Method for predicting and estimating ADS-B track in real time through satellite platform

Abstract

The invention relates to a method for predicting and estimating an ADS-B track in real time by a satellite platform, and belongs to the field of satellite control systems. Aligning the time of the satellite-borne ADS-B receiver with the time of the satellite platform; after the satellite platform accumulatively collects five packets of effective data, based on a least square algorithm, adopting a quadratic polynomial to slide and fit an aircraft track curve; calculating a constant term coefficient ah, a monomial term coefficient bh and a quadratic term coefficient ch of a quadratic polynomial; in each control period of the satellite platform, predicting the position unow of the aircraft at the current moment; based on an extended Kalman filtering algorithm of a CA model, filtering estimation is carried out on the predicted position unow of the aircraft at the current moment, and flight path information, subjected to prediction and filtering estimation, of the aircraft at the current moment is obtained; the satellite platform performs continuous tracking and pointing on the aircraft based on the flight path information of the aircraft at the current moment; accurate data are provided for follow-up satellite high-precision and high-stability aircraft tracking, the algorithm is simple, the principle is clear, and satellite in-orbit implementation is easy.

Description

Technical Field

[0001] This invention belongs to the field of satellite control systems and relates to a method for real-time prediction and estimation of ADS-B tracks for satellite platforms. Background Technology

[0002] The ADS-B (Automatic Dependent Surveillance-Broadcast) system works by having aircraft update their position information (longitude, latitude, altitude, time), aircraft identity and category information, and operational status information through various onboard sensors. This information is then packaged into ADS-B data frames according to specific requirements and broadcast at a certain rate via onboard ADS-B transmitters. Spaceborne ADS-B systems primarily utilize ADS-B receivers mounted on low-Earth orbit communication satellites, eliminating reliance on ground-based equipment. Therefore, they can cover areas not covered by land-based ADS-B systems, forming a comprehensive global ADS-B network for tracking and real-time monitoring of aircraft worldwide. Furthermore, due to the high accuracy, high-frequency multi-mode operation, and wide coverage of spaceborne ADS-B systems, operating costs are reduced, and system accuracy and reliability are improved, leading to their increasingly widespread application.

[0003] The raw point tracking data provided by the onboard ADS-B receiver is characterized by its discontinuous nature (uniform time intervals) and low signal-to-noise ratio. According to traditional onboard data protocols, the satellite platform transmits the raw point tracking data back to the ground via a data transmission channel, with only a small amount of payload telemetry data being transmitted in real-time via telemetry. Upon receiving the point tracking information, the ground typically employs a Kalman filter algorithm based on a "current" statistical model to predict and estimate the aircraft's trajectory. To achieve real-time tracking of civil aircraft by low-Earth orbit satellites, the satellite platform needs to calculate the current aircraft position in real-time on-orbit based on the point tracking data from the onboard ADS-B receiver. The "current" statistical model used for ground data processing is particularly suitable for highly maneuverable targets and can meet the requirements for high-precision tracking of target aircraft. However, due to limitations in the capacity and computing power of current onboard computers, the data processing method based on the "current" statistical model cannot be practically applied on the satellite platform due to its complex formulas and high computational load. Summary of the Invention

[0004] The technical problem solved by this invention is to overcome the shortcomings of the prior art and propose a method for real-time prediction and estimation of ADS-B tracks by a satellite platform, which provides accurate data for subsequent high-precision and high-stability tracking of aircraft by satellites. The algorithm is simple, the principle is clear, and it is easy to implement on satellite in orbit.

[0005] The solution of the present invention is:

[0006] A method for real-time prediction and estimation of ADS-B tracks using a satellite platform includes:

[0007] Step 1: Align the time of the onboard ADS-B receiver with the time of the satellite platform;

[0008] Step 2: The onboard ADS-B receiver sends one data packet to the satellite platform at fixed intervals; after the satellite platform has collected n valid data packets, it uses a univariate quadratic polynomial sliding fit to fit the aircraft trajectory curve based on the least squares algorithm; the constant term coefficient 'a' of the univariate quadratic polynomial is calculated. h coefficient b of the linear term h The coefficient of the quadratic term c h n is a positive integer;

[0009] Step 3: Within each control cycle of the satellite platform, based on the constant term coefficient a h coefficient b of the linear term h The coefficient of the quadratic term c h Predict the current position of the aircraft. now ;

[0010] Step 4: Using the extended Kalman filter algorithm based on the CA model to predict the aircraft's current position u. now Filtering and estimation are performed to obtain the aircraft's trajectory information at the current moment after prediction and filtering estimation; the satellite platform continuously tracks and points the aircraft based on the aircraft's trajectory information at the current moment.

