Aircraft landing control method and system and aircraft

Abstract

The invention discloses an aircraft landing control method and system and an aircraft, and relates to the technical field of flight control. The aircraft landing control method comprises the following steps: acquiring a current scene image when an aircraft lands; comparing and analyzing the current scene image and the reference scene image to obtain an effective scene image; training the effective scene image in a machine learning model to obtain a current pitch angle and an expected pitch angle of the aircraft; whether the current pitch angle is matched with the expected pitch angle or not is judged, when the current pitch angle is not matched with the expected pitch angle, the pitch angle control link is achieved, an elevator skewness instruction is obtained, and the aircraft is stably transited to the grounding state from the descending state in the aircraft landing process. Moreover, in the leveling stage of aircraft landing, subjective judgment of a pilot is replaced with objective computer automatic judgment, and the risks of heavy landing, too long flat floating and the like are reduced.

Description

Technical Field

[0001] This application relates to the field of flight control technology, and in particular, to an aircraft landing control method, system, and aircraft. Background Art

[0002] Safe, predictable, and reliable landing is an important aspect of an effective aircraft transportation system. In many cases, relying on the pilot's visual ability to complete the descent operation is difficult to ensure accuracy because it is easily affected by lighting conditions, the pilot's personal conditions, etc. Although some external navigation systems, such as Instrument Landing System (ILS) and Global Positioning System (GPS), can be used to help the pilot complete the flare maneuver visually. However, external navigation systems are vulnerable to interruptions and interference. In addition, the availability and feasibility of dedicated equipment at different airports vary, which pose technical challenges to automatic landing and assisted driving.

[0003] Therefore, smoothly transitioning the aircraft from the descent state to the grounded state during the landing process is an urgent problem to be solved currently. Summary of the Invention

[0004] The purpose of the present invention is to provide an aircraft landing control method, system, and aircraft to achieve a smooth transition of the aircraft from the descent state to the grounded state during the landing process.

[0005] In a first aspect, this application provides an aircraft landing control method, including the following steps:

[0006] Obtain the current scene image when the aircraft lands;

[0007] Compare and analyze the current scene image with the reference scene image to obtain an effective scene image;

[0008] Apply the effective scene image to the machine learning model training to obtain the current pitch angle and the desired pitch angle of the aircraft;

[0009] Judge whether the current pitch angle and the desired pitch angle match. When the current pitch angle and the desired pitch angle do not match, implement the pitch angle control link to obtain the elevator deflection command.

[0010] In some embodiments, the step of comparing and analyzing the current scene image with the reference scene image to obtain an effective scene image includes:

[0011] Obtain the reference feature information of the reference scene image and the current feature information of the current scene image;

[0012] Compare and analyze the current feature information with the reference feature information to obtain effective feature information, and determine the effective scene image through the effective feature information.

[0013] In some embodiments, the reference feature information includes reference aircraft attitude, reference aircraft speed, and reference environmental information;

[0014] The current feature information includes current aircraft attitude, current aircraft speed, and current environmental information;

[0015] The effective feature information respectively includes effective aircraft attitude, effective aircraft speed, and effective environmental information.

[0016] In some embodiments, the steps of comparing and analyzing the current feature information with the reference feature information to obtain the effective feature information, and determining the effective scene image through the effective feature information include:

[0017] Comparing and analyzing the reference environmental information with the current environmental information to determine whether the current scene image has effective environmental information;

[0018] When the current scene image has effective environmental information, then comparing and analyzing the reference aircraft attitude with the current aircraft attitude, and the reference aircraft speed with the current aircraft speed respectively to determine whether the current scene image has effective aircraft attitude and effective aircraft speed;

[0019] When the current scene image has effective environmental information, effective aircraft attitude, and effective aircraft speed at the same time, then determining that the current scene image is an effective scene image.

[0020] In some embodiments, the steps of applying the effective scene image to machine learning model training to obtain the current pitch angle and the desired pitch angle of the aircraft include:

[0021] Obtain the current pitch angle;

[0022] According to the effective environmental information, effective aircraft attitude, and effective aircraft speed, and combining with the leveling law that the predetermined leveling trajectory should follow, calculate to obtain the desired pitch angle.

[0023] In some embodiments, before the steps of determining whether the current pitch angle and the desired pitch angle match, and when they do not match, implementing the pitch angle control link to obtain the elevator deflection command, further includes:

[0024] Implement the expected value generation link of the lift speed and lift acceleration to obtain the expected lift speed v yg (t) and the expected lift acceleration a yg (t);

[0025] According to the altitude h0 when the aircraft accesses the leveling control and the lift speed v y0 、the expected aircraft touchdown speed v yjd 、the aircraft altitude h(t) during the leveling process, and the aircraft lift speed v y(t), obtain the expected vertical speed v yg (t) and the expected vertical acceleration a yg (t):

[0026]

[0027] According to the open-loop correction amount ω of the expected pitch angle rate zeSemi and the integral coefficient K of the vertical speed control LI , obtain the open-loop integral correction amount v of the vertical speed ygSemi ;

[0028] The open-loop integral correction amount v of the vertical speed ygSemi is as follows:

[0029] v ygSemi = ω zeSemi / K LI ;

[0030] Implement the vertical speed control loop. According to the vertical speed v of the aircraft during the flare y (t), the vertical acceleration a of the aircraft during the flare y (t), the expected vertical speed v yg (t), the expected vertical acceleration a yg (t), the open-loop integral correction amount v of the vertical speed ygSemi and the control parameters K L , K LI , K LD , obtain the expected pitch angle θ g (t):

[0031] Among them, the control parameters K L , K LI , K LD are selected according to the current pitch angle control loop.

