Aircraft attitude control method and system based on semiconductor microcomputer system

Abstract

The invention discloses an aircraft attitude control method and device based on a semiconductor microcomputer system, and the method comprises the steps: collecting the multi-source sensor data of an aircraft, calculating an attitude error vector, and constructing a three-dimensional error space coordinate system; taking the three-dimensional error space coordinate system as a reference, adjusting the model in real time, and solving the optimal attitude adjustment amount in a prediction window period; the optimal attitude adjustment amount is input into an execution mechanism distribution module, and a thrust adjustment instruction sequence and a control surface deflection angle instruction queue of each vector propeller are generated through calculation in combination with a propeller response matrix and a control surface efficiency coefficient; dynamically correcting and predicting weight parameters of the controller through closed-loop verification; and updating the multi-objective optimization model based on the corrected parameters, adjusting the instruction distribution proportion through a dynamic reconfiguration interface, and writing the updated parameters into the attitude database to form a complete closed-loop period. According to the invention, high-precision real-time control of the attitude of the aircraft is realized, and the flight stability and the energy efficiency ratio in a complex environment are improved.

Description

Technical Field

[0001] The present invention relates to the technical fields of aerospace technology, automatic control technology, and semiconductor microcomputer system applications, and particularly relates to a method for controlling the attitude of an aircraft based on a semiconductor microcomputer system. Background Art

[0002] In the modern aerospace field, the attitude control of an aircraft plays a crucial role in ensuring flight safety, improving flight performance, and completing various tasks. Traditional aircraft attitude control systems mainly rely on mechanical sensors and independent control systems, and have many limitations. The data acquisition accuracy and real-time performance of traditional sensors are difficult to meet the requirements of high-precision attitude control, resulting in large attitude errors and affecting the stability and controllability of the aircraft; the control algorithms of traditional attitude control systems are relatively simple and difficult to cope with complex flight environments and changing mission requirements, and cannot achieve refined attitude adjustment; the system integration level is low, and the coordination between components is poor, resulting in low efficiency and reliability of the entire system. With the continuous development of aerospace technology, the performance requirements for aircraft attitude control systems are getting higher and higher, and a more accurate, efficient, and intelligent control method and system are needed to overcome the deficiencies of the existing technology. Summary of the Invention

[0003] A method and system for controlling the attitude of an aircraft based on a semiconductor microcomputer system, comprising:

[0004] S1. The semiconductor microcomputer system collects multi-source sensor data of the aircraft in real time, calculates the attitude error vector through a multi-modal data fusion algorithm, and constructs a three-dimensional error space coordinate system;

[0005] S2. Based on the three-dimensional error space coordinate system, a real-time attitude adjustment model is designed by using a non-linear predictive controller, and the optimal attitude adjustment amount is solved within the prediction window period through a rolling time domain optimization method;

[0006] S3. The optimal attitude adjustment amount is input into the actuator allocation module, and combined with the current thruster response matrix and rudder surface effectiveness coefficient of the aircraft, a thrust adjustment instruction sequence for each vector thruster and a rudder surface deflection angle instruction queue are generated through calculation;

[0007] S4. During the execution of the instructions, the closed-loop verification module calculates the attitude control residual index and energy consumption efficiency coefficient in real time, and dynamically corrects the rolling optimization weight parameters of the predictive controller;

[0008] S5. The multi-objective optimization model is updated according to the corrected weight parameters, the control instruction allocation ratio of the actuator is adjusted through a dynamic reconfiguration interface, and at the same time, the updated control parameters are written into the aircraft attitude database to form an adaptive closed-loop control cycle.

[0009] A method for controlling the attitude of an aircraft based on a semiconductor microcomputer system as described above, wherein the semiconductor microcomputer system collects multi-source sensor data of the aircraft in real time, calculates the attitude error vector through a multi-modal data fusion algorithm, and constructs a three-dimensional error space coordinate system, including the following sub-steps:

[0010] Perform noise suppression and time-domain synchronization processing on the multi-source sensor data, and extract the error components of the attitude angular velocity, acceleration, and angular displacement;

[0011] Map the error components to a unified error space through a weighted fusion strategy to generate an attitude error vector;

[0012] According to the dynamic change trend of the error vector, construct a three-dimensional error space coordinate system with the error amplitude, direction, and change rate as the coordinate axes.