[0011] In the above-described method for real-time prediction and estimation of ADS-B tracks using a satellite platform, the time alignment method in step one is as follows:

[0012] The satellite platform and the onboard ADS-B receiver use the same time reference and are calibrated regularly.

[0013] In the above-mentioned method for real-time prediction and estimation of ADS-B tracks by a satellite platform, in step two, the onboard ADS-B receiver sends one data packet to the satellite platform every 1 second; n is 5; each data packet includes the aircraft ID number, data packaging time, data update status, longitude, latitude, and altitude information.

[0014] In the aforementioned method for real-time prediction and estimation of ADS-B tracks using a satellite platform, after receiving the data, the satellite platform determines whether it is the same aircraft as the one previously received based on the aircraft's ID number. The satellite platform needs to continuously track the same aircraft, so if it is not the same aircraft, the data packet is considered invalid. The satellite platform also needs to determine the data update status; if the status word shows no update, the data packet is also considered invalid.

[0015] In the aforementioned method for real-time prediction and estimation of ADS-B tracks using a satellite platform, a quadratic polynomial curve is used to fit the aircraft track curve. The method for determining the coefficients of the quadratic polynomial curve is as follows:

[0016] Let u(i) be the 5 most recent valid data packets, i = 1, 2, 3, 4, 5; where u(1) is the latest data packet; let t(i) be the time corresponding to u(i), and let T be the control period of the satellite platform. c Then the data storage matrix A h for:

[0017]

[0018] Let the time interval between two valid location information data be Δt. i ;

[0019] To reduce the fitting error, the time t(5) corresponding to the valid data of the 5th packet is forcibly converted to a smaller fitting time T(5); let T(5) = T c +1, then the fitting time T(i) corresponding to the remaining 4 location information data points is T(i-1)=T(i)+Δt i i = 2, 3, 4, 5;

[0020] The corresponding data storage matrix A h Convert to A h ′:

[0021]

[0022] Independent variable matrix B h and the dependent variable matrix C h They are respectively:

[0023] B h (i,1)=T(i) 0

[0024] B h (i,2)=T(i) 1 C h (i,1)=u(i)

[0025] B h (i,3)=T(i) 2

[0026] Calculate the coefficient vector D h :

[0027]

[0028] Let the coefficient of the constant term of the quadratic polynomial be a. h The coefficient of the linear term is bh The coefficient of the quadratic term is c. h ,but:

[0029] a h =D h (1)

[0030] b h =D h (2)

[0031] c h =D h (3)

[0032] If no valid data packet is received in this control cycle, the fitting coefficients will not be updated in this cycle; if a valid data packet is received, the data storage matrix A will be updated. h Following the first-in, first-out principle, the data is updated continuously, and the fitting coefficients are recalculated.

[0033] In the aforementioned method for real-time prediction and estimation of ADS-B tracks using a satellite platform, the interval is Δt. i The calculation method is as follows:

[0034] Δt i = t(i-1) - t(i)

[0035] In the formula, i = 2, 3, 4, 5.

[0036] In the aforementioned method for real-time prediction and estimation of ADS-B tracks using a satellite platform, the current position of the aircraft, u, is predicted. now The method is as follows:

[0037] Let the current control time be t. now At that moment t now The time difference between t(5) and t is Δt. now :

[0038] Δt now =t now -t(5)

[0039] t now Corresponding fitting time T now for:

[0040] T now =T(5)+Δt now

[0041] The aircraft position u at that moment was calculated. now :

[0042]

[0043] In the aforementioned method for real-time prediction and estimation of ADS-B tracks using a satellite platform, the location u... now It includes longitude, latitude, and altitude; among which longitude and latitude have a range of values ​​and are subject to amplitude limiting according to their physical meaning.

[0044] In the aforementioned method for real-time prediction and estimation of ADS-B tracks using a satellite platform, the longitude range is (-180°, 180°), meaning that when the result calculated from the fitted curve exceeds 180°, 360° is subtracted; conversely, when it is less than 180°, 360° is added. The latitude range is [-90°, 90°], meaning that when the result calculated from the fitted curve exceeds 90°, 360° is subtracted; conversely, when it is less than 90°, 360° is added.

[0045] In the aforementioned method for real-time prediction and estimation of ADS-B tracks using a satellite platform, step four involves predicting the aircraft's current position u. now The method for performing filter estimation is as follows:

[0046] Establish the continuous-time state equations for the CA model:

[0047]

[0048] In the formula, X(t) is the position information matrix of the aircraft at time t; Where, x(t), These are the aircraft's position, velocity, and acceleration information at time t;

[0049] To find the derivative of X(t);

[0050] A is the transition matrix;

[0051] ω(t) has a mean of zero and a variance of σ. 2 Gaussian white noise;

[0052] The discrete form F corresponding to the transition matrix A is:

[0053]

[0054] Set the system noise variance matrix Q in the Kalman filter algorithm as follows:

[0055]

[0056] Where, T c For system control cycle;

[0057] q is the noise figure;

[0058] make The change in acceleration within the control period is equal to the amount of q that can be obtained.

[0059] Measurement information in the Kalman filter algorithm The predicted position u in step three now Based on the previous estimate and the current measurement. Perform residual calculation to obtain the system's information error; perform data filtering based on the information error to update the state and obtain estimated values ​​of the state variables; and save the estimated values ​​for the next filtering estimation.

[0060] The advantages of this invention compared to the prior art are:

[0061] (1) This invention eliminates the time delay effect of point information by unifying the time base and calibrating the time periodically, optimizing the state equation of the extended Kalman filter algorithm, and using fitting prediction and filtering estimation for the track.

[0062] (2) This invention simplifies the data processing formula, realizes the satellite platform's requirement for accurate estimation of target information, and provides a data source for subsequent high-precision and high-stability tracking;

[0063] (3) This invention provides accurate data for subsequent high-precision and high-stability satellite tracking of aircraft. The algorithm is simple, the principle is clear, and it is easy to implement on satellite in orbit. Attached Figure Description

[0064] Figure 1 This is a flowchart illustrating the real-time prediction and estimation of ADS-B tracks by the satellite platform of this invention. Detailed Implementation

[0065] The present invention will be further described below with reference to the embodiments.

[0066] This invention provides a method for real-time prediction and estimation of ADS-B tracks by a satellite platform. By unifying the time base and performing time-based calibration, optimizing the state equation of the extended Kalman filter algorithm, and employing fitting prediction and filtering estimation methods for the tracks, the method eliminates the time lag effect of point information, simplifies data processing formulas, and meets the requirement of accurate target information estimation by the satellite platform. This provides a data source for subsequent high-precision and high-stability tracking.

[0067] Commercial aircraft typically operate along fixed routes, exhibiting significant maneuvering acceleration only during ascent and descent, with relatively stable speeds during cruising. Furthermore, the cruising altitude of commercial aircraft is usually around 10 km, while low-Earth orbit satellites operate at altitudes of approximately 400 km to 900 km. Therefore, the distance between the satellite and the aircraft differs significantly from the distance from the ground, meaning that from a satellite perspective, the aircraft's motion appears as slow. Based on these two points, satellite platforms can treat aircraft motion as weak motion when tracking aircraft, allowing for simpler on-orbit real-time processing of ADS-B data.

[0068] Methods for real-time prediction and estimation of ADS-B tracks using satellite platforms, such as Figure 1 As shown, the specific steps include the following:

[0069] Step 1: Align the time of the onboard ADS-B receiver with the time of the satellite platform. The time alignment method is as follows:

[0070] The satellite platform and the onboard ADS-B receiver use the same time reference and are calibrated regularly.

[0071] Step 2: The onboard ADS-B receiver sends one data packet to the satellite platform every 1 second; after the satellite platform has collected n valid data packets, it uses a univariate quadratic polynomial sliding fit to fit the aircraft trajectory curve based on the least squares algorithm; the constant term coefficient 'a' of the univariate quadratic polynomial is calculated. h coefficient b of the linear term h The coefficient of the quadratic term c h In this invention, n is set to 5.

[0072] In step two, each data packet includes the aircraft's ID number, the time the data was packaged, the data update status, longitude, latitude, and altitude information. Upon receiving this data, the satellite platform determines whether it belongs to the same aircraft as the one previously received. Since the satellite platform needs to continuously track the same aircraft, if it does not belong to the same aircraft, the data packet is considered invalid. The satellite platform also needs to check the data update status; if the status indicates no update, the data packet is also considered invalid.

[0073] The aircraft trajectory curve is fitted using a univariate quadratic polynomial curve. The method for determining the coefficients of the univariate quadratic polynomial curve is as follows:

[0074] Let u(i) be the 5 most recent valid data packets, i = 1, 2, 3, 4, 5; where u(1) is the latest data packet; let t(i) be the time corresponding to u(i), and let T be the control period of the satellite platform. c Then the data storage matrix A h for:

[0075]

[0076] Let the time interval between two valid location information data be Δt. i .

[0077] The interval time is Δt i The calculation method is as follows:

[0078] Δt i = t(i-1) - t(i)

[0079] In the formula, i = 2, 3, 4, 5.

[0080] To reduce the fitting error, the time t(5) corresponding to the valid data of the 5th packet is forcibly converted to a smaller fitting time T(5); let T(5) = T c +1, then the fitting time T(i) corresponding to the remaining 4 location information data points is T(i-1)=T(i)+Δt i i = 2, 3, 4, 5;

[0081] The corresponding data storage matrix A h Convert to A h ′:

[0082]

[0083] Independent variable matrix B h and the dependent variable matrix C h They are respectively:

[0084] B h (i,1)=T(i) 0

[0085] B h (i,2)=T(i) 1 C h (i,1)=u(i)

[0086] B h (i,3)=T(i) 2

[0087] Calculate the coefficient vector D h :

[0088]

[0089] Let the coefficient of the constant term of the quadratic polynomial be a. h The coefficient of the linear term is b h The coefficient of the quadratic term is c. h ,but:

[0090] a h =D h (1)

[0091] b h =D h (2)

[0092] c h =D h (3)

[0093] If no valid data packet is received in this control cycle, the fitting coefficients will not be updated in this cycle; if a valid data packet is received, the data storage matrix A will be updated. h Following the first-in, first-out principle, the data is updated continuously, and the fitting coefficients are recalculated.

[0094] Step 3: Within each control cycle of the satellite platform, based on the constant term coefficient a h coefficient b of the linear term h The coefficient of the quadratic term c h Predict the current position of the aircraft. now .

[0095] In step three, the current position u of the aircraft is predicted. now The method is as follows:

[0096] Let the current control time be t. now At that moment t now The time difference between t(5) and t is Δt. now :

[0097] Δt now =t now -t(5)

[0098] t now Corresponding fitting time T now for:

[0099] T now =T(5)+Δt now

[0100] The aircraft position u at that moment was calculated. now :

[0101]

[0102] Position u nowThis includes longitude, latitude, and altitude; both longitude and latitude have value ranges and are limited according to their physical meaning. The value range for longitude is (-180°, 180°], meaning that if the result calculated from the fitted curve exceeds 180°, then 360° is subtracted; conversely, if it is less than 180°, then 360° is added. The value range for latitude is [-90°, 90°], meaning that if the result calculated from the fitted curve exceeds 90°, then 360° is subtracted; conversely, if it is less than 90°, then 360° is added.

[0103] Step 4: Using the extended Kalman filter algorithm based on the CA model to predict the aircraft's current position u. now Filtering and estimation are performed to obtain the aircraft's trajectory information at the current moment after prediction and filtering estimation; the satellite platform continuously tracks and points the aircraft based on the aircraft's trajectory information at the current moment.

[0104] In step four, the predicted current aircraft position u is calculated. now The method for performing filter estimation is as follows:

[0105] Establish the continuous-time state equations for the CA model:

[0106]

[0107] In the formula, X(t) is the position information matrix of the aircraft at time t; Where, x(t), These are the aircraft's position, velocity, and acceleration information at time t;

[0108] To find the derivative of X(t);

[0109] A is the transition matrix;

[0110] ω(t) has a mean of zero and a variance of σ. 2 Gaussian white noise;

[0111] The discrete form F corresponding to the transition matrix A is:

[0112]

[0113] Set the system noise variance matrix Q in the Kalman filter algorithm as follows:

[0114]

[0115] Where, T c For system control cycle;

[0116] q is the noise figure;

[0117] make It equals the change in acceleration within the control period, and then q is obtained.

[0118] Measurement information in the Kalman filter algorithm The predicted position u in step three now Based on the previous estimate and the current measurement. Perform residual calculation to obtain the system's information error; perform data filtering based on the information error to update the state and obtain estimated values ​​of the state variables; and save the estimated values ​​for the next filtering estimation.

[0119] Based on the previous estimate and the current measurement The residual is calculated to obtain the system's innovation error. This innovation error is then used for data filtering to update the state and obtain estimates of the state variables. These estimates are then saved for the next filtering cycle.

[0120] Thus, by fitting a sliding curve, the current flight path information (longitude, latitude, and altitude) of the aircraft can be obtained after prediction and filtering estimation. The satellite platform can then continuously track and point the aircraft based on this flight path information.

[0121] This invention eliminates the time lag effect of point information by unifying the time base and calibrating the time period, optimizing the state equation of the extended Kalman filter algorithm, and using fitting prediction and filtering estimation for the track; it simplifies the data processing formula, realizes the satellite platform's requirement for accurate estimation of target information, and provides a data source for subsequent high-precision and high-stability tracking;

[0122] This invention provides accurate data for subsequent high-precision and high-stability satellite tracking of aircraft. The algorithm is simple, the principle is clear, and it is easy to implement on satellite in orbit.

[0123] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications to the technical solutions of the present invention by utilizing the methods and techniques disclosed above without departing from the spirit and scope of the present invention. Therefore, any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solutions of the present invention shall fall within the protection scope of the technical solutions of the present invention.

Claims

1. A method for real-time prediction and estimation of ADS-B tracks by a satellite platform, characterized in that: include: Step 1: Align the time of the onboard ADS-B receiver with the time of the satellite platform; Step 2: The satellite-borne ADS-B receiver sends a packet of data to the satellite platform at a fixed interval; after the satellite platform has accumulated n packets of valid data, it uses a quadratic polynomial sliding fit to fit the aircraft track curve based on the least squares algorithm; Calculate the constant term coefficient a of a quadratic polynomial h , linear coefficient b h , quadratic coefficient c h ; n is a positive integer; Step 3: In each control cycle of the satellite platform, according to the constant coefficient a h , linear coefficient b h , quadratic coefficient c h , predict the current position of the aircraft u now ; Step 4: Based on the extended Kalman filter algorithm of the CA model, predict the current aircraft position u now Perform filtering estimation to obtain the track information of the aircraft at the current moment after prediction and filtering estimation; The satellite platform continuously tracks and points to the aircraft based on the aircraft's current track information.

2. The method for real-time prediction and estimation of ADS-B tracks by a satellite platform according to claim 1, characterized in that: In step 1, the time alignment method is: The satellite platform and the onboard ADS-B receiver use the same time reference and are regularly calibrated.

3. The method for real-time prediction and estimation of ADS-B tracks by a satellite platform according to claim 1, characterized in that: In the step 2, the satellite-borne ADS-B receiver sends a data packet to the satellite platform every 1 second; n is 5; each data packet includes the aircraft ID number, data packaging time, data update status, longitude, latitude, and altitude information.

4. The method for real-time prediction and estimation of ADS-B tracks by a satellite platform according to claim 3, characterized in that: After receiving the data, the satellite platform determines whether it is the same aircraft as the last received one based on the aircraft ID number; the satellite platform needs to continuously track the same aircraft, so when it is not the same aircraft, the package data is considered invalid; the satellite platform also needs to determine the data update status, and when the status word shows that it has not been updated, the reported data is also considered invalid.

5. The method for real-time prediction and estimation of ADS-B tracks by a satellite platform according to claim 4, characterized in that: A quadratic polynomial curve is used to fit the aircraft track curve. The coefficients of the quadratic polynomial curve are determined as follows: Suppose the last five packets of valid data are u(i), i = 1, 2, 3, 4, 5; where u(1) is the latest data; let the time corresponding to u(i) be t(i), and let the control period of the satellite platform be T c , then the data storage matrix A h for: Let the interval between two valid position information data be Δt i ; In order to reduce the fitting error, the time t(5) corresponding to the fifth packet of valid data is forcibly converted to a smaller fitting time T(5); let T(5) = T c +1, then the fitting time T(i) corresponding to the remaining 4 position information data points is T(i-1)=T(i)+Δt i ,i=2,3,4,5; The corresponding data is stored in matrix A h Convert to A h ′: Independent variable matrix B h And the dependent variable matrix C h They are: B h (i,1)=T(i) 0 B h (i,2)=T(i) 1 ,C h (i,1)=u(i) B h (i,3)=T(i) 2 Calculate the coefficient vector D h : Set the constant coefficient of the quadratic polynomial to a h The coefficient of the first-order term is b h , the coefficient of the quadratic term is c h ,but: a h =D h (1) b h =D h (2) c h =D h (3) If no valid data packet is received in this control cycle, the fitting coefficients will not be updated in this beat; if a valid data packet is received, the data storage matrix A h According to the first-in-first-out principle, the data is updated in a rolling manner and the fitting coefficients are recalculated.

6. The method for real-time prediction and estimation of ADS-B tracks by a satellite platform according to claim 5, characterized in that: The interval time is Δt i The calculation method is: Δt i =t(i-1)-t(i) In the formula, i=2,3,4,5.

7. The method for real-time prediction and estimation of ADS-B tracks by a satellite platform according to claim 6, characterized in that: In step 3, the position u of the aircraft at the current moment is predicted now The method is: Assume the current control time is t now , the time t now The time difference between t(5) and t(5) is Δt now : Δt now =t now -t(5) t now The corresponding fitting time T now for: T now =T(5)+Δt now Calculate the aircraft position u at that moment now :

8. The method for real-time prediction and estimation of ADS-B tracks by a satellite platform according to claim 7, characterized in that: The position u now It includes longitude, latitude and altitude; longitude and latitude have value ranges and are limited according to the physical meaning.

9. The method for real-time prediction and estimation of ADS-B tracks by a satellite platform according to claim 8, characterized in that: The range of the longitude is (-180°, 180°], that is, when the result calculated according to the fitting curve exceeds 180°, 360° is subtracted; conversely, when it is less than 180°, 360° is added; the range of the latitude is [-90°, 90°], that is, when the result calculated according to the fitting curve exceeds 90°, 360° is subtracted; conversely, when it is less than 90°, 360° is added.

10. The method for real-time prediction and estimation of ADS-B tracks by a satellite platform according to claim 1, characterized in that: In step 4, the current aircraft position u is predicted now The method for filtering estimation is: Establish the continuous-time state equation of the CA model: Where X(t) is the position information matrix of the aircraft at time t; Among them, x(t), They are the position information, velocity information, and acceleration information of the aircraft at time t respectively; is the derivative of X(t); A is the transfer matrix; ω(t) has a mean of zero and a variance of σ 2 Gaussian white noise; The discrete form F corresponding to the transfer matrix A is: Set the system noise variance matrix Q in the Kalman filter algorithm to: Where, T c is the system control cycle; q is the noise factor; make Equal to the acceleration change during the control period and then calculate q; Measurement Information in Kalman Filter Algorithm is the position u predicted in step 3 now ; Based on the last estimated value and this measured value Perform residual calculation to obtain the information error of the system; perform data filtering operations based on the information error to achieve state update and obtain the estimated value of the state quantity; at the same time, save the estimated value for the next filtering estimation.

Images