[0032] In some embodiments, it is determined whether the current pitch angle and the expected pitch angle match. When the current pitch angle and the expected pitch angle do not match, the steps of implementing the pitch angle control loop to obtain the elevator deflection command include:

[0033] Implement the current pitch angle control loop to obtain the elevator deflection command δ z (t);

[0034] According to the current pitch angle θ(t) of the aircraft, the expected pitch angle θg(t), the pitch angle rate ω of the aircraft z (t) and the corresponding control parameters K g , obtain the elevator deflection command δ z (t);

[0035]

[0036] δ z (t) serves as the input of the elevator loop, and realizes the flare control of the aircraft by controlling the elevator deflection.

[0037] In a second aspect, the present application provides an aircraft landing automatic flare control system, including an on-board detection module, which is used to obtain the current scene image when the aircraft lands and compare and analyze the current scene image with the reference scene image to obtain an effective scene image;

[0038] An on-board core network module, which is used to judge whether the current pitch angle matches the desired pitch angle;

[0039] An on-board control module, which is used to implement the pitch angle control link and obtain the elevator deflection command when the current pitch angle does not match the desired pitch angle.

[0040] In some embodiments, the on-board detection module includes:

[0041] An image acquisition unit, which is used to obtain the current scene image when the aircraft lands;

[0042] A landing posture and speed model unit, which is connected to the image acquisition unit and is used to store the reference scene image;

[0043] A data preprocessing unit, which is connected to the image acquisition unit and the landing posture and speed model unit. The data preprocessing unit is used to compare and analyze the current scene image with the reference scene image to obtain an effective scene image.

[0044] In some embodiments, the on-board detection module further includes:

[0045] A network interface unit, which is connected to the on-board core network module and is used to transmit the effective scene image to the on-board core network module.

[0046] In some embodiments, the on-board core network module includes:

[0047] A machine learning unit, which is used to train the effective scene image to obtain the current pitch angle and the desired pitch angle of the aircraft;

[0048] An integrated processing unit, which is used to judge whether the current pitch angle matches the desired pitch angle.

[0049] In a third aspect, the present application further provides an aircraft, including the above-mentioned aircraft landing control system.

[0050] The technical effect of the present invention is to provide an aircraft landing control method, system, and aircraft. By acquiring the current scene image when the aircraft lands, comparing and analyzing the current scene image with the reference scene image to obtain an effective scene image, training the effective scene image with a machine learning model, and analyzing the effective scene image to obtain the current pitch angle and the desired pitch angle. Thus, according to the difference between the current pitch angle and the desired pitch angle, the pitch angle control link is realized, and the elevator deflection command is obtained, so that the aircraft can be smoothly transitioned from the descending state to the grounded state during the landing process. Moreover, in the flare stage of the aircraft landing, this application uses objective computer automatic judgment to replace the subjective judgment of the pilot, reducing the risk of problems such as hard landing and excessive floating. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The following will, with reference to the accompanying drawings, through a detailed description of the specific embodiments of the present application, make the technical solutions and other beneficial effects of the present application obvious.

[0052] Figure 1 It is a flowchart of the aircraft landing control method provided by the embodiment of the present application.

[0053] Figure 2 It is a functional diagram of the aircraft landing control system provided by the embodiment of the present application.

[0054] Figure 3 It is a functional diagram of the airborne detection module provided by the embodiment of the present application.

[0055] Figure 4 It is a schematic structural diagram of the electronic device provided by the embodiment of the present application in the folded state.

[0056] The identification of the components in the drawings is as follows:

[0057] 100 Airborne detection module; 200 Airborne core network module; 300 Airborne control module;

[0058] 101 Image acquisition unit; 102 Landing posture and speed model unit; 103 Data preprocessing unit; 104 Network interface unit;

[0059] 201 Machine learning unit; 202 Comprehensive processing unit. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] The following will, with reference to the accompanying drawings in the embodiments of the present application, clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present application.

[0061] Safe, predictable, and reliable landings are an important aspect of an effective aircraft transportation system. Since the approach speed of an aircraft during landing is generally not less than 1.3 times the stall speed, this will result in a relatively large vertical descent speed (about -3.5 to -4.5 m / s, which does not meet the requirement of the vertical speed at the moment of touchdown required for the safe grounding of the aircraft . Therefore, it is required that when the aircraft is descending, it gradually reduces its flight path inclination angle θ and pulls up along a curved trajectory to make the flight speed vector as parallel to the ground as possible - flaring. The landing process of an aircraft generally includes several stages such as constant altitude, descent, flare, touchdown, and rollout. Among them, constant altitude, descent, flare, and rollout are necessary, and the flight speed at the end of the flare is the touchdown speed of the landing. It can be seen that the aircraft automatic flare landing system is an important flight control system during automatic landing, and the quality of its control will have an important impact on the landing quality of the aircraft.