[0013] A method for controlling the attitude of an aircraft based on a semiconductor microcomputer system as described above, wherein, based on the three-dimensional error space coordinate system, a non-linear predictive controller is used to design a real-time attitude adjustment model, and the optimal attitude adjustment amount is solved within the prediction window period through a rolling time domain optimization method, including the following sub-steps:

[0014] Establish an attitude dynamics state equation based on the three-dimensional error space coordinate system;

[0015] Discretely sample the attitude adjustment amount within the prediction window period, and evaluate the error convergence and energy consumption cost of each sampling point through a cost function;

[0016] Use the gradient descent method to iteratively solve the minimum value of the cost function to determine the optimal attitude adjustment amount.

[0017] A method for controlling the attitude of an aircraft based on a semiconductor microcomputer system as described above, wherein the optimal attitude adjustment amount is input into the actuator allocation module, combined with the current thruster response matrix and rudder surface effectiveness coefficient of the aircraft, and a thrust adjustment instruction sequence for each vector thruster and a rudder surface deflection angle instruction queue are generated through calculation, including the following sub-steps:

[0018] Calculate the thrust allocation priority of each vector thruster according to the thruster response matrix to generate a thrust adjustment instruction sequence;

[0019] Perform saturation limit verification on the rudder surface deflection angle based on the rudder surface effectiveness coefficient to generate a rudder surface deflection angle instruction queue;

[0020] Send the thrust adjustment instruction and the rudder surface deflection instruction to the corresponding actuator synchronously according to the time stamp.

[0021] A flight vehicle attitude control method based on a semiconductor microcomputer system as described above, wherein, during the execution of instructions, the closed-loop verification module calculates the attitude control residual index and the energy consumption efficiency coefficient in real time, and dynamically corrects the rolling optimization weight parameters of the predictive controller, including the following sub-steps:

[0022] Judge the current control accuracy according to the attitude control residual index, and adjust the weight coefficient of the attitude tracking error;

[0023] Evaluate the energy consumption status of the actuator according to the energy consumption efficiency coefficient, and optimize the weight distribution of the energy consumption item;

[0024] Feed back the adjusted weight coefficient to the non-linear predictive controller in real time.

[0025] A flight vehicle attitude control method based on a semiconductor microcomputer system as described above, wherein, judging the current control accuracy according to the attitude control residual index and adjusting the weight coefficient of the attitude tracking error includes the following sub-steps:

[0026] When the attitude control residual index exceeds the preset threshold, increase the weight of the attitude tracking error term;

[0027] When the residual index is lower than the threshold, gradually restore the default weight distribution.

[0028] A flight vehicle attitude control method based on a semiconductor microcomputer system as described above, wherein, updating the multi-objective optimization model according to the corrected weight parameters, adjusting the control instruction allocation ratio of the actuator through the dynamic reconfiguration interface, and at the same time writing the updated control parameters into the flight vehicle attitude database to form an adaptive closed-loop control cycle, including the following sub-steps:

[0029] According to the updated multi-objective optimization model, re-allocate the priorities of the thrust adjustment instruction and the rudder surface deflection instruction;

[0030] When the thruster fails or the rudder surface is saturated, activate the instruction compensation logic of the redundant actuator;

[0031] Feed back the adjusted instruction allocation ratio to the non-linear predictive controller to update the optimization model parameters of the next control cycle.

