in Figures 3a and 3b. For each training dataset, flight data and corresponding reference output data were extracted by averaging two-second data streams for each data type. Therefore, for each training dataset, each type of flight data (pitot tube data, static pressure port data, vertical speed data, and pitch attitude angle data) and each type of reference output data (i.e., reference airspeed and reference static pressure) were obtained by averaging and filtering two seconds of the data stream for each type of flight data measured under each flight condition and in each flight configuration. This resulted in a 50% overlap between the two-second time windows of adjacent training datasets. Alternatively, the time window over which each type of data stream is filtered by averaging can be selected to be different from two seconds. For example, this time window can be selected to be 0.5 seconds, 5 seconds, or 10 seconds. Furthermore, instead of filtering by averaging the data stream for each type of data, another filtering method, such as low-pass filtering, can be employed.

[0115] The described training data was used to train the regressor 3, calibrate the airspeed and altitude determination module 2, and calibrate the air data display device 1. The regressor 3 is a fully connected neural network regressor that includes two hidden layers and an output layer. The transfer functions of the two hidden layers are relu transfer functions, and the transfer function of the output layer is a hyperbolic tangent (tanh) function. Each of the two hidden layers has 256 neurons. The regressor 3 used to calculate the corrected airspeed (CAS) can be expressed as follows:

[0116]

number

[0117] where C1, C2, and C3 are matrices with the weights of the first hidden layer, the second hidden layer, and the output layer, respectively, and I1, I2, and I3 are constant vectors of the first hidden layer, the second hidden layer, and the output layer, respectively.

[0118] The air data display 1 is calibrated as described above by training the regressor 3 using the training data. To verify the calibration of the air data display 1, a third test flight was conducted using the helicopter 50. For this purpose, the helicopter 50 was again physically in an intermediate center of gravity configuration. During the test flight, various maneuvers were performed, ranging from level flight at different constant airspeeds, climbs at different constant climb rates, and descents at different descent rates to an unpowered descent with autorotation. Additionally, these maneuvers were flown with different forward and aft accelerations to simulate various flight configurations with different centers of gravity, ranging from a fully forward longitudinal center of gravity configuration to a fully aft longitudinal center of gravity configuration. In this manner, a full range of flight conditions and flight configurations was flown.

[0119] Data recorded during this third test flight is shown in Figure 4. Accordingly, Figure 4 shows six charts illustrating data recorded during the third test flight as a function of the time in seconds of the third test flight, in a similar arrangement to that shown in Figures 3a and 3b for the first two test flights. The top chart, similar to Figures 3a and 3b, shows uncorrected airspeed in knots calculated from the pitot tube data and static port data obtained from the pitot tube device 51 and static port device 52. Additionally, a second chart shows vertical speed in feet per minute calculated from the uncorrected static pressure, a third chart shows pitch attitude angle data obtained from the ADAHRS 56, and a fourth chart shows uncorrected altitude in feet calculated from the static pressure obtained from the static port device 52 in the same manner as shown in Figures 3a and 3b.

[0120] However, in contrast to Figures 3a and 3b, Figure 4 shows a comparison of airspeed in knots determined by various methods and a comparison of altitude in feet determined by various methods, respectively, in the bottom two charts. Here, the dashed lines are the corrected reference airspeed and corrected reference altitude in feet, respectively, obtained from a reference pitot tube device 54 and a reference static port device 55 on the nose boom 53 of the helicopter 50, as known in the art. These dashed lines therefore correspond to the reference output data of the third test flight. The values ​​of these dashed lines are known to be well within the accuracy required by regulatory specifications. The dotted lines are the airspeed corrected using a conventional GPS-based correction method known in the art, and the altitude corrected using a conventional GPS-based correction method known in the art, both calculated from pitot tube data from the pitot tube device 51 and static port data from the static port device 52, and corrected using a look-up table. This involved using the GPS-based calibration method described in the paper "GPS-Based Airspeed Calibration for Rotorcraft: General Application to All Flight Regimes" by Denis Hamel and Alex Kolarich, presented at the Vertical Flight Society's 76th Annual Forum & Technology Display, held virtually October 6-8, 2020. The dotted lines therefore serve as a comparison with known, conventional correction methods. The solid lines represent the airspeed determined using a calibrated air data display device 1 and thus determined with the aid of trained regressor 3, and the altitude determined using a calibrated air data display device 1. That is, the solid lines were determined as described above with the aid of trained regressor 3 from flight data including pitot tube data from pitot tube device 51, static port data from static port device 52, vertical speed data, and pitch attitude angle data.

[0121] As can be seen from Figure 4, in the bottom two charts, the solid line (from the calibrated air data display device 1) and the dashed line (reference output data) are in excellent agreement, while the dotted line (from the conventional correction) deviates the most from the other two. The residual airspeed error is generally close to zero, reaching ±2 knots in the worst case at the extremes of flight conditions and rotorcraft configurations. This demonstrates that the method for calibrating the air data display device 1 according to the present invention works very well. In particular, the third test flight covered the full range of flight conditions and flight configurations. After determining the regressor 3 as described above, the nose boom 53 can now be removed from the helicopter 50. This is because the air data display device 1, based on the pitot tube data from the pitot tube device 51 and the static port data from the static port device 52, can operate without the nose boom 53 to obtain the airspeed and altitude of the helicopter 50. Furthermore, the regressor 3 can also be used in other helicopter air data displays of the same type of helicopter as the helicopter 50 to which the pitot tube device 51 and static port device 52 are attached.

[0122] To more fully illustrate how well the method and air data display device 1 according to the present invention perform, Figure 5 shows detailed views of data during three maneuvers flown during the third test flight. These charts represent enlarged clippings of the data shown in Figure 4. The six charts shown on the left side of Figure 5 show a powered descent with a fully forward CG configuration, while the six charts shown in the center of Figure 5 show data during an autorotational descent with a fully forward CG configuration. Additionally, the six charts on the right side of Figure 5 show a climb first with a fully forward CG configuration, followed by a climb with a fully aft CG configuration.

[0123] In summary, it should be noted that an air data display device related to the technical field mentioned at the outset and a method for calibrating such an air data display device are provided, which enables the air data display device to provide correct values ​​of airspeed and altitude for all flight configurations and flight conditions with the accuracy required by regulatory specifications, and which also enables a cheaper and safer design procedure for designing VTOL aircraft.

Claims

1. An air data display device (1) for providing information about the airspeed of a vertical take-off and landing aircraft, in particular for a helicopter (50), and information about the altitude of said vertical take-off and landing aircraft, comprising: The vertical take-off and landing aircraft comprises: a Pitot tube device (51) for determining stagnation air pressure at its location and providing Pitot tube data including information regarding the stagnation air pressure at said location; a static pressure port device (52) for determining a static pressure at its location and providing static pressure port data including information regarding the static pressure at its location; the air data display device (1) is connectable to the pitot tube device (51) to receive the pitot tube data provided by the pitot tube device (51), and connectable to the static port device (52) to receive the static port data provided by the static port device (52); The airdata display device (1) includes an airspeed and altitude determination module (2) adapted to determine the airspeed of the VTOL aircraft and the altitude of the VTOL aircraft in real time from flight data including at least a) the pitot tube data, b) the static port data, c) vertical speed data including information about the vertical speed of the VTOL aircraft, and d) pitch attitude angle data including information about the pitch attitude angle of the VTOL aircraft, by using a regressor (3) obtained by training an artificial neural network with training data.

2. 2. An air data display device (1) according to claim 1, characterized in that the flight data includes roll attitude angle data containing information about the roll attitude angle of the VTOL aircraft.

3. 2. An air data display device (1) according to claim 1, characterized in that the flight data includes sideslip angle data containing information regarding the sideslip angle of the VTOL aircraft.

4. 2. The air data display device (1) of claim 1, wherein the flight data includes lateral acceleration data containing information about the lateral acceleration of the VTOL aircraft, in particular the amount of lateral acceleration of the VTOL aircraft.

5. A vertical take-off and landing aircraft comprising an air data display device (1) according to any one of claims 1 to 4, in particular including a helicopter (50).

6. the airspeed and altitude determination module (2) of the air data display device (1) is calibrated by obtaining the regressor (3) by training the neural network with training data comprising a training data set; Each training data set is associated with a flight condition and includes flight data acquired during flights in the respective flight conditions with a type of vertical take-off and landing aircraft, in particular a helicopter (50), for which the air data display device (1) is calibrated. each training data set includes reference output data corresponding to a desired output of the regressor (3) for a VTOL aircraft of said type while flying said VTOL aircraft in each of said flight conditions; The flight data includes at least a) Pitot tube data obtained from the Pitot tube device (51) of a vertical take-off and landing aircraft of the type during flight in each flight condition by the vertical take-off and landing aircraft of the type; b) static port data obtained from the static port device (52) of a vertical take-off and landing aircraft of the type during flight by the vertical take-off and landing aircraft of the type under each flight condition; and c) vertical speed data including information regarding the vertical speed of a vertical take-off and landing aircraft of the type during flight by the vertical take-off and landing aircraft of the type in each flight condition; d) pitch attitude angle data including information regarding pitch attitude angles of a vertical take-off and landing aircraft of the type during flight by the vertical take-off and landing aircraft of the type in each flight condition; A method for calibrating an air data display device (1) according to any one of claims 1 to 4, comprising:

7. 7. The method of claim 6, wherein, for each training data set, the reference output data corresponding to the desired output of the regressor (3) while flown in each of the flight conditions by the VTOL aircraft of the type is obtained from one or more reference sensors (54, 55) of the VTOL aircraft of the type while flown in each of the flight conditions.

8. the training data includes, for at least one flight condition of a list of flight conditions, at least one training data set associated with each flight condition; The list of flight conditions includes: a) horizontal flight at a first horizontal flight speed; b) climbing at a first rate of climb and a first climb speed; c) descending at a first descent rate and a first descent flight speed; 7. The method of claim 6, comprising:

9. each training data set is associated with a flight configuration and includes flight data obtained during flight in a respective flight configuration and flight condition for which each training data set is associated with said vertical take-off and landing aircraft of said type for which said air data display device (1) is calibrated; The training data is a) a middle center of gravity; b) a fully forward longitudinal center of gravity; c) a fully aft longitudinal center of gravity; d) A slight forward longitudinal center of gravity; e) A slight aft longitudinal center of gravity; 7. The method of claim 6, further comprising at least one training data set associated with at least one of a list of flight configurations including:

10. 7. The method according to claim 6, characterized in that the regressor (3) is a neural network regressor, in particular a fully connected neural network regressor including an output layer.

11. 11. The method of claim 10, wherein the output layer has a hyperbolic tangent (tanh) transfer function.

12. The method of claim 10 , wherein the regressor includes at least two hidden layers.

13. 13. The method of claim 12, wherein each of the at least two hidden layers includes at least 32 neurons.

14. 13. The method of claim 12, wherein a first hidden layer of the at least two hidden layers has a relu transfer function.

15. 7. The method of claim 6, wherein in each training data set, the data for each type of flight data is a value obtained by filtering a data stream of at least 0.5 seconds of the flight data of each type measured under each flight condition.

Images