[0062] In many cases, it is difficult to ensure the accuracy of the descent operation relying on the pilot's visual ability because it is easily affected by lighting conditions, the pilot's personal conditions, etc. Although some external navigation systems can be used, such as the Instrument Landing System (ILS) and the Global Positioning System (GPS), to help the pilot complete the flare action visually. However, external systems are vulnerable to interruptions and interferences. In addition, the availability and feasibility of the dedicated equipment at different airports vary, which pose technical challenges to automatic landing and assisted driving.

[0063] Therefore, the present application provides an aircraft landing control method, system, and aircraft. By receiving a series of images captured by a camera carried by the aircraft and applying the received images to a trained machine learning model to detect the runway or runway markings in the images. Comparing the captured image features with the on-board system data to obtain the current aircraft attitude, the lift and descent speed and acceleration of the aircraft, and combining the flare law that the predetermined flare trajectory should follow, calculating the desired speed and the desired aircraft attitude. Finally, the pitch angle control is achieved by judging whether the current pitch angle is consistent with the desired pitch angle. The following will be described in detail.

[0064] Some embodiments of the present application provide an aircraft landing control method, including the following steps:

[0065] S1. Obtain the current scene image when the aircraft lands.

[0066] The current scene image can be: when the aircraft lands, the current scene of the airport during the aircraft landing is captured by a camera carried by the aircraft.

[0067] In step S1, an infrared camera and a color camera carried on the aircraft are used to photograph the scene of the aircraft landing to obtain an image sequence. The image sequence includes at least one current image.

[0068] S2. Compare and analyze the current scene image and the reference scene image to obtain the effective scene image.

[0069] In step S2, first obtain the reference feature information of the reference scene image and the current feature information of the current scene image, then compare and analyze the current feature information with the reference feature information to obtain the effective feature information, and finally determine the effective scene image through the effective feature information.

[0070] The reference scene image can be: when the aircraft lands, the historical image associated with the current scene captured by the camera on the aircraft.

[0071] The reference scene image includes reference feature information, and the current scene image includes current feature information.

[0072] Specifically, the reference feature information includes the reference aircraft attitude, the reference aircraft speed, and the reference environment information. The current feature information includes the current aircraft attitude, the current aircraft speed, and the current environment information. The effective feature information respectively includes the effective aircraft attitude, the effective aircraft speed, and the effective environment information.

[0073] In one embodiment, the effective scene image can be determined by the following method.

[0074] Specifically, first, compare and analyze the reference environment information with the current environment information to determine whether the current scene image has effective environment information.

[0075] It should be noted that the environment information includes features such as anchor points, runways, and obstacles. The runway features include runway boundaries, runway stripes, runway entry points, runway exit points, runway centerlines, apron markings, profile markings, and parking position markings.

[0076] In this step, compare and analyze the reference environment information with the current environment information to eliminate some obviously invalid images and screen out the effective images with features such as anchor points, runways, and obstacles, that is, perform pre-screening processing on the current scene image.

[0077] Second, when the current scene image has effective environment information, then compare and analyze the reference aircraft attitude with the current aircraft attitude and the reference aircraft speed with the current aircraft speed respectively to determine whether the current scene image has an effective aircraft attitude and an effective aircraft speed.

[0078] The aircraft attitude includes features such as pitch angle. The aircraft speed includes lift speed and lift acceleration.

[0079] In this step, compare and analyze according to the reference aircraft attitude and the current aircraft attitude to determine whether the current aircraft attitude and the reference aircraft attitude match each other.

[0080] Specifically, by comparing the pixel parameters of the reference scene image with those of the current scene image one by one, it is determined whether the pixel parameters of the reference scene image exactly match those of the current scene image, or whether there are errors. The reference pitch angle is obtained through the pixel parameters of the reference scene image, and the current pitch angle is obtained through the pixel parameters of the current scene image. When the reference pitch angle exactly matches the current pitch angle, the reference aircraft speed is then continuously compared and analyzed with the current aircraft speed. When there is an error between the reference pitch angle and the current pitch angle, and the error value is within the error range, such as between ±0.5 and ±1, the reference aircraft speed can also be continuously compared and analyzed with the current aircraft speed. Conversely, if the error value exceeds the error range, there is no need to compare and analyze the reference aircraft speed with the current aircraft speed.

[0081] The reference aircraft speed is compared and analyzed with the current aircraft speed to determine whether they match each other. The reference lift speed and reference lift acceleration are obtained through the pixel parameters of the reference scene image, and the current lift speed and current lift acceleration are obtained through the pixel parameters of the current scene image. When the current lift speed exactly matches the reference lift speed and the current lift acceleration exactly matches the reference lift acceleration, it is determined that the current scene image has a valid aircraft attitude and a valid aircraft speed. When there is an error between the current lift speed and the reference lift speed and / or between the current lift acceleration and the reference lift acceleration, and the error range is between ±0.5 and ±1, it is determined that the current scene image has a valid aircraft attitude and a valid aircraft speed. Conversely, the current scene image does not have a valid aircraft attitude and a valid aircraft speed.