[0032] A flight vehicle attitude control system based on a semiconductor microcomputer system, wherein, it includes:

[0033] 8. A flight vehicle attitude control system based on a semiconductor microcomputer system, characterized in that it includes:

[0034] Data acquisition and processing module: The semiconductor microcomputer system collects multi-source sensor data of the flight vehicle in real time, calculates the attitude error vector through a multi-modal data fusion algorithm, and constructs a three-dimensional error space coordinate system;

[0035] Predictive control algorithm module: Based on the three-dimensional error space coordinate system, a nonlinear predictive controller is used to design a real-time attitude adjustment model, and the optimal attitude adjustment amount is solved within the prediction window period through the rolling horizon optimization method;

[0036] Actuator allocation module: Input the optimal attitude adjustment amount into the actuator allocation module. Combining the current thruster response matrix and the rudder surface effectiveness coefficient of the aircraft, a thrust adjustment instruction sequence for each vector thruster and a rudder surface deflection angle instruction queue are generated through calculation;

[0037] Closed-loop verification and dynamic adjustment module: During the execution of the instructions, the closed-loop verification module calculates the attitude control residual index and the energy consumption efficiency coefficient in real time, and dynamically corrects the rolling optimization weight parameters of the predictive controller;

[0038] Dynamic configuration management module: Update the multi-objective optimization model according to the corrected weight parameters, adjust the control instruction allocation ratio of the actuator through the dynamic reconfiguration interface, and at the same time write the updated control parameters into the aircraft attitude database to form an adaptive closed-loop control cycle.

[0039] A computer storage medium, characterized in that it includes: at least one memory and at least one processor;

[0040] The memory is used to store one or more program instructions;

[0041] The processor is used to run one or more program instructions to execute an aircraft attitude control method based on a semiconductor microcomputer system as described in any one of the above.

[0042] The beneficial effects achieved by the present invention are as follows:

[0043] Through the deep collaboration of multimodal data fusion and nonlinear predictive control, the present invention significantly improves the comprehensive effectiveness of aircraft attitude control. Based on the efficient architecture of the semiconductor microcomputer system, precise synchronization and fusion processing of multi-source sensor data are achieved, greatly optimizing the calculation accuracy of the attitude error vector and effectively overcoming the error accumulation problem caused by data heterogeneity in traditional methods. By introducing rolling horizon optimization and nonlinear dynamic model predictive technology, a highly robust attitude adjustment strategy is quickly generated in complex dynamic environments, significantly enhancing the response ability and stability of the aircraft to sudden disturbances. The dynamic allocation mechanism of the actuator, combined with closed-loop energy consumption feedback, realizes the refined matching of thruster and rudder control commands, significantly reducing energy consumption while ensuring control accuracy. The adaptive parameter correction module continuously optimizes the control weight allocation strategy through real-time residual analysis and linkage with historical data, ensuring the stable operation of the system during long-term tasks. This technical solution provides reliable support for the autonomous control of aircraft in extreme environments, combining the comprehensive advantages of control accuracy, real-time performance, and energy efficiency optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0045] Figure 1 It is a flowchart of an aircraft attitude control method based on a semiconductor microcomputer system provided by an embodiment of the present application.

[0046] Figure 2 It is a schematic diagram of an aircraft attitude control system based on a semiconductor microcomputer system provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0048] Embodiment 1

[0049] As Figure 1 shown, an aircraft attitude control method based on a semiconductor microcomputer system in an embodiment of the present application includes:

[0050] Step S1, the semiconductor microcomputer system collects multi-source sensor data of the aircraft in real time, calculates the attitude error vector through a multimodal data fusion algorithm, and constructs a three-dimensional error space coordinate system, which specifically includes the following sub-steps:

[0051] Step S11, performing noise suppression and time domain synchronization processing on multi-source sensor data to extract error components of attitude angular velocity, acceleration and angular displacement;

[0052] The multi-source sensor data is subjected to noise suppression, and a low-pass filtering algorithm is used to weaken the high-frequency noise interference; then, through time domain synchronization processing, the sensor data with mismatched timestamps is expressed by the following formula:

[0053]

[0054] Among them, y(t n ) represents the data of sensor s in a unified time series {t n} n The value after synchronization at the moment; M represents the number of sensor data points involved in synchronization; β m represents the weight coefficient, which is used to measure the contribution weight of the mth sensor data point in the synchronization calculation; y s (m) represents the value of the mth raw data point of sensor s; φ represents the cubic spline interpolation basis function, which is used to process the interpolation calculation of the time difference; t s,m represents the original time corresponding to the mth data point of sensor s; Δt represents the time interval parameter; λ represents the weight coefficient, which is used to balance the optimization weight of the cubic spline interpolation term and the dynamic time warping term; y s (i) represents the value of the i-th data point in the raw data sequence of sensor s; Represents the value of the jth data point in the reference data of sensor s.

[0055] According to the time difference and value difference of the data at adjacent moments, the value of the intermediate moment is estimated proportionally to eliminate the timing deviation. The error components of the attitude angular velocity, acceleration and angular displacement are extracted respectively. The error component is the difference between the actual measured value of each sensor and the expected reference value of the current attitude of the aircraft.

[0056] Step S12: Mapping the error components to a unified error space through a weighted fusion strategy to generate a posture error vector;

[0057] After normalizing each error component, the fusion weights are dynamically allocated based on the sensor confidence: the weight of the angular velocity error increases with the increase of the gyroscope signal-to-noise ratio, the weight of the acceleration error increases during the steady-state flight phase, and the weight of the angular displacement error is adjusted according to the magnetometer calibration status. The Euler angle-quaternion hybrid coordinate system is used as the unified error space, and each component error is converted to this space through orthogonal projection. Finally, the comprehensive attitude error vector is generated by weighted superposition according to the weight coefficients:

[0058]

[0059] where \(E\) represents the comprehensive attitude error vector; \(S\) represents the number of sensors participating in the fusion; are the weight coefficients corresponding to the angular velocity error, acceleration error, and angle projection error of the \(s\)-th sensor, respectively, which are used to adjust the contribution ratio of different error terms; are the angular velocity error vector and acceleration error vector of the \(s\)-th sensor, respectively; represents the angle projection error vector of the \(s\)-th sensor; \(\lambda\) represents the weight coefficient; \(\sum\) s represents the covariance matrix of the \(s\)-th sensor; \(W\) s represents the weight matrix of the \(s\)-th sensor.

[0060] Step S13: According to the dynamic change trend of the error vector, construct a three-dimensional error space coordinate system with the error amplitude, direction, and change rate as the coordinate axes;

[0061] Analyze the statistical characteristics and dynamic evolution laws of the error vector within a continuous time window. The error amplitude is characterized by the vector norm, the direction is parameterized by the polar angle of the unit vector, and the change rate is calculated by the sliding window difference method to obtain the instantaneous derivative. When constructing the coordinate system, the amplitude is mapped to the radial axis by polar coordinate transformation, the direction is decomposed into the pitch-yaw biaxial, and the change rate generates a continuous gradient field through cubic spline interpolation to form a three-dimensional error space with a dynamic orthogonal basis.

[0062] Step S2: Based on the three-dimensional error space coordinate system, design a real-time attitude adjustment model using a non-linear predictive controller, and solve for the optimal attitude adjustment amount within the prediction window period through the rolling time domain optimization method, which specifically includes the following sub-steps:

[0063] Step S21: Establish an attitude dynamics state equation based on the three-dimensional error space coordinate system;

[0064] Based on the three-dimensional error space coordinate system, taking the error amplitude, direction, and rate of change as the core state variables, combining the rigid body dynamics characteristics and nonlinear coupling effects of the aircraft, an attitude dynamics state equation is established. By describing the law of the evolution of state variables over time, the dynamic changes of errors are associated with control inputs, forming a continuous nonlinear system model, and clarifying the functional relationship between the derivative of state variables and control quantities, providing a mathematical framework for subsequent predictive control.

[0065] Step S22: Discretely sample the attitude adjustment amount within the prediction window period, and evaluate the error convergence and energy consumption cost of each sampling point through a cost function.

[0066] Within the prediction window, the continuous time is discretized into a time series with a fixed interval, and the attitude adjustment amount at each discrete moment is sampled to generate a candidate control sequence. For each sampling point, a comprehensive cost function is designed to evaluate its performance:

[0067]

[0068] where, J represents the comprehensive cost function; N represents the length of the prediction window, and n = 0, 1,..., N - 1 represents the discrete moments within the prediction window; represents the square of the error amplitude at the nth moment, quantifying the magnitude of the attitude error; ω φ represents the error direction term is the weight coefficient; represents the square of the attitude error at the nth moment in a certain direction; λ represents the weight coefficient of the energy penalty term, dynamically balancing the priority of error convergence and energy consumption; represents the weighted sum of squares of the control quantity, where R is the weight matrix, used to constrain the magnitude of the control quantity to avoid actuator overload; u n represents the control input vector at the nth moment; ρ represents the weight coefficient of the terminal cost term, strengthening the constraint on the final attitude error convergence effect; X N represents the system state vector at the end of the prediction window; X ref represents the desired reference state vector, representing the ideal attitude error convergence target;

[0069] The error convergence is quantified by the cumulative squared value of the error amplitude, reflecting the control effect; the energy consumption is characterized by the sum of squares of the control quantity to avoid actuator overload. The two are dynamically balanced through preset weight coefficients, forming an optimization goal that takes into account both accuracy and energy consumption. During the rolling optimization process, the state trajectory within the prediction window is updated in real time and the cost is re-evaluated to ensure that the control strategy adapts to dynamic error changes.

[0070] Step S23: Use the gradient descent method to iteratively solve the minimum value of the cost function to determine the optimal attitude adjustment amount.

[0071] The gradient descent method is used to iteratively search for the minimum value of the cost function. After initially setting the control sequence, the error evolution path is predicted by forward integrating the state equation, and the gradient information of the cost function with respect to the control quantity is calculated backward. In each iteration, the control quantity is adjusted along the gradient descent direction according to the learning rate, gradually approaching the optimal solution. The iteration termination condition is set to the gradient magnitude being lower than the threshold or reaching the maximum number of iterations. Finally, the first control quantity in the prediction window is extracted from the optimization result as the real-time optimal adjustment instruction to drive the actuator to achieve rapid attitude stabilization while balancing the system energy consumption and response speed.

[0072] Step S3: Input the optimal attitude adjustment amount into the actuator allocation module. Combining with the current thruster response matrix and the rudder surface effectiveness coefficient of the aircraft, generate the thrust adjustment instruction sequence for each vector thruster and the rudder surface deflection angle instruction queue through calculation, specifically including the following sub-steps:

[0073] Step S31: Calculate the thrust allocation priority for each vector thruster according to the thruster response matrix and generate the thrust adjustment instruction sequence;

[0074] Based on the thruster layout and thrust direction characteristics of the aircraft, construct the thruster response matrix. The matrix elements reflect the contribution of the unit thrust of each thruster to the attitude adjustment amount in the three-dimensional error space coordinate system. By solving the matrix pseudo-inverse or constrained optimization problem, decompose the optimal attitude adjustment amount into the thrust instructions for each vector thruster: calculate the weight coefficients for thrust allocation to ensure that the total thrust vector sum is consistent with the target adjustment amount while minimizing the total energy consumption or thrust change amplitude of the thrusters; for redundant propulsion systems, introduce a priority strategy to preferentially call thrusters with high response efficiency and sufficient remaining fuel to generate a smooth thrust adjustment instruction sequence that meets the mechanical response delay constraints in terms of timing.

[0075] Step S32: Perform saturation limit verification on the rudder surface deflection angle based on the rudder surface effectiveness coefficient and generate the rudder surface deflection angle instruction queue;

[0076] According to the aerodynamic effectiveness coefficient of the rudder surface and the current flight state, dynamically calculate the correction ability of the deflection angle of each rudder surface to the error space direction axis. Ensure that the rudder surface instructions do not exceed the mechanical limit and the aerodynamic stability boundary through saturation limit verification: based on the non-linear curve of rudder surface deflection angle - effectiveness, back-calculate the minimum deflection angle that meets the target correction amount and superimpose the safety margin threshold, where the target direction correction amount is Δφ cmd The corresponding rudder surface instruction solution is:

[0077]

[0078] where δ represents the optimal deflection instruction of the rudder surface obtained by optimization solution; It means to find the solution that minimizes the subsequent expression among the values of δ; j = 1 represents the index range of summation, where j represents the number of the control surface, indicating that the calculation object is from the 1st to the nth control surface; ΔM j represents the change in moment generated when the jth control surface deflects; Δφ cmd represents the target direction correction amount; e φ represents the unit vector of the target correction direction; λ represents the weight coefficient, which is used to dynamically balance the priorities of the "deviation constraint term" and the "energy constraint term"; represents the sum of the squares of the deflection amounts of the first n control surfaces.

[0079] If the calculated value exceeds the physical limit, the deflection angle is scaled proportionally to the maximum allowable value, and at the same time, a thrust compensation strategy is triggered to maintain the overall adjustment amount, and finally a control surface instruction queue that takes into account both response speed and stability is generated.

[0080] Step S33: Synchronously send the thrust adjustment instruction and the control surface deflection instruction to the corresponding actuators according to the time stamp;

[0081] Using the aircraft on-board clock synchronization mechanism, a unified time stamp is added to the thrust adjustment instruction sequence and the control surface deflection instruction queue to ensure that the actions of multiple actuators are strictly aligned in time sequence. The instructions are distributed to the corresponding thrusters and control surface controllers through the real-time bus, and conflict detection is performed before sending: checking the physical interference risk of each actuator within the overlapping time interval of the instruction time sequence, and inserting a small time offset or adjusting the instruction amplitude if necessary, and finally realizing the coordinated control of multiple actuators with high precision and low disturbance.

[0082] Step S4: During the execution of the instruction, the closed-loop verification module calculates the attitude control residual index and the energy consumption efficiency coefficient in real time, and dynamically corrects the rolling optimization weight parameters of the predictive controller, which specifically includes the following sub-steps:

[0083] Step S41: Judge the current control accuracy according to the attitude control residual index, and adjust the weight coefficient of the attitude tracking error;

[0084] The closed-loop verification module calculates the attitude control residual index, the moving window mean of the square of the residual, by collecting the difference between the actual attitude parameters and the desired attitude parameters of the aircraft in real time: where R is the moving window mean of the square of the residual; N is the length of the moving window, that is, the number of samples included in the window; k is the time index variable, which is used to traverse each moment within the moving window, and the value range is from t - N + 1 to t; e(k) is the residual vector of the system at moment k. The formula is used to quantify the current control accuracy.

[0085] (1). When the attitude control residual index exceeds the preset threshold, increase the weight of the attitude tracking error term;

[0086] When the residual index exceeds the dynamic threshold, gradually increase the weight coefficient of the attitude tracking error in the cost function based on the proportional-integral regulation logic, and strengthen the priority of the controller for error convergence;

[0087] (2) When the residual index is lower than the threshold, gradually restore the default weight allocation;

[0088] If the residual continuously remains lower than the threshold, reduce the weight according to the exponential decay law to avoid over-control. At the same time, introduce historical residual trend analysis to prevent frequent weight oscillations and ensure a smooth and stable adjustment process.

[0089] Step S42: Evaluate the energy consumption status of the actuator according to the energy consumption efficiency coefficient, and optimize the weight allocation of the energy consumption item;

[0090] The energy consumption efficiency coefficient is calculated by the ratio of the real-time energy consumption data of the actuator to the error convergence rate, reflecting the improvement amplitude of the control effect corresponding to unit energy consumption. When the efficiency coefficient decreases, adopt a piecewise linear interpolation strategy to reduce the weight coefficient of the energy consumption item, allowing a temporary increase in energy consumption to maintain control accuracy; if the efficiency coefficient is higher than the reference value, limit the minimum value of the energy consumption weight according to the saturation function to prevent energy waste. During the weight adjustment process, combine the remaining energy reserve of the aircraft and the requirements of the mission phase, and dynamically set the adaptive correction amplitude of the efficiency coefficient to achieve task context awareness of the energy consumption strategy.

[0091] Step S43: Feed back the adjusted weight coefficient to the non-linear predictive controller in real time;

[0092] The adjusted weight coefficient is injