A method and system for determining coefficients of a smooth switching state transition matrix

Abstract

The application relates to the technical field of aircraft simulation control, and provides a determination method and system for smoothly switching state transition matrix coefficients, and solves the problems of attitude estimation misalignment, control instruction oscillation and hardware impact in the prior art. The method comprises the following steps: acquiring flight attitude change data, and generating a characteristic vector through space-time coupling processing; collecting three-dimensional angular motion data by using an inertial navigation unit array, implementing non-uniform weight distribution on high-frequency components, and outputting a stable state estimation quantity; constructing a double-threshold characteristic boundary based on the deviation degree of the characteristic vector relative to the stable state estimation quantity, determining a maneuver mutation opportunity when a continuous state transition trajectory crosses the boundary, and generating a switching trigger signal; and performing a smooth switching operation of state transition matrix coefficients in response to the signal. Through the fusion of multi-source sensor data, the construction of a dynamic characteristic boundary and the accurate capture of a maneuver mutation opportunity, the smooth and adaptive switching of state transition matrix coefficients is realized.

Description

Technical Field

[0001] The present application relates to the field of aircraft simulation control technology, and in particular to a method and system for determining coefficients of a smooth switching state transfer matrix. Background Art

[0002] When reproducing highly maneuverable flight maneuvers, flight simulators must dynamically adjust the state transition matrix coefficients of the motion simulation algorithm based on attitude changes to accurately convey the instantaneous acceleration sensation. These scenarios require the algorithm to respond quickly, switching coefficients within a very short timeframe. Furthermore, smooth and continuous acceleration commands must be ensured during the switching process to avoid sensory distortion. Furthermore, the risk of hydraulic platform overload caused by the switching shock must be completely eliminated to ensure safe hardware operation.

[0003] The current mainstream solution uses a pre-set maneuver rule library and a linear transition strategy. By establishing a mapping library between typical flight patterns and filter coefficients, sensors identify the maneuver type and call the corresponding coefficient set. When a maneuver switch is detected, the system performs a linear interpolation transition between the old and new coefficients within a fixed time window, attempting to mitigate output jumps through gradual changes.

[0004] The current mainstream solution is limited by the fixed scenario coverage capability of the preset rule library, and it is difficult to adapt to the dynamic coupling effects of complex maneuvers, resulting in inaccurate coefficient matching; the linear interpolation process ignores the system dynamic constraints and cannot maintain the continuity of the first-order derivative of acceleration, causing output command jumps and platform mechanical vibrations; at the same time, the delay in rule matching and interpolation calculation significantly exceeds the real-time requirements of high-maneuverability scenarios, causing simulated action lags, and the impact load at the switching moment continues to threaten the safety of the hardware structure. Summary of the Invention

[0005] The present application provides a method and system for determining the coefficients of a smooth switching state transfer matrix, which is used to solve the problems of attitude estimation inaccuracy, control command oscillation and hardware impact caused by delayed or uneven switching of state transfer matrix coefficients when the aircraft is violently maneuvering in the prior art.

[0006] In a first aspect, the present application provides a method for determining coefficients of a smooth switching state transfer matrix, comprising:

[0007] Acquiring flight attitude change data through an inertial measurement unit, performing spatiotemporal coupling processing on gyroscope angular velocity data in the flight attitude change data and satellite positioning data to generate a feature vector reflecting the attitude change trend;

[0008] Using an inertial navigation unit array to collect three-dimensional angular motion data, performing non-uniform weight distribution on high-frequency components in the three-dimensional angular motion data, and outputting a steady-state estimate via a dynamic weighted average processor;

[0009] constructing a dual-threshold characteristic boundary based on the degree of deviation of the characteristic vector relative to the stable state estimate, and determining a maneuver mutation timing and generating a switching trigger signal when a continuous state transition trajectory crosses the dual-threshold characteristic boundary;

[0010] In response to the switching trigger signal, a smooth switching operation of the state transition matrix coefficients is performed.

[0011] Optionally, the process of performing a smooth switching operation of the state transfer matrix coefficients includes:

[0012] Obtain the old coefficient set of the current state transfer matrix and the new coefficient set of the target state transfer matrix;

[0013] Taking the switching time point as a reference point, based on the old coefficient set and the current state vector, a boundary condition vector that satisfies the continuity of the output acceleration command and its first-order derivative is calculated;

[0014] Constructing a state transition relationship equation using the new coefficient set, taking the boundary condition vector as the right-hand side term of the equation, and solving the state transition relationship equation to obtain a state vector after switching;

[0015] The switched state vector and the new coefficient set are synchronously loaded into the motion simulation filter to complete the smooth switching of the state transfer matrix coefficients.

[0016] Optionally, the calculating of the boundary condition vector satisfying the continuity of the output acceleration command and its first-order derivative based on the old coefficient set and the current state vector with the switching time point as the reference point includes:

[0017] Determine the switching time point of the state transfer matrix;

[0018] Acquire a first state vector corresponding to the switching time point, where the first state vector includes a first position component and a first velocity component;

[0019] Taking the first velocity component as the output acceleration component;

[0020] Based on the low-frequency coefficients and high-frequency coefficients in the old coefficient set and in combination with the first state vector, an output acceleration change component is calculated: the first position component is multiplied by the low-frequency coefficient to obtain a first intermediate value; the first velocity component is multiplied by the high-frequency coefficient to obtain a second intermediate value; the first intermediate value is added to the second intermediate value and the negative value is taken to obtain the output acceleration change component;

[0021] The output acceleration component and the output acceleration change component are combined in a predetermined order to form a boundary condition vector.

[0022] Optionally, constructing a state transition relationship equation using the new coefficient set, taking the boundary condition vector as the right-hand side of the equation, and solving the state transition relationship equation to obtain a post-switching state vector includes:

[0023] Constructing a coefficient matrix based on the low-frequency coefficients and the high-frequency coefficients in the new coefficient set;

[0024] Set the first row and first column elements of the coefficient matrix to a fixed value of 0, set the first row and second column elements of the coefficient matrix to a fixed value of 1, set the second row and first column elements of the coefficient matrix to the negative value of the low-frequency coefficient, and set the second row and second column elements of the coefficient matrix to the negative value of the high-frequency coefficient;

[0025] Using the boundary condition vector as a constant on the right side of the linear relationship equation;

[0026] A linear relationship equation with the coefficient matrix as the coefficient matrix and the boundary condition vector as the right-hand constant is solved to obtain a state vector after switching, wherein the state vector after switching includes a second position component and a second velocity component.

[0027] Optionally, the step of synchronously loading the switched state vector and the new coefficient set into a motion simulation filter to complete smooth switching of state transfer matrix coefficients includes:

[0028] Writing the second position component of the switched state vector into the position component storage space of the dynamic simulation filter;

[0029] Writing the second velocity component of the switched state vector into the velocity component storage space of the dynamic simulation filter;

[0030] Writing the low-frequency coefficients in the new coefficient set into the low-frequency coefficient storage space of the dynamic simulation filter;

[0031] Writing the high frequency coefficients in the new coefficient set into the high frequency coefficient storage space of the dynamic simulation filter;

[0032] When the writing operation of all storage spaces is completed, the dynamic simulation filter starts running refresh based on the value of the current storage space, and the switching of the state transfer matrix coefficients is completed.

[0033] Optionally, the method of collecting three-dimensional angular motion data using an inertial navigation unit array, performing non-uniform weight distribution on high-frequency components in the three-dimensional angular motion data, and outputting a steady-state estimate via a dynamic weighted average processor includes:

[0034] Three-dimensional angular motion data is collected synchronously through multiple inertial measurement nodes distributed in space. Each inertial measurement node outputs angular velocity components and linear acceleration components along three orthogonal axes.

[0035] Perform motion feature analysis on the angular velocity components output by each inertial measurement node, identify motion data segments containing rapidly changing features, and mark the identified motion data segments with dynamic feature identifiers;

[0036] According to the dynamic feature identifier, assigning a first weight coefficient to the marked motion data segment and assigning a second weight coefficient to the unmarked motion data segment, wherein the first weight coefficient is greater than the second weight coefficient;

[0037] The weighted angular velocity components are input into a dynamic weighted average processor, which performs the following processing: weighted fusion of the same axial angular velocity components from different inertial measurement nodes; motion consistency verification of each fused axial angular velocity component; and output of the verified three-dimensional angular motion fusion result as a stable state estimate.

[0038] Optionally, constructing a dual-threshold characteristic boundary based on the degree of deviation of the characteristic vector from the stable state estimate, and determining a maneuver mutation timing and generating a switching trigger signal when a continuous state transition trajectory crosses the dual-threshold characteristic boundary, includes:

[0039] Convert the pitch rate and roll acceleration parameters contained in the eigenvector into the first eigencoordinate, convert the platform displacement and velocity parameters contained in the stable state estimate into the second eigencoordinate, and establish a dynamic projection relationship between the first and second eigencoordinates;

[0040] An adjustable boundary threshold is set in the dynamic projection relationship, and the boundary threshold range is dynamically adjusted according to historical motion data to form a dual-threshold feature boundary including an upper boundary and a lower boundary;

[0041] Calculating the projection distance between the feature coordinates at the current moment and the stable state coordinates, and recording the projection distance change trajectory for multiple consecutive sampling periods, so as to mark the crossing point and crossing direction of the trajectory crossing the dual-threshold feature boundary through the projection distance change trajectory;

[0042] When the trajectory is detected to cross the same feature boundary three times in a row, the timestamp of the last crossing is recorded as the mutation moment, and a switching trigger signal containing the mutation moment, crossing point, and crossing direction is generated.

[0043] In a second aspect, the present application provides a system for determining coefficients of a smooth switching state transfer matrix, comprising:

[0044] an acquisition module, configured to acquire flight attitude change data through an inertial measurement unit, perform spatiotemporal coupling processing on gyroscope angular velocity data in the flight attitude change data and satellite positioning data, and generate a feature vector reflecting the attitude change trend;

[0045] an output module configured to collect three-dimensional angular motion data using an inertial navigation unit array, perform non-uniform weight distribution on high-frequency components in the three-dimensional angular motion data, and output a steady-state estimate via a dynamic weighted average processor;

[0046] a generating module, configured to construct a dual-threshold characteristic boundary based on the degree of deviation of the characteristic vector relative to the stable state estimate, and to determine a maneuver mutation timing and generate a switching trigger signal when a continuous state transition trajectory crosses the dual-threshold characteristic boundary;

[0047] An execution module is used to execute a smooth switching operation of the state transfer matrix coefficients in response to the switching trigger signal.

[0048] In a third aspect, an embodiment of the present application provides a computing device comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for determining the coefficients of a smooth switching state transfer matrix as described in the first aspect above.

[0049] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, it implements a method for determining the coefficients of a smooth switching state transfer matrix as described in the first aspect.

[0050] In an embodiment of the present application, flight attitude change data is acquired through an inertial measurement unit, and the gyroscope angular velocity data in the flight attitude change data is subjected to spatiotemporal coupling processing with the satellite positioning data to generate a feature vector reflecting the attitude change trend; an inertial navigation unit array is used to collect three-dimensional angular motion data, and non-uniform weight distribution is performed on the high-frequency components in the three-dimensional angular motion data, and a stable state estimate is output through a dynamic weighted average processor; based on the degree of deviation of the feature vector relative to the stable state estimate, a dual-threshold characteristic boundary is constructed, and when the continuous state transition trajectory crosses the dual-threshold characteristic boundary, the maneuver mutation timing is determined and a switching trigger signal is generated; in response to the switching trigger signal, a smooth switching operation of the state transfer matrix coefficients is performed.

[0051] The technical solution of this application has the following beneficial effects:

[0052] This application generates attitude change feature vectors by fusing inertial measurement units and satellite positioning data, and outputs stable state estimates by combining dynamic weighted processing of high-frequency motion data by the inertial navigation unit array. It uses dual-threshold feature boundaries to accurately capture the deviation state between the feature vector and the estimate, and instantly triggers maneuver mutation judgment when the continuous trajectory crosses the boundary, ultimately achieving smooth adaptive switching of the state transfer matrix coefficients, significantly improving the attitude estimation stability and control continuity of the aircraft in intense maneuvering scenarios.

[0053] The old coefficient set and target new coefficient set of the current state transfer matrix are further obtained; based on the switching time point, the boundary condition vector that satisfies the continuity of the output acceleration command and its first-order derivative is calculated based on the old coefficients and the current state vector; the state transfer relationship equation is constructed using the new coefficients, and the boundary condition vector is solved as the right-hand term to obtain the state vector after switching; finally, the new coefficients and the state vector after switching are synchronously loaded into the dynamic simulation filter to achieve coefficient switching. This solution ensures a smooth transition of the acceleration command at the switching moment through boundary condition vector constraints, and uses the state transfer equation solution to ensure that the system dynamic response is free of jumps. Combined with the synchronous loading of the dynamic simulation filter, it completely eliminates the mechanical vibration and trajectory mutation caused by traditional coefficient switching, achieving millisecond-level disturbance-free switching and providing continuous and stable dynamic support for aircraft maneuvering control.

[0054] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0056] Figure 1 A flow chart showing a method for determining a smooth switching state transfer matrix coefficient provided by the present application is shown;

[0057] Figure 2 A scenario diagram showing a method for determining the coefficients of a smooth switching state transfer matrix provided by the present application;

[0058] Figure 3 A schematic diagram of the structure of a system for determining the coefficients of a smooth switching state transfer matrix provided by the present application is shown;

[0059] Figure 4 A schematic structural diagram of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0060] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0061] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0062] Research shows that current flight simulator dynamic simulation algorithms generally adopt a coefficient switching strategy driven by a preset rule library when dealing with high-maneuverability scenarios. The core flaw is that the limited scenario coverage of the rule library makes it difficult to adapt to the strong nonlinear dynamic characteristics of complex maneuvers, resulting in inaccurate matching of the state transfer matrix coefficients; and although the linear interpolation transition scheme attempts to alleviate output jumps, it breaks the coupling relationship between the system dynamics equation and the coefficient switching, and cannot maintain the continuity of the acceleration command and its first-order derivative, causing mechanical vibration and hydraulic shock loads on the platform. At the same time, the delay in rule matching and interpolation calculation seriously restricts the response performance, forming an irreconcilable contradiction among response speed, output smoothness and hardware safety.

[0063] To address the above issues, the present invention proposes a method for determining the coefficients of the smooth state transition matrix. Its core technical approach is as follows: first, through spatiotemporal coupling, gyroscope and satellite positioning data are integrated to generate an attitude change feature vector. This is then combined with non-uniformly weighted noise reduction of the inertial navigation array to output a stable state estimate. A dual-threshold dynamic boundary is then constructed for the deviation between the feature vector and the estimate, accurately capturing the timing of sudden maneuver changes when continuous trajectories cross the boundary and generating a switching trigger signal. Finally, by solving the state transition equation under the boundary conditions, a strictly continuous transition of the acceleration command and its first-order derivative is achieved during the switching process between the old and new coefficients. This method completely eliminates the output step and mechanical vibration of traditional solutions, breaks through the scenario coverage limitations of the rule library, compresses the switching response time, and simultaneously resolves the three technical contradictions of response delay, somatosensory distortion, and hardware impact.

[0064] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0065] Figure 1 A flow chart of a method for determining a smooth switching state transfer matrix coefficient is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0066] 101. Acquire flight attitude change data through an inertial measurement unit, perform spatiotemporal coupling processing on gyroscope angular velocity data in the flight attitude change data and satellite positioning data, and generate a feature vector reflecting an attitude change trend;

[0067] In the above solution, the inertial measurement unit (IMU) refers to a sensor module integrated with a gyroscope and accelerometer. It measures the dynamic data of the aircraft's three-dimensional angular velocity and linear acceleration, including the instantaneous motion characteristics in the body coordinate system, and is used to construct the original input for attitude solution. Gyroscope angular velocity data refers to a quantized signal reflecting the aircraft's rotation rate around the X / Y / Z axes, including the angular change rate of the body in the roll, pitch, and yaw directions, and is used to detect sudden changes in attitude during maneuvers. Satellite positioning data refers to dynamic geospatial information obtained through the global satellite navigation system, including latitude and longitude coordinates, altitude, and three-dimensional velocity vectors. It is used to provide an absolute motion reference to correct the accumulated errors of inertial sensors. Spatiotemporal coupling processing refers to the technical means of resolving the differences in spatiotemporal references of multi-source sensor data. It includes interpolation synchronization in the time dimension and coordinate system conversion in the spatial dimension, used to generate a unified spatiotemporal fusion data stream.

[0068] In the embodiment of the present application, the inertial measurement unit first collects the original attitude data of the aircraft, wherein the gyroscope continuously outputs the instantaneous angular velocity values ​​of the pitch axis, roll axis, and yaw axis of the aircraft, while the satellite positioning module provides the spatial position and movement speed information of the aircraft. Since the gyroscope sampling frequency is much higher than the satellite data update rate, time alignment processing is required: a linear interpolation algorithm is used to interpolate the satellite data frames, and the interpolation speed of the intermediate moment is calculated based on the time ratio of the satellite speed values ​​of two adjacent frames, so that all data obtain a unified timestamp. The interpolation speed calculation formula is as follows:

[0069] ;

[0070] in, and For two adjacent frames, for The corresponding satellite velocity value, for The corresponding satellite velocity value, for For example, when the satellite outputs 300km / h at t1=0ms and 310km / h at t2=100ms, the interpolation speed of the gyroscope at t=50ms is calculated by the formula Synchronization completed.

[0071] Next, the time-aligned data enters the spatial coordinate conversion stage: the direction cosine matrix is ​​constructed using the current aircraft's pitch angle, roll angle, and yaw angle, and the satellite's northeast celestial coordinate system velocity vector is converted to the body coordinate system.

[0072] The specific conversion formula is as follows:

[0073] ;

[0074] ;

[0075] ;

[0076] in, 、 、 are the north velocity vector, east velocity vector and earth velocity vector in the northeast celestial coordinate system respectively, 、 、 are the front and rear velocity vector, left and right velocity vector, and up and down velocity vector in the body coordinate system, 、 、 are pitch angle, roll angle and yaw angle respectively. For example, when 、 Satellite data Convert to .

[0077] Finally, the converted body velocity is fused with the gyroscope angular velocity, and the angular acceleration is calculated by taking the differential derivative of the angular velocity of two consecutive frames. With the linear velocity vector The trajectory curvature is calculated by dividing the cross product modulus by the cube of the velocity. Finally, the angular acceleration, trajectory curvature and other indicators are combined into a feature vector to fully characterize the posture change trend. For example, when the angular acceleration is , when the trajectory curvature is 0.12, the output feature vector is [5000,0.12].

[0078] In actual applications, when a certain type of flight simulator was reproducing an emergency evasive maneuver, the inertial measurement unit (IMU) recorded a gyroscope roll rate of positive 120 degrees per second. Simultaneously, the satellite positioning module reported that the aircraft was moving northeast at 600 kilometers per hour and descending at a rate of 15 meters per second. Time alignment processing first linearly interpolates the satellite data, for example, calculating the velocity value at the intermediate moment between two adjacent frames of satellite data to synchronize the satellite data with the gyroscope data sampled 1,000 times per second. The direction cosine matrix is ​​then used to convert the satellite's velocity vector in the northeast celestial coordinate system to the aircraft's body coordinate system. Specifically, the northeast velocity of 600 kilometers per hour is combined with the current pitch angle of -10 degrees, roll angle of 45 degrees, and yaw angle of 30 degrees to calculate the aircraft's X-axis velocity of 420 kilometers per hour and Y-axis velocity of 310 kilometers per hour. The fusion process then combines the linear velocity of the aircraft coordinate system with the gyroscope's angular velocity, calculating a roll angular acceleration of 8000 degrees per second squared. The cross product of the angular velocity vector and the linear velocity vector yields a trajectory curvature of 0.28. The resulting feature vector [8000, 0.28] combines key metrics such as angular acceleration and trajectory curvature, accurately characterizing the aircraft's high-speed roll-down maneuver.

[0079] The above-mentioned 101 overall solution eliminates the information fragmentation problem caused by sampling frequency differences and coordinate system mismatch in traditional methods by uniformly processing the time and space benchmarks of multi-source sensor data; the generated fusion feature vector can more comprehensively capture the dynamic coupling characteristics of the flight attitude, providing a high-precision, low-noise input basis for subsequent mutation detection, and improving the reliability of state judgment from the source.

[0080] 102. Using an inertial navigation unit array to collect three-dimensional angular motion data, performing non-uniform weight distribution on high-frequency components in the three-dimensional angular motion data, and outputting a steady-state estimate via a dynamic weighted average processor;

[0081] Optionally, step 102 may specifically include the following steps:

[0082] 1021. Synchronously collect three-dimensional angular motion data through multiple spatially distributed inertial measurement nodes, each inertial measurement node outputting angular velocity components and linear acceleration components along three orthogonal axes;

[0083] 1022. Perform motion feature analysis on the angular velocity components output by each inertial measurement node, identify motion data segments containing rapidly changing features, and mark the identified motion data segments with dynamic feature identifiers;

[0084] 1023. Assign a first weight coefficient to the marked motion data segment and a second weight coefficient to the unmarked motion data segment according to the dynamic feature identifier, wherein the first weight coefficient is greater than the second weight coefficient;

[0085] 1024. Input the weighted angular velocity components into a dynamic weighted average processor, which performs the following processing: weighted fusion of the same axial angular velocity components from different inertial measurement nodes; motion consistency verification of each fused axial angular velocity component; and outputting the verified three-dimensional angular motion fusion result as a stable state estimate.

[0086] In the above scheme, the inertial navigation unit array refers to a spatially distributed sensor network that reflects the global motion state of the aircraft and can be used to build a motion perception system that is resistant to local interference. Three-dimensional angular motion data is a vector set representing the rotational dynamics of the aircraft. It includes instantaneous angular velocity components in the three orthogonal directions of pitch, roll, and yaw, and can be used to determine attitude change trends and maneuver intensity. High-frequency components specifically refer to fluctuations in motion data that reflect rapid mechanical vibrations or instantaneous impacts. They include millisecond-level mutations in angular velocity and can be used to identify the boundary between sensor noise and true maneuvers. Non-uniform weight allocation is a processing mechanism that dynamically adjusts data credibility based on motion characteristics. It includes a differentiated coefficient strategy that assigns high weights to rapidly changing segments and low weights to stable segments, enhancing valid signals and suppressing noise interference. The dynamic weighted average processor is a computational unit that optimizes multi-source data fusion. It includes dual modules for weighted fusion and physical consistency verification, and can output reliable motion estimates that are resistant to abnormal disturbances. The steady-state estimator is a description of the motion state after filtering out high-frequency noise. It includes smoothed three-dimensional angular velocity values ​​verified by global sensor collaboration and can be used to provide a highly robust attitude control reference input.

[0087] In the embodiment of the present application, after the inertial navigation unit array is started in step 1021, multiple spatially distributed inertial measurement nodes synchronously collect three-dimensional angular motion data. Each node independently outputs complete motion information containing three orthogonal axes: angular velocity components and linear acceleration components. For example, when the aircraft performs a sharp turn, the nose node A records the roll angular velocity. , the left wing node B detects the roll angular velocity , the right wing node C outputs abnormal roll angular velocity due to local vibration , all data are synchronously time-stamped and transmitted to the processor at a frequency of 1000Hz.

[0088] Next, the processor in step 1022 performs motion feature analysis on the angular velocity components of each node: a sliding window difference algorithm is used to calculate the angular acceleration changes of 10 consecutive frames of data. If the slope of a data segment exceeds a preset threshold, it is marked as a dynamic feature segment and a red mark is added. Taking the aforementioned sharp turn as an example, the roll angular velocity of the left wing node B changes from 0.5 to 0.52 seconds. Jump to , the instantaneous acceleration is calculated to be Superthreshold , the 0.02 second data segment is marked in red; while the right-wing node C only receives becomes , acceleration If the threshold is not reached, no mark will be made.

[0089] Then, in step 1023, non-uniform weights are assigned based on the dynamic feature identifiers: the data of the red dynamic feature segment is assigned a high weight coefficient of 0.9, i.e., the first weight coefficient, and the unmarked stable segment is assigned a low weight coefficient of 0.1, i.e., the second weight coefficient. Continuing with the previous example, the 0.5-0.52 second dynamic segment of the left wing node B is assigned a weight of 0.9 because it contains a red identifier, and the rest of the time segment is assigned a weight of 0.1; the right wing node C is unmarked throughout, so all data has a weight of 0.1; the nose node A appears between 0.5-0.51 seconds. Jump, calculate acceleration When the threshold is reached, the 0.01 second segment is marked red and assigned a weight of 0.9.

[0090] Finally, the dynamic weighted average processor in step 1024 first performs weighted fusion on the same axial data: take the roll angular velocity of the three nodes, according to the formula Output fusion value, where 、 、 are the rolling angular velocities of the three nodes respectively, 、 、 is the weight assigned to it.

[0091] For example, the roll angular velocity of node A Weight 0.9, B node roll angular velocity Weight 0.9, C node roll angular velocity The weight is 0.1, calculated by the formula Then the motion consistency check is performed to verify the physical rationality of the three-axis angular velocity after fusion. The final output is the stable state estimate that has passed the check, for example , where the low-speed data of abnormal node C is suppressed by the high-weight mechanism.

[0092] In actual applications, when a certain flight simulator simulates a sharp turn in a fighter jet, it first collects motion data simultaneously using three measuring instruments installed on the nose, left wing, and right wing. Instrument A records a left-right twisting speed of 150 degrees per second, instrument B detects 145 degrees per second, and instrument C only displays 10 degrees per second due to mechanical vibration interference. All data are marked with time points accurate to one thousandth of a second. The intensity of the data change is then analyzed: Instrument B displays a twisting speed of 30 degrees per second at 0.5 seconds, and suddenly rises to 180 degrees per second at 0.52 seconds. The calculated change intensity is 7500 degrees per second squared, exceeding the preset alarm value of 1000 degrees per second squared. Therefore, this time period is marked as a critical dynamic segment. During the same period, instrument C only changes from 8 degrees per second to 12 degrees per second, with an intensity of 200 degrees per second squared, which does not meet the standard and is therefore not marked. Then, the importance coefficients were assigned to the marked segments: the key segment from 0.5 to 0.52 seconds for instrument B was assigned a high coefficient of 0.9, and the rest of the segment was assigned a low coefficient of 0.1; instrument C had no markers throughout the entire process, and all data had a coefficient of 0.1; instrument A increased its speed from 170 degrees per second to 175 degrees per second from 0.5 to 0.51 seconds, with an intensity of 500 degrees per second squared, and its coefficient remained at 0.1. To perform data synthesis: the twisting speed at 0.505 seconds was calculated. Finally, check whether the relationship between the three directions of movement is reasonable, such as whether the combination of left and right twisting and up and down rotation exceeds the safe range. After confirmation, output the stable motion state value , where the interference data impact of instrument C is greatly reduced.

[0093] The overall solution described above, 102, synchronously collects three-dimensional angular motion data from multiple spatially distributed inertial measurement nodes. It dynamically assigns high weights to rapid maneuvering segments and low weights to stable segments. A dynamic weighted average processor fuses the multi-node data and verifies its physical plausibility, ultimately outputting a steady-state estimate. This method effectively suppresses local sensor anomalies and mechanical vibration interference, improving data consistency while preserving true maneuvering characteristics. This provides a highly robust motion reference input for subsequent state switching decisions, resolving the maneuvering feature flooding and noise sensitivity issues associated with traditional uniform weighting.

[0094] 103. Constructing a dual-threshold characteristic boundary based on the degree of deviation of the characteristic vector relative to the stable state estimate, and determining a maneuver mutation timing and generating a switching trigger signal when a continuous state transition trajectory crosses the dual-threshold characteristic boundary;

[0095] Optionally, step 103 may specifically include the following steps:

[0096] 1031. Convert the pitch rate and roll acceleration parameters included in the eigenvector into first eigencoordinates, convert the platform displacement and velocity parameters included in the stable state estimate into second eigencoordinates, and establish a dynamic projection relationship between the first eigencoordinates and the second eigencoordinates.

[0097] 1032. Setting an adjustable boundary threshold in the dynamic projection relationship and dynamically adjusting the boundary threshold range based on historical motion data to form a dual-threshold feature boundary including an upper boundary and a lower boundary;

[0098] 1033. Calculate the projection distance between the feature coordinates at the current moment and the stable state coordinates, and record the projection distance change trajectory for multiple consecutive sampling periods, so as to mark the crossing point and crossing direction of the trajectory crossing the dual-threshold feature boundary through the projection distance change trajectory;

[0099] 1034. When the trajectory is detected to cross the same feature boundary three times in a row, the timestamp of the last crossing is recorded as the mutation moment, and a switching trigger signal including the mutation moment, crossing point, and crossing direction is generated.

[0100] In the above scheme, the dual-threshold characteristic boundary refers to a dynamic monitoring area that reflects the risk of sudden changes in the motion state. It includes upper and lower warning lines that are adaptively adjusted based on historical data and can be used to accurately determine the critical point of abnormal deviation in the aircraft's maneuvering behavior. The continuous state transition trajectory represents a data sequence that represents the temporal variation of the motion state. It contains a line connecting the projected distances between the characteristic coordinates and the stable coordinates at multiple consecutive sampling moments and can be used to track the continuous evolution of the maneuvering trend. The switching trigger signal is a decision instruction that indicates the occurrence of a sudden change in the maneuver. It contains three core parameters: the precise timestamp of the change moment, the coordinates of the boundary crossing position, and the crossing direction. It can be used to drive the non-disturbance switching control of the state transition matrix coefficients.

[0101] In the embodiment of the present application, the pitch angular rate and roll angular acceleration in the feature vector are first converted into the first feature coordinates through step 1031: the angular rate is divided by 100 through linear scaling to obtain the X-axis value, and the angular acceleration is divided by 1000 to obtain the Y-axis value, which are combined into a coordinate point. For example, the pitch angular rate detection value is 120 degrees per second and the roll angular acceleration is 5000 degrees per second squared, which are converted into the first feature coordinate point (1.2, 5.0). At the same time, the platform displacement and velocity in the stable state estimate are converted into the second feature coordinates: the displacement is multiplied by 0.5 and the velocity is divided by 100, which are combined into the second feature coordinate point. For example, the platform moves 3 meters forward and backward with a horizontal speed of 300 kilometers per hour, which is converted into the second feature coordinate point (1.5, 3.0). Then, a dynamic projection relationship is established: the least squares method is used to fit the mapping equation of the two coordinate points. For example, the calculated projection formula y=2.5x means that for every 1 unit increase in the feature coordinate, the stable coordinate increases by 2.5 units.

[0102] Next, step 1032 dynamically constructs a dual-threshold feature boundary based on historical motion data: extract the past 100 sets of projection distance data and calculate the average distance value and fluctuation range Initially set the upper boundary , the lower boundary When continuous violent maneuvers are detected, the coefficient is adaptively increased from 1.5 to 2.0, forming a dynamically expanding monitoring zone. For example, extracting the past 100 sets of projection distance data to obtain an average distance value μ = 2.0 and a fluctuation range σ = 0.4, then the initial upper boundary , the initial lower boundary When the distance of the maneuver is greater than 2.5 for five consecutive times, the adaptive coefficient increases to 2.0, and the boundaries are updated to the upper boundary U=2.0+2.0×0.4=2.8 and the lower boundary L=2.0-2.0×0.4=1.2.

[0103] Then, calculate the projection distance between the current characteristic coordinate and the stable coordinate through step 1033: obtain the characteristic vector conversion value and the stable state conversion value, substitute them into the dynamic projection equation y=2.5x to calculate the theoretical value, and calculate the projection distance based on the deviation between the theoretical value and the actual value. Then, the distance values ​​of 10 consecutive sampling points are continuously recorded to form a trajectory sequence. Using the trajectory crossing judgment algorithm, when two adjacent points satisfy dt≤U and dt+1>U, the trajectory is judged to cross the upper boundary upward, and the original data of the crossing moment and the direction "upward" are recorded. For example, the distance value forms a trajectory sequence as ,in and , it is determined to be crossing the boundary upwards, and the crossing time is recorded Corresponding pitch angle rate , platform speed is 385km / h and crossing direction is "upward".

[0104] Finally, a continuous crossing determination mechanism is established through step 1034: the crossing counter is initialized and set to 1 when a single crossing is detected. Subsequent trajectories are continuously monitored: if the next sampling point crosses in the same direction, the counter increases to 2; if there is also a crossing in the same direction at the next moment, the counter reaches 3. When the condition of three consecutive crossings in the same direction is met, the exact time of the last crossing timestamp is extracted, and the original data of the crossing point at that moment and the direction "up" are synchronously recorded, and finally a triplet switching trigger signal is generated. For example, if the condition of three consecutive crossings in the same direction is met, the timestamp of the last crossing is extracted. seconds, and simultaneously record the pitch angular velocity of the crossing point at that moment as rate , the platform speed is 385km / h and the direction is "up", and then a triplet switching trigger signal is generated: mutation time: 0.850 seconds, crossing point coordinates: (152,385), crossing direction: upward.

[0105] In actual applications, when a certain flight simulator reproduces a fighter jet's dive recovery maneuver, it first scales the pitch angular velocity detection value of -90 degrees per second in the eigenvector to the first coordinate X-axis value of -0.9. At the same time, it scales the platform sinking speed of 200 kilometers per hour in the steady-state estimate to the second coordinate Y-axis value of 2.0. Through data fitting, the projection relationship equation between the two is established: Y equals -2.2 times X. Then, based on 100 sets of historical motion data, the average projection distance of 1.8 and the fluctuation range of 0.3 are calculated. Based on this, the initial monitoring zone is set: the upper boundary , lower boundary Then the current projection distance is detected. The moment is 1.2, which is below the upper boundary. It rose to 2.3 at that moment, exceeding the upper limit of 2.25, and the trajectory was determined to have crossed the monitoring zone upwards and the coordinates of the crossing point were recorded (-0.95, 2.4). Time distance 2.4, When the distance 2.5 crosses the same boundary twice in a row, When the third crossing is equal to 0.6 seconds, a complete switching trigger signal including the mutation time "0.6 seconds", the crossing position (-0.95, 2.4) and the direction "up" is generated, driving the system to execute the state matrix switching.

[0106] The overall solution described in 103 establishes a dynamic projection relationship between the eigenvector and the steady-state estimate, adaptively generates a dual-threshold monitoring boundary based on historical data, and tracks the crossing behavior of continuous-state trajectories. When a trajectory crosses the same boundary three times in a row, the timing of the maneuver mutation is accurately determined, and a switching trigger signal containing a timestamp, position, and direction is generated. This method effectively avoids false triggering due to noise, solves the problem of inaccurate or delayed switching timing caused by traditional single-threshold determination, and provides a highly reliable decision-making basis for smooth state matrix switching, significantly improving the control stability and safety of the aircraft during intense maneuvers.

[0107] 104. In response to the switching trigger signal, perform a smooth switching operation of state transfer matrix coefficients.

[0108] Optionally, the process of performing the smooth switching operation of the state transfer matrix coefficients in step 104 includes:

[0109] 1041. Obtain an old coefficient set of a current state transfer matrix and a new coefficient set of a target state transfer matrix;

[0110] 1042. Taking the switching time point as a reference point, based on the old coefficient set and the current state vector, calculate a boundary condition vector that satisfies the continuity of the output acceleration command and its first-order derivative;

[0111] Among them, step 1042 may specifically include the following processes: determining the switching time point of the state transfer matrix; obtaining the first state vector corresponding to the switching time point, the first state vector including a first position component and a first velocity component; using the first velocity component as the output acceleration component; based on the low-frequency coefficient and the high-frequency coefficient in the old coefficient set, combined with the first state vector, calculating the output acceleration change component: multiplying the first position component by the low-frequency coefficient to obtain a first intermediate value; multiplying the first velocity component by the high-frequency coefficient to obtain a second intermediate value; adding the first intermediate value and the second intermediate value and taking the negative value to obtain the output acceleration change component; combining the output acceleration component and the output acceleration change component in a predetermined order to form a boundary condition vector.

[0112] 1043. Construct a state transition relationship equation using the new coefficient set, use the boundary condition vector as the right-hand side of the equation, and solve the state transition relationship equation to obtain a post-switching state vector.

[0113] Among them, step 1043 may specifically include the following processes: constructing a coefficient matrix based on the low-frequency coefficients and high-frequency coefficients in the new coefficient set; setting the first row and first column elements of the coefficient matrix to a fixed value of 0, setting the first row and second column elements of the coefficient matrix to a fixed value of 1, setting the second row and first column elements of the coefficient matrix to the negative value of the low-frequency coefficient, and setting the second row and second column elements of the coefficient matrix to the negative value of the high-frequency coefficient; using the boundary condition vector as the right-hand constant of the linear relationship equation; solving the linear relationship equation with the coefficient matrix as the coefficient matrix and the boundary condition vector as the right-hand constant to obtain a state vector after switching, wherein the state vector after switching includes a second position component and a second velocity component.

[0114] 1044. Synchronously load the switched state vector and the new coefficient set into the motion simulation filter to complete smooth switching of the state transfer matrix coefficients.

[0115] Among them, step 1044 may specifically include the following processes: writing the second position component of the state vector after switching into the position component storage space of the dynamic simulation filter; writing the second velocity component of the state vector after switching into the velocity component storage space of the dynamic simulation filter; writing the low-frequency coefficients in the new coefficient set into the low-frequency coefficient storage space of the dynamic simulation filter; writing the high-frequency coefficients in the new coefficient set into the high-frequency coefficient storage space of the dynamic simulation filter; when completing the write operation of all storage spaces, the dynamic simulation filter starts running refresh based on the value of the current storage space to complete the switching of the state transfer matrix coefficients.

[0116] In the above scheme, the old coefficient set refers to a combination of dynamic parameters that reflects the current motion simulation characteristics. It includes low-frequency smoothing coefficients and high-frequency response coefficients that maintain the existing dynamic effect and are used to ensure the stability of the motion output before the switch. The new coefficient set represents the optimized parameter set for the target maneuver scenario. It includes low-frequency attenuation coefficients and high-frequency enhancement coefficients adjusted to accommodate sudden maneuvering requirements and is used to achieve precise adaptation of the motion simulation after the switch. The boundary condition vector is a mathematical boundary condition that ensures motion continuity. It includes the acceleration baseline value and its rate of change threshold at the switching instant, which can be used to eliminate motion jumps during state switching. The state transition relationship equation describes the matrix mathematical model of the motion state transition. It includes a transfer matrix constructed from the new coefficients and a right-hand constant term composed of the boundary conditions. It can be used to solve the post-switch motion state under the acceleration continuity constraint. The motion simulation filter is an embedded hardware unit that performs motion state calculations. It contains position memory, velocity memory, and coefficient configuration registers and is used to generate smooth motion output without jerkiness or jitter.

[0117] In the embodiment of the present application, steps 1041 to 1044 implement smooth switching of the state transfer matrix coefficients through the following process:

[0118] First, taking the transfer function of a second-order system as an example, its general form can be written as:

[0119] ;

[0120] in and These are the state transfer matrix coefficients, which together determine how the platform responds to control commands. The corresponding state transfer matrix can be set as:

[0121] ;

[0122] in , For status, is the input acceleration, Output acceleration command.

[0123] The direct switching of the state transfer matrix coefficients will lead to discontinuous output problems: if The state transfer matrix coefficients are From the old coefficient Mutation to new coefficients , and the state quantity remains unchanged, the output acceleration There is usually a jump.

[0124] Based on this, step 1041 is executed to obtain the old coefficient set of the current state transfer matrix and the new coefficient set of the target state transfer matrix, that is, at the switching time pointt Get the old coefficient set currently in effect and the new coefficient set to be switched This step is the preparation for the switch and establishes the parameter basis for the subsequent smooth switch. Ensure that the system masters the motion control rules before and after the switch to avoid interruption of control logic during the switch. For example, step 1041 simultaneously obtains the old coefficients during smooth flight: , And the new coefficients for maneuver mode: , These two sets of parameters.

[0125] Next, execute step 1042. The first step is to determine the switching time point of the state transfer matrix. , this time point serves as the synchronous triggering moment for switching the state and state transfer matrix coefficients.

[0126] The second step is to obtain the first state vector corresponding to the switching time point t, which contains the first position component representing the displacement-related state. and the first velocity component representing the velocity-related state According to the state transfer matrix , the first velocity component Directly as the output acceleration component .

[0127] In the third step, based on the low-frequency coefficients and high-frequency coefficients in the old coefficient set and combined with the first state vector, the output acceleration change component is calculated, the first position component is multiplied by the low-frequency coefficient to obtain a first intermediate value; the first velocity component is multiplied by the high-frequency coefficient to obtain a second intermediate value; the first intermediate value and the second intermediate value are added and the negative value is taken to obtain the output acceleration change component.

[0128] Specifically, the low-frequency coefficient is processed as follows: the first position component ( ) and the low-frequency coefficients in the old coefficient set that control motion smoothness Multiply and take the negative value to get the first product term , to characterize the low-frequency cumulative effect of displacement on acceleration changes.

[0129] High frequency coefficient processing: The first velocity component representing the velocity state ( ) and the high-frequency coefficients in the old coefficient set that control the motion response speed Multiply and take the negative value to get the second product term , to characterize the high-frequency transient impact of velocity on acceleration changes.

[0130] Then a summation operation is performed: sum the first product term and the second product term , get the output acceleration change component , characterizes the output acceleration change rate at the switching time point, that is, , whose mathematical basis comes from the state transfer equation ,When the input command u(t) = 0 at the switching time point t, the rate of ,acceleration change is directly determined by the linear combination of ,displacement and velocity.

[0131] The fourth step is to convert the output acceleration component of the actual acceleration value at the switching moment obtained by the above calculation into The output acceleration change component that represents the acceleration change trend Combine them in a predetermined order to form the boundary condition vector b, namely:

[0132] ;

[0133] The core value of the boundary condition vector b is to ensure smooth output of the switching point by using dual physical constraints, where Ensure the continuity of acceleration value, so that the output acceleration of the new and old systems are completely consistent at the moment of switching, avoiding "value jump". Ensure the continuity of the acceleration change rate, make the acceleration change trend transition smoothly, and eliminate "sudden changes in change trend".

[0134] Next, step 1043 is executed. The first step is based on the low-frequency coefficients in the new coefficient set. and high frequency coefficients , construct the coefficient matrix Specifically, the first row and first column element of the coefficient matrix is ​​set to a fixed value of 0, the first row and second column element of the coefficient matrix is ​​set to a fixed value of 1, and the second row and first column element of the coefficient matrix is ​​set to the negative value of the low-frequency coefficient. , set the second row and second column element of the coefficient matrix to the negative value of the high frequency coefficient ,Right now:

[0135] ;

[0136] Among them, the second row Define the law of how acceleration changes with displacement and velocity, which together constitute the mathematical skeleton of the new motion control rules.

[0137] The second step is to use the boundary condition vector b as the right-hand constant of the linear relationship equation, solve the linear relationship equation with the coefficient matrix as the coefficient matrix and the boundary condition vector as the right-hand constant, and obtain the state vector after switching. Contains the second position component and the second velocity component ,Right now:

[0138] ;

[0139] in, is the state vector after switching to be solved, and b is the boundary condition vector calculated in step 1042.

[0140] Assumptions Reversible, that is, determinant When is always satisfied, the state vector after switching is solved by deriving the equation:

[0141] ;

[0142] This step is the mathematical core of the new state vector, the equation Ensure that the output after switching and derivatives Continuous with the switching, solve Adjust the state to compensate for the coefficient change. The essence is to use the state quantity Accurate correction, offset coefficient switching The resulting motion mutation makes the platform run smoothly at the moment of coefficient switching as if it has not undergone any change in control rules, completely avoiding the "motion setback" phenomenon of traditional switching.

[0143] Finally, execute step 1044 to write the second position component of the switched state vector into the position component storage space of the dynamic simulation filter; write the second velocity component of the switched state vector into the velocity component storage space of the dynamic simulation filter; write the low-frequency coefficients in the new coefficient set into the low-frequency coefficient storage space of the dynamic simulation filter; write the high-frequency coefficients in the new coefficient set into the high-frequency coefficient storage space of the dynamic simulation filter; when the write operation of all storage spaces is completed, the dynamic simulation filter starts running refresh based on the value of the current storage space to complete the switching of the state transfer matrix coefficients.

[0144] Specifically, the compensation state vector obtained by solving step 1043 is With the new coefficient set At the same time, write the physical storage unit of the dynamic analog filter: the second position component , that is, the compensated displacement is written into the position storage register of the filter, which records the actual moving distance of the platform; the second velocity component , that is, the compensated speed is written into the speed storage register, which controls the acceleration change rate; the low-frequency coefficient Write the smooth coefficient register to determine the platform's follow-up accuracy for the smooth instruction; Write to the response coefficient register to control the platform's response speed to sudden actions, achieving smooth transitions and eliminating output jumps.

[0145] The filter starts at the switching time point t, triggering the hardware synchronous refresh instruction, and uses the new state vector and new coefficient set to run the state transfer matrix. The system state transfer equation switches to:

[0146] ;

[0147] When the switching condition of the filter coefficient is triggered, it is calculated according to the above formula , and use it as the new state of the filter, so the output There will be no jumps or spikes.

[0148] For example, under certain working conditions, the initial filter coefficients , when the time reaches 0.8 seconds, switch to By comparing the effect of directly switching the coefficients, it can be found that the coefficients switched by this method have smooth output and no jump occurs.

[0149] In actual application, a flight simulator triggers a switch when reproducing a fighter dive recovery maneuver for 0.5 seconds: the system synchronously obtains the old coefficient set currently in effect, and the stable control coefficient , maneuver response coefficient and the target new coefficient set , extract the state vector at the switching moment ; Calculate boundary condition vector: acceleration reference value , acceleration rate of change ,generate ; Construct a new coefficient matrix And solve Compensation status Atomized loading filter: write 18.75m to the position register, 50m / s to the speed register, 1.8 to the smoothness coefficient register, and 0.9 to the response coefficient register. After triggering the synchronous refresh, the new equation is run. The output acceleration smoothly transitions from 50m / s without any jumps, completely avoiding the drastic jumps caused by directly switching coefficients in traditional solutions.

[0150] The above-mentioned overall solution of 104 synchronously obtains the coefficients of the new and old state transfer matrices, accurately calculates the continuity boundary conditions of the acceleration and its rate of change based on the motion state at the switching moment, uses the new coefficients to construct the state transfer equation to solve the compensation state vector, and finally uses hardware-level atomic operations to synchronously load the four types of parameters, namely displacement state quantity, velocity state quantity, smooth control coefficient and maneuver response coefficient, into the filter register within milliseconds and trigger instantaneous refresh. From a mathematical level, it forces the zero-order and first-order continuity constraints of the acceleration value and its rate of change to be met, completely eliminating the output jump, mechanical vibration and hardware damage problems caused by asynchronous parameter switching in traditional solutions, so that the flight simulator can achieve natural motion transition without shock and jitter during violent maneuvers.

[0151] The following is a complete example for steps 101 to 104. Figure 2 As shown in the figure, when a certain type of flight simulator reproduces a fighter jet barrel roll maneuver, it first collects flight attitude data through the onboard inertial measurement unit. The gyroscope outputs a peak roll angular velocity of 180 degrees per second, and the satellite positioning module feedbacks a horizontal velocity of 550 kilometers per hour and a climb rate of 30 meters per second. The two are then temporally coupled, and the 10Hz satellite data is aligned to the 1000Hz gyroscope timestamp through linear interpolation. The direction cosine matrix is ​​then used to convert the satellite's northeast celestial coordinate system velocity into the aircraft coordinate system, obtaining the aircraft's X-axis velocity of 420 kilometers per hour and Z-axis velocity of 180 kilometers per hour. Finally, the roll angular acceleration of 8500 degrees per second squared and the trajectory curvature of 0.32 are integrated and calculated to generate the feature vector [8500, 0.32], which accurately represents the compound maneuver trend of the fuselage's high-speed rotation and upward pitch.

[0152] Next, the inertial navigation nodes distributed on the nose and wings synchronously collected three-dimensional angular motion data. The left wing node detected that the roll angular velocity jumped from 40 degrees per second to 175 degrees per second within 0.3 seconds, marked it as a dynamic feature segment and assigned a high weight of 0.9; the right wing node output abnormally stable data due to sensor failure and was assigned a low weight of 0.1; the dynamic weighted processor fused and calculated the data at the 0.25-second moment: the head node value was 160 degrees per second × weight 0.9, the left wing node was 170 degrees per second × 0.9, and the right wing node was 15 degrees per second × 0.1. The weighted average was 156.6 degrees per second. After motion consistency verification, the stable state estimate [156.6 degrees per second, 38.2 degrees per second, -3.1 degrees per second] was output, effectively suppressing interference from the right wing node.

[0153] Then, the roll angular acceleration of 8500 degrees per second squared in the eigenvector is converted to the first eigencoordinate value of 8.5, and the horizontal speed of 420 kilometers per hour in the steady state estimate is converted to the second eigencoordinate value of 4.2; the projection relationship equation is established ; Based on historical data average , standard deviation Generate double threshold boundaries 、 ; Monitor the projection distance of three consecutive frames Crossing the upper boundary, the third crossing time 𝑡 = 0.6 seconds is determined to be the mutation opportunity, and a switching signal is generated.

[0154] Finally, in response to the 0.6 second switching signal, the old coefficient set is obtained With the new coefficient set ; Extract the switching point state vector ; Calculate boundary condition vector: acceleration value , rate of change ,have to ; Construct a new coefficient matrix And solve Atomic write filter register: displacement 28 meters, speed 60 meters per second, smoothness coefficient 1.5, response coefficient 1.2. After synchronous refresh, it runs according to the new equation, and the output acceleration smoothly transitions from 60 meters per second, completely avoiding the theoretical jump of 215 meters per second squared caused by direct switching of traditional solutions.

[0155] Figure 3 The present invention provides a schematic diagram of a system for determining a smooth switching state transfer matrix coefficient. Figure 3 As shown, the system includes:

[0156] an acquisition module 31 for acquiring flight attitude change data through an inertial measurement unit, performing spatiotemporal coupling processing on gyroscope angular velocity data in the flight attitude change data and satellite positioning data to generate a feature vector reflecting the attitude change trend;

[0157] an output module 32 for collecting three-dimensional angular motion data using an inertial navigation unit array, performing non-uniform weight distribution on high-frequency components in the three-dimensional angular motion data, and outputting a steady-state estimate via a dynamic weighted average processor;

[0158] A generating module 33 is configured to construct a dual-threshold characteristic boundary based on the degree of deviation of the characteristic vector relative to the stable state estimate, and to determine a maneuver mutation timing and generate a switching trigger signal when a continuous state transition trajectory crosses the dual-threshold characteristic boundary;

[0159] The execution module 34 is configured to execute a smooth switching operation of the state transfer matrix coefficients in response to the switching trigger signal.

[0160] Figure 3 The system for determining the coefficients of the smooth switching state transfer matrix can be executed Figure 1The implementation principle and technical effects of the method for determining the smooth switching state transfer matrix coefficients described in the illustrated embodiment will not be elaborated upon. The specific manner in which the various modules and units perform operations in the system for determining the smooth switching state transfer matrix coefficients in the above-described embodiment have been described in detail in the embodiments of the method and will not be further elaborated upon here.

[0161] In one possible design, Figure 3 A system for determining the coefficients of a smooth switching state transfer matrix of the embodiment shown can be implemented as a computing device, such as Figure 4 As shown, the computing device may include a storage component 41 and a processing component 42;

[0162] The storage component 41 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 42 .

[0163] The processing component 42 is used for the above Figure 1 The embodiment provides a method for determining coefficients of a smooth switching state transfer matrix.

[0164] The processing component 42 may include one or more processors to execute computer instructions to perform all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0165] The storage component 41 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0166] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0167] The input / output interface provides an interface between the processing component and the peripheral interface module, which can be an output device, an input device, etc.

[0168] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0169] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0170] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The embodiment shown is a method for determining coefficients of a smooth switching state transfer matrix.

[0171] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0172] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0173] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer or server) to execute the methods described in each embodiment or certain portions of the embodiments.

[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for determining the coefficients of a smooth switching state transfer matrix, characterized in that: include: Acquiring flight attitude change data through an inertial measurement unit, performing spatiotemporal coupling processing on gyroscope angular velocity data in the flight attitude change data and satellite positioning data to generate a feature vector reflecting the attitude change trend; An inertial navigation unit array is used to collect three-dimensional angular motion data, and non-uniform weight distribution is performed on the high-frequency components in the three-dimensional angular motion data. The stable state estimate is output through a dynamic weighted average processor. The three-dimensional angular motion data represents a vector set of the rotational dynamic characteristics of the body, including the instantaneous angular velocity components in the three orthogonal directions of pitch, roll, and yaw, which are used to resolve attitude change trends and maneuver intensity. Based on the degree of deviation of the characteristic vector from the stable state estimate, a dual-threshold characteristic boundary is constructed. When the continuous state transition trajectory crosses the dual-threshold characteristic boundary, the timing of the maneuver mutation is determined and a switching trigger signal is generated. The continuous state transition trajectory represents a data sequence of temporal changes in the motion state, and includes a line connecting the projected distances of the characteristic coordinates and the stable coordinates at multiple consecutive sampling moments, which is used to track the continuous evolution of the maneuver trend. In response to the switching trigger signal, performing a smooth switching operation of the state transfer matrix coefficients; The process of performing a smooth switching operation of the state transfer matrix coefficients includes: Obtain the old coefficient set of the current state transfer matrix and the new coefficient set of the target state transfer matrix; Taking the switching time point as a reference point, based on the old coefficient set and the current state vector, a boundary condition vector that satisfies the continuity of the output acceleration command and its first-order derivative is calculated; Constructing a state transition relationship equation using the new coefficient set, taking the boundary condition vector as the right-hand side term of the equation, and solving the state transition relationship equation to obtain a state vector after switching; The switched state vector and the new coefficient set are synchronously loaded into the motion simulation filter to complete the smooth switching of the state transfer matrix coefficients.

2. The method for determining the smooth switching state transfer matrix coefficients according to claim 1, characterized in that: The step of calculating the boundary condition vector satisfying the continuity of the output acceleration command and its first-order derivative based on the old coefficient set and the current state vector with the switching time point as the reference point includes: Determine the switching time point of the state transfer matrix; Acquire a first state vector corresponding to the switching time point, where the first state vector includes a first position component and a first velocity component; Taking the first velocity component as the output acceleration component; Based on the low-frequency coefficients and high-frequency coefficients in the old coefficient set and in combination with the first state vector, an output acceleration change component is calculated: the first position component is multiplied by the low-frequency coefficient to obtain a first intermediate value; the first velocity component is multiplied by the high-frequency coefficient to obtain a second intermediate value; the first intermediate value is added to the second intermediate value and the negative value is taken to obtain the output acceleration change component; The output acceleration component and the output acceleration change component are combined in a predetermined order to form a boundary condition vector.

3. The method for determining the smooth switching state transfer matrix coefficients according to claim 1, wherein: The step of constructing a state transition relationship equation using the new coefficient set, taking the boundary condition vector as the right-hand side of the equation, and solving the state transition relationship equation to obtain a state vector after switching includes: Constructing a coefficient matrix based on the low-frequency coefficients and the high-frequency coefficients in the new coefficient set; Set the first row and first column elements of the coefficient matrix to a fixed value of 0, set the first row and second column elements of the coefficient matrix to a fixed value of 1, set the second row and first column elements of the coefficient matrix to the negative value of the low-frequency coefficient, and set the second row and second column elements of the coefficient matrix to the negative value of the high-frequency coefficient; Using the boundary condition vector as a constant on the right side of the linear relationship equation; A linear relationship equation with the coefficient matrix as the coefficient matrix and the boundary condition vector as the right-hand constant is solved to obtain a state vector after switching, wherein the state vector after switching includes a second position component and a second velocity component.

4. The method for determining the smooth switching state transfer matrix coefficients according to claim 3, characterized in that: The step of synchronously loading the switched state vector and the new coefficient set into the motion simulation filter to complete the smooth switching of the state transfer matrix coefficients includes: Writing the second position component of the switched state vector into the position component storage space of the dynamic simulation filter; Writing the second velocity component of the switched state vector into the velocity component storage space of the dynamic simulation filter; Writing the low-frequency coefficients in the new coefficient set into the low-frequency coefficient storage space of the dynamic simulation filter; Writing the high frequency coefficients in the new coefficient set into the high frequency coefficient storage space of the dynamic simulation filter; When the writing operation of all storage spaces is completed, the dynamic simulation filter starts running refresh based on the value of the current storage space, and the switching of the state transfer matrix coefficients is completed.

5. The method for determining the smooth switching state transfer matrix coefficients according to claim 1, characterized in that: The method comprises: collecting three-dimensional angular motion data using an inertial navigation unit array, performing non-uniform weight distribution on high-frequency components in the three-dimensional angular motion data, and outputting a steady-state estimate via a dynamic weighted average processor, including: Three-dimensional angular motion data is collected synchronously through multiple inertial measurement nodes distributed in space. Each inertial measurement node outputs angular velocity components and linear acceleration components along three orthogonal axes. Perform motion feature analysis on the angular velocity components output by each inertial measurement node, identify motion data segments containing rapidly changing features, and mark the identified motion data segments with dynamic feature identifiers; According to the dynamic feature identifier, assigning a first weight coefficient to the marked motion data segment and assigning a second weight coefficient to the unmarked motion data segment, wherein the first weight coefficient is greater than the second weight coefficient; The weighted angular velocity components are input into a dynamic weighted average processor, which performs the following processing: weighted fusion of the same axial angular velocity components from different inertial measurement nodes; motion consistency verification of each fused axial angular velocity component; and output of the verified three-dimensional angular motion fusion result as a stable state estimate.

6. The method for determining the smooth switching state transfer matrix coefficients according to claim 1, characterized in that: The step of constructing a dual-threshold characteristic boundary based on the degree of deviation of the characteristic vector from the stable state estimate, and determining a maneuver mutation timing and generating a switching trigger signal when a continuous state transition trajectory crosses the dual-threshold characteristic boundary, includes: Convert the pitch rate and roll acceleration parameters contained in the eigenvector into the first eigencoordinate, convert the platform displacement and velocity parameters contained in the stable state estimate into the second eigencoordinate, and establish a dynamic projection relationship between the first and second eigencoordinates; An adjustable boundary threshold is set in the dynamic projection relationship, and the boundary threshold range is dynamically adjusted according to historical motion data to form a dual-threshold feature boundary including an upper boundary and a lower boundary; Calculating the projection distance between the feature coordinates at the current moment and the stable state coordinates, and recording the projection distance change trajectory for multiple consecutive sampling periods, so as to mark the crossing point and crossing direction of the trajectory crossing the dual-threshold feature boundary through the projection distance change trajectory; When the trajectory is detected to cross the same feature boundary three times in a row, the timestamp of the last crossing is recorded as the mutation moment, and a switching trigger signal containing the mutation moment, crossing point, and crossing direction is generated.

7. A system for determining coefficients of a smooth switching state transfer matrix, characterized in that: include: an acquisition module, configured to acquire flight attitude change data through an inertial measurement unit, perform spatiotemporal coupling processing on gyroscope angular velocity data in the flight attitude change data and satellite positioning data, and generate a feature vector reflecting the attitude change trend; An output module, configured to collect three-dimensional angular motion data using an inertial navigation unit array, perform non-uniform weighting on high-frequency components in the three-dimensional angular motion data, and output a steady-state estimate via a dynamic weighted average processor. The three-dimensional angular motion data is a vector set representing the rotational dynamic characteristics of the aircraft, including instantaneous angular velocity components in three orthogonal directions: pitch, roll, and yaw, for calculating attitude change trends and maneuver intensity. a generation module configured to construct a dual-threshold characteristic boundary based on the degree of deviation of the characteristic vector from the stable state estimate, and to determine a maneuver mutation timing and generate a switching trigger signal when a continuous state transition trajectory crosses the dual-threshold characteristic boundary, wherein the continuous state transition trajectory represents a data sequence representing a temporal variation law of the motion state, and includes a line connecting the projected distances between characteristic coordinates and stable coordinates at multiple consecutive sampling moments, and is used to track the continuous evolution of the maneuver trend; an execution module, configured to execute a smooth switching operation of the state transfer matrix coefficients in response to the switching trigger signal; The process of performing a smooth switching operation of the state transfer matrix coefficients includes: Obtain the old coefficient set of the current state transfer matrix and the new coefficient set of the target state transfer matrix; Taking the switching time point as a reference point, based on the old coefficient set and the current state vector, a boundary condition vector that satisfies the continuity of the output acceleration command and its first-order derivative is calculated; Constructing a state transition relationship equation using the new coefficient set, taking the boundary condition vector as the right-hand side term of the equation, and solving the state transition relationship equation to obtain a state vector after switching; The switched state vector and the new coefficient set are synchronously loaded into the motion simulation filter to complete the smooth switching of the state transfer matrix coefficients.

Images