[0082] In another embodiment, the comparison and analysis order of "reference aircraft attitude and current aircraft attitude" and "reference aircraft speed and current aircraft speed" can be swapped, that is, by comparing whether the reference aircraft speed and the current aircraft speed match each other, and then determining whether to compare the reference aircraft attitude and the current aircraft attitude. This will not be elaborated here.

[0083] Third, when the current scene image simultaneously has valid environmental information, a valid aircraft attitude, and a valid aircraft speed, it is determined that the current scene image is a valid scene image.

[0084] Therefore, after confirming that the current scene image has valid environmental information, the aircraft attitude and aircraft speed are compared, and then it is determined whether the current scene image is a valid scene image according to the comparison result.

[0085] S3. Apply the valid scene image to machine learning model training to obtain the current pitch angle and desired pitch angle of the aircraft.

[0086] In step S3, first, directly obtain the current pitch angle in the machine learning model. Then, calculate according to the effective environmental information, effective aircraft attitude, and effective aircraft speed, and combine with the leveling law that the predetermined leveling trajectory should follow to obtain the desired pitch angle.

[0087] Specifically, in the vertical plane of the aircraft, the longitudinal movement trajectory from the glide to the actual landing point is called the leveling trajectory, and generally an exponential curve is used as the leveling trajectory. When the aircraft levels off with an exponential curve, the vertical lift-off speed of the aircraft decreases correspondingly as the altitude decreases, and the lift-off speed at each instant of the aircraft is proportional to its current altitude. If the runway plane line of the aircraft is used as the asymptote of the exponential leveling trajectory and the lift-off speed is zero, then the aircraft can only land when the time and leveling distance tend to infinity, which is obviously not allowed. Let the distance between the runway plane and the asymptote of the leveling trajectory be hc. At this time, the relationship between the desired lift-off speed of the aircraft and the altitude is shown in Equation (1):

[0088]

[0089] In Equation (1), τ is the exponential curve time constant, v yjd is the landing speed of the aircraft.

[0090] From the altitude h0 and lift-off speed v y0 when the aircraft accesses the leveling control, as well as the desired lift-off speed v yjd of the aircraft at landing, the exponential curve time constant τ can be determined, as shown in Equation (2):

[0091]

[0092] Substitute the time constant τ obtained from Equation (2) into Equation (1), then the expected value v yg of the lift-off speed during the entire leveling process is obtained. Take the derivative of the expected value of the lift-off speed with respect to time to obtain the corresponding expected value a yg (t) of the lift-off acceleration, as shown in Equation (3):

[0093]

[0094] Therefore, through the above, the generation link of the expected values of the lift-off speed and lift-off acceleration can be realized, and the expected value v yg of the lift-off speed and the expected value a yg of the lift-off acceleration are obtained; according to the altitude h0 and lift-off speed v y0 when the aircraft accesses the leveling control, the desired landing speed v yjd of the aircraft, the altitude h(t) of the aircraft during the leveling process, and the lift-off speed v y (t) of the aircraft during the leveling process, the expected value v yg(t) and the expected value of the elevator acceleration a yg (t).

[0095] According to the open-loop correction amount ω of the expected pitch rate zeSemi and the integral coefficient K of the elevator speed control LI , the open-loop integral correction amount v of the elevator speed is obtained ygSemi ;

[0096] The open-loop integral correction amount v of the elevator speed ygSemi , as shown in Equation (4):

[0097] v ygSemi = ω zeSemi / K LI .

[0098] When adopting the PID control structure, the basic idea of automatic control is based on the feedback of the deviation. Introducing the integral signal of the deviation amount can eliminate the steady-state error of the system. However, the elimination of its steady-state error requires a certain adjustment time. Since the expected value of the elevator speed of the aircraft changes with altitude during the flare process, and the actual movement of the controlled aircraft always lags behind the expected movement, the elevator speed when the actual aircraft touches down is not the expected value. Therefore, considering the safety requirements when the aircraft lands, the open-loop correction amount is introduced and enters the integral link together with the closed-loop feedback amount.

[0099] The open-loop correction amount ω of the pitch rate in Equation (4) zeSemi means how the pitch angle of the expected aircraft changes at a certain rate during the original flare process. The elevator speed of the aircraft during the glide process is generally about -3 m / s, and the allowable elevator speed when the aircraft touches down is generally -0.3 m / s to -0.6 m / s. The aircraft is in the process of gradually raising its head during the entire flare process. Also, because the movement of the controlled aircraft always lags behind the expected movement, during the flare process, the expected pitch angle should be deflected in advance at a certain angular rate on its original basis, that is, the open-loop correction amount ω of the pitch rate zeSemi is generally selected as a small positive number, such as 0.05 - 0.2 ° / s. The open-loop correction amount ω of the pitch rate zeSemi is initially selected based on the longitudinal motion characteristics of the aircraft and finally determined after verification using the full-scale mathematical model of the UAV.

[0100] To implement the elevator speed control link, according to the elevator speed v of the aircraft y (t) during the flare process, the elevator acceleration a of the aircraft y (t) during the flare process, the expected value v of the elevator speed yg (t), the expected value a of the elevator acceleration yg (t), the open-loop integral correction amount v of the elevator speed ygSemi and the control parameter K L, K LI , K LD , the desired pitch angle θ g (t) is obtained as shown in Equation (5):

[0101] θ g = K L (v yg (t) - v y (t)) + K LI t ∫ t0 (v yg (t) + v ygSemi - v y (t))dt + K LD (a yg (t) - a y (t)), where, according to

[0102] the current pitch angle control loop, the control parameters K L , K LI , K LD are selected and determined. In the integral channel of the lift speed control link, a lift speed open-loop integral correction amount v ygSemi is introduced, so that the aircraft tracks the desired lift speed under the combined action of closed-loop feedback control and open-loop correction. Under the condition that the control parameters K L , K LI , K LD are all the same, ω zeSemi = 0.1° / s is selected, that is, the desired pitch angle increases at a rate of 0.1° / s on the original basis. When applying the existing aircraft landing control method, since the actual lift speed of the aircraft lags behind the desired lift speed, there is a certain deviation between the lift speed at the moment of touchdown and the expected value. When applying the aircraft landing control method provided by the embodiments of the present application, there is almost no deviation between the two at the moment of touchdown, and the lift speed tracks the ideal curve well.

[0103] S4. Determine whether the current pitch angle and the desired pitch angle match. When they do not match, implement the pitch angle control link to obtain the elevator deflection command.

[0104] Specifically, according to the current pitch angle θ(t) of the aircraft, the desired pitch angle θ g (t), the pitch angle rate ω z (t) of the aircraft, and the corresponding control parameters K g , K ωz , the elevator deflection command δ z (t) is obtained, that is, as shown in Equation (6):

[0105] δ z (t) = K θ(θ(t) - θ g (t)) + K ωz ω z (t);

[0106] δ z (t) is used as the input of the elevator loop, and the leveling control of the aircraft is achieved by controlling the deflection of the elevator.

[0107] Therefore, for the aircraft landing control method provided in this application, during the descent of the aircraft, the stick pilot activates the automatic leveling landing system. The system starts to capture images of the runway and obtains the anchor point information on the runway. And through the method of machine learning, the current aircraft attitude, lift speed, and lift acceleration are obtained, so as to obtain the current pitch angle. At the same time, the system calculates the desired pitch angle and compares the current pitch angle with the desired pitch angle. If they are the same, the flight continues; if not, the system takes control of the elevator and adjusts the angle to the desired pitch angle, thus ensuring the accuracy of the leveling action and the safety of the aircraft's descent.

[0108] As Figure 2 shown, some embodiments of this application also provide an aircraft landing control system, including an airborne detection module 100, an airborne core network module 200, and an airborne control module 300.

[0109] The airborne detection module 100 is used to obtain the current scene image when the aircraft lands and compare and analyze the current scene image with the reference scene image to obtain the effective scene image.

[0110] As Figure 3 shown, the airborne detection module 100 includes an image acquisition unit 101, a landing posture and speed model unit 102, a data preprocessing unit 103, and a network interface unit 104.

[0111] The image acquisition unit 101 is used to obtain the current scene image when the aircraft lands. The landing posture and speed model unit 102 is connected to the image acquisition unit and is used to store the reference scene image. The data preprocessing unit 103 is connected to the image acquisition unit 101 and the landing posture and speed model unit 102. The data preprocessing unit 103 is used to compare and analyze the current scene image with the reference scene image to obtain the effective scene image. The network interface unit 104 is connected to the airborne core network module 200 and is used to transmit the effective scene image to the airborne core network module 200.

[0112] The above-mentioned current scene image can be: when the aircraft lands, the current scene of the airport when the aircraft lands is captured by the camera carried by the aircraft.

[0113] The above-mentioned reference scene image can be: when the aircraft lands, the historical image associated with the current scene captured by the camera on the aircraft.

[0114] The above reference scene image includes reference feature information, and the current scene image includes current feature information.

[0115] The reference feature information includes reference aircraft attitude, reference aircraft speed, and reference environment information. The current feature information includes current aircraft attitude, current aircraft speed, and current environment information. The effective feature information respectively includes effective aircraft attitude, effective aircraft speed, and effective environment information.

[0116] The reference feature information, the current feature information, and the effective scene image respectively include aircraft attitude, aircraft speed, and environment information. The aircraft attitude includes pitch angle, the aircraft speed includes speed and acceleration, and the environment information includes aircraft state, runway, obstacles, etc.

[0117] The data preprocessing unit first obtains the reference feature information of the reference scene image and the current feature information of the current scene image, then obtains the effective feature information through comparative analysis of the current feature information and the reference feature information, and finally determines the effective scene image through the effective feature information.

[0118] In one embodiment, the effective scene image can be determined by the following method.

[0119] Specifically, first, the reference environment information is compared and analyzed with the current environment information to determine whether the current scene image has effective environment information.

[0120] It should be noted that the environment information includes features such as anchor points, runways, and obstacles. The runway features include runway boundaries, runway stripes, runway entry points, runway exit points, runway centerlines, apron markings, profile markings, and parking position markings.

[0121] In this step, the reference environment information is compared and analyzed with the current environment information to eliminate some obviously invalid images and screen out effective images with features such as anchor points, runways, and obstacles, that is, to perform pre-screening processing on the current scene image.

[0122] Second, when the current scene image has effective environment information, the reference aircraft attitude is compared and analyzed with the current aircraft attitude, and the reference aircraft speed is compared and analyzed with the current aircraft speed respectively to determine whether the current scene image has effective aircraft attitude and effective aircraft speed.

[0123] The aircraft attitude includes features such as pitch angle. The aircraft speed includes lift speed and lift acceleration.

[0124] In this step, the reference aircraft attitude is compared and analyzed with the current aircraft attitude to determine whether the current aircraft attitude matches the reference aircraft attitude.

[0125] Specifically, by comparing the pixel parameters of the reference scene image with those of the current scene image one by one, it is determined whether the pixel parameters of the reference scene image and the current scene image are exactly the same, or whether there are errors. The reference pitch angle is obtained through the pixel parameters of the reference scene image, and the current pitch angle is obtained through the pixel parameters of the current scene image. When the reference pitch angle and the current pitch angle are exactly the same, the comparison and analysis of the reference aircraft speed and the current aircraft speed are continued. When there is an error between the reference pitch angle and the current pitch angle, and the error value is within the error range, such as between ±0.5 and ±1, the comparison and analysis of the reference aircraft speed and the current aircraft speed can also be continued. Conversely, if the error value exceeds the error range, there is no need to compare and analyze the reference aircraft speed and the current aircraft speed.

[0126] The reference aircraft speed and the current aircraft speed are compared and analyzed to determine whether they match each other. The reference lift speed and the reference lift acceleration are obtained through the pixel parameters of the reference scene image, and the current lift speed and the current lift acceleration are obtained through the pixel parameters of the current scene image. When the current lift speed and the reference lift speed, and the current lift acceleration and the reference lift acceleration are exactly the same, it is determined that the current scene image has a valid aircraft attitude and a valid aircraft speed. When there is an error between the current lift speed and the reference lift speed and / or between the current lift acceleration and the reference lift acceleration, and the error range is between ±0.5 and ±1, it is determined that the current scene image has a valid aircraft attitude and a valid aircraft speed. Conversely, the current scene image does not have a valid aircraft attitude and a valid aircraft speed.

[0127] In another embodiment, the comparison and analysis order of "reference aircraft attitude and current aircraft attitude" and "reference aircraft speed and current aircraft speed" can be reversed, that is, by comparing whether the reference aircraft speed and the current aircraft speed match each other, and then determining whether to compare the reference aircraft attitude and the current aircraft attitude. This will not be elaborated here.

[0128] Third, when the current scene image has valid environmental information, a valid aircraft attitude, and a valid aircraft speed at the same time, it is determined that the current scene image is a valid scene image.

[0129] Therefore, after confirming that the current scene image has valid environmental information, the aircraft attitude and the aircraft speed are compared, and then it is determined whether the current scene image is a valid scene image according to the comparison result.

[0130] The above airborne detection module can be devices such as an infrared camera or a color camera.

[0131] Such as Figure 4As shown, the airborne core network module 200 is used to determine whether the current pitch angle matches the desired pitch angle. The airborne core network module 200 includes a machine learning unit 201 and an integrated processing unit 202. The machine learning unit 201 is used to train the effective scene images to obtain the current information and desired information of the aircraft. The current information includes the current pitch angle, and the desired information includes the desired pitch angle. The integrated processing unit 202 is used to determine whether the current pitch angle matches the desired pitch angle.

[0132] First, directly obtain the current pitch angle in the machine learning model. Then, calculate according to the effective environmental information, effective aircraft attitude, and effective aircraft speed, and combine the leveling law that the predetermined leveling trajectory should follow to obtain the desired pitch angle.

[0133] Specifically, in the vertical plane of the aircraft, the longitudinal movement trajectory from the glide to the actual landing point is called the leveling trajectory, and generally an exponential curve is used as the leveling trajectory. When the aircraft levels off with an exponential curve, the vertical lift speed of the aircraft decreases correspondingly as the height decreases, and the lift speed at each instant of the aircraft is proportional to its current height. If the runway plane line of the aircraft is used as the asymptote of the exponential leveling trajectory and the lift speed is zero, then the aircraft can only land when the time and leveling distance tend to infinity, which is obviously not allowed. Let the distance between the runway plane and the asymptote of the leveling trajectory be hc. At this time, the relationship between the desired lift speed of the aircraft and the height is shown in Equation (1):

[0134]

[0135] In Equation (1), τ is the exponential curve time constant, and v yjd is the touchdown speed of the aircraft.

[0136] From the height h0 and lift speed v when the aircraft accesses the leveling control y0 and the desired touchdown lift speed v of the aircraft yjd the exponential curve time constant τ can be determined, as shown in Equation (2):

[0137]

[0138] Substitute the time constant τ obtained from Equation (2) into Equation (1), then the expected value v of the lift speed during the entire leveling process is obtained yg (t). Take the derivative of the expected value of the lift speed with respect to time to obtain the corresponding expected value a of the lift acceleration yg (t), as shown in Equation (3):

[0139]

[0140] Therefore, through the above, the generation link of the expected values of the lifting speed and the lifting acceleration can be achieved, and the expected value of the lifting speed v yg (t) and the expected value of the lifting acceleration a yg (t).

[0141] Specifically, according to the altitude h0 and the lifting speed v y0 at the time when the aircraft accesses the flare control, the expected touchdown speed v yjd of the aircraft, the altitude h(t) of the aircraft during the flare process, and the lifting speed v y (t) of the aircraft during the flare process, the expected value of the lifting speed v yg (t) and the expected value of the lifting acceleration a yg (t) are obtained as follows:

[0142]

[0143] According to the open-loop correction amount ω zeSemi of the expected pitch rate and the integral coefficient K LI of the lifting speed control, the open-loop integral correction amount v ygSemi of the lifting speed is obtained;

[0144] The open-loop integral correction amount v ygSemi of the lifting speed is as follows:

[0145] v ygSemi = ω zeSemi / K LI .

[0146] When adopting the PID control structure, the basic idea of automatic control is based on the feedback of the deviation. Introducing the integral signal of the deviation amount can eliminate the steady-state error of the system. However, the elimination of its steady-state error requires a certain adjustment time. Since the expected value of the aircraft's lifting speed changes with altitude during the flare process, and the actual movement of the controlled aircraft always lags behind the expected movement, the lifting speed at the actual aircraft touchdown is not the expected value. Therefore, considering the safety requirements during aircraft landing, an open-loop correction amount is introduced and enters the integral link together with the closed-loop feedback amount.

[0147] To implement the lifting speed control link, according to the lifting speed v y (t) of the aircraft during the flare process, the lifting acceleration a y (t) of the aircraft during the flare process, the expected value of the lifting speed v yg (t), the expected value of the lifting acceleration a yg (t), the open-loop integral correction amount v ygSemi of the lifting speed, and the control parameters K L , K LI , K LD , the expected pitch angle θ g (t) is obtained:

[0148] Among them, the control parameter K L 、K LI 、K LD is selected according to the current pitch angle control loop.

[0149] As Figure 2 shown, the airborne control module 300 is used to implement the pitch angle control link and obtain the elevator deflection command when the current pitch angle does not match the desired pitch angle.

[0150] Specifically, according to the current pitch angle θ(t) of the aircraft, the desired pitch angle θ g (t), the pitch rate ω z (t) of the aircraft, and the corresponding control parameters K g 、K ωz , the elevator deflection command δ z (t) is obtained;

[0151] δ z (t) = K θ (θ(t) - θ g (t)) + K ωz ω z (t);

[0152] δ z (t) is used as the input of the elevator loop, and the aircraft is leveled by controlling the elevator deflection.

[0153] Therefore, in the aircraft landing control system provided by this application, during the descent of the aircraft, the pilot who holds the control stick turns on the automatic landing leveling system. The system starts to photograph the runway and captures the anchor point information on the runway. And through the method of machine learning, the current aircraft attitude, lift speed, and lift acceleration are obtained, so as to obtain the current pitch angle. At the same time, the system calculates the desired pitch angle and compares the current pitch angle with the desired pitch angle. If they are consistent, the flight continues; if they are inconsistent, the system takes control of the elevator and adjusts the angle to the desired pitch angle, so as to ensure the accuracy of the leveling action and thus ensure the safety of the aircraft during descent.

[0154] This application also provides an aircraft, including the above-mentioned aircraft landing control system. During the leveling stage of the aircraft landing, the subjective judgment of the pilot is replaced by the objective computer automatic judgment, reducing the risk of problems such as hard landing and excessive floating, so as to ensure the stability and safety of the aircraft.

[0155] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0156] The above has introduced in detail a method, system, and aircraft for aircraft landing control provided by the embodiments of the present application. Specific examples are used in this article to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the technical solution and its core idea of the present application; those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. An aircraft landing control method, characterized in that, It includes the following steps: Obtain the current scene image when the aircraft lands; Compare and analyze the current scene image and the reference scene image to obtain an effective scene image; Apply the effective scene image to the training of a machine learning model to obtain the current pitch angle and the desired pitch angle of the aircraft; Judge whether the current pitch angle and the desired pitch angle match. When the current pitch angle and the desired pitch angle do not match, implement the pitch angle control link to obtain the elevator deflection command.

2. The aircraft landing control method according to claim 1, wherein The step of comparing and analyzing the current scene image and the reference scene image to obtain an effective scene image includes: Obtain the reference feature information of the reference scene image and the current feature information of the current scene image; Compare and analyze the current feature information with the reference feature information to obtain effective feature information, and determine the effective scene image through the effective feature information.

3. The aircraft landing control method according to claim 2, wherein The reference feature information includes the reference aircraft attitude, the reference aircraft speed, and the reference environmental information; The current feature information includes the current aircraft attitude, the current aircraft speed, and the current environmental information; The effective feature information respectively includes the effective aircraft attitude, the effective aircraft speed, and the effective environmental information.

4. The aircraft landing control method according to claim 3, wherein The step of comparing and analyzing the current feature information with the reference feature information to obtain effective feature information, and determining the effective scene image through the effective feature information includes: Compare and analyze the reference environmental information with the current environmental information to determine whether the current scene image has the effective environmental information; When the current scene image has the effective environmental information, then compare and analyze the reference aircraft attitude with the current aircraft attitude, and the reference aircraft speed with the current aircraft speed respectively to determine whether the current scene image has the effective aircraft attitude and the effective aircraft speed; When the current scene image has both the effective environmental information, the effective aircraft attitude, and the effective aircraft speed, then determine that the current scene image is the effective scene image.

5. The aircraft landing control method according to claim 4, wherein The step of applying the effective scene image to the training of a machine learning model to obtain the current pitch angle and the desired pitch angle of the aircraft includes: Obtain the current pitch angle; Calculate according to the effective environmental information, the effective aircraft attitude, and the effective aircraft speed, and combine the leveling law that the predetermined leveling trajectory should follow to obtain the desired pitch angle.

6. The aircraft landing control method according to claim 1, wherein Before the step of judging whether the current pitch angle and the desired pitch angle match, and implementing the pitch angle control link to obtain the elevator deflection command when the current pitch angle and the desired pitch angle do not match, Implement the generation link of the expected values of the lifting speed and the lifting acceleration to obtain the expected value of the lifting speed v yg (t) and the expected value of the lifting acceleration a yg (t); According to the altitude h0 and the vertical speed v of the aircraft during the flare control y0 and the desired touchdown speed v yjd of the aircraft, the altitude h(t) of the aircraft during the flare process, and the vertical speed v y (t) of the aircraft during the flare process, the expected value v yg (t) of the vertical speed and the expected value a yg (t) of the vertical acceleration are obtained: According to the desired pitch angle rate open-loop correction amount ω zeSemi and the elevator speed control integral coefficient K LI , the elevator speed open-loop integral correction amount v ygSemi ; Open-loop integral correction amount v of the lifting speed ygSemi As shown in the following formula: Iv ygSemi = ω zeSemi / K LI ; Implement the lift-off speed control loop, based on the aircraft's lift-off speed v y (t), the aircraft's lift-off acceleration a y (t), the expected lift-off speed v yg (t), the expected lift-off acceleration a yg (t), the open-loop integral correction of the lift-off speed v ygSemi i and the control parameters K L 、K LI 、K LD , to obtain the expected pitch angle θg(t): Among them, the control Parameter K L , K LI , K LD Selected according to the current pitch angle control loop.

7. The aircraft landing control method according to claim 6, wherein the step of determining whether the current pitch angle and the desired pitch angle match, and when the current pitch angle and the desired pitch angle do not match, implementing a pitch angle control loop to obtain an elevator deflection command includes: Implement the current pitch angle control link to obtain the elevator deflection command δ z (t); According to the current pitch angle θ(t) of the aircraft, the desired pitch angle θ g (t), the pitch angle rate ω z (t) of the aircraft, and the corresponding control parameters K g , the elevator deflection command δ z (t) is obtained; δ z (t) serves as the input of the elevator loop, and the aircraft's flaring control is achieved by controlling the deflection of the elevator.

8. An automatic landing flare control system for an aircraft, characterized in that, including: an airborne detection module, configured to obtain a current scene image during aircraft landing and compare and analyze the current scene image with a reference scene image to obtain a valid scene image; an airborne core network module, configured to determine whether the current pitch angle and the desired pitch angle match; an airborne control module, configured to implement a pitch angle control loop to obtain an elevator deflection command when the current pitch angle and the desired pitch angle do not match.

9. The aircraft landing control system according to claim 8, wherein the airborne detection module includes: an image acquisition unit, configured to obtain the current scene image during aircraft landing; a landing posture and speed model unit, connected to the image acquisition unit, configured to store the reference scene image; a data preprocessing unit, connected to the image acquisition unit and the landing posture and speed model unit, the data preprocessing unit being configured to compare and analyze the current scene image with the reference scene image to obtain the valid scene image.

10. The aircraft landing control system according to claim 8, wherein the airborne detection module further includes: a network interface unit, connected to the airborne core network module, configured to transmit the valid scene image to the airborne core network module.

11. The aircraft landing control system according to claim 8, wherein the airborne core network module includes: a machine learning unit, configured to train the valid scene image to obtain the current pitch angle and the desired pitch angle; a comprehensive processing unit, configured to determine whether the current pitch angle and the desired pitch angle match.

12. An aircraft, characterized in that, including the aircraft landing control system according to any one of claims 8 to 11.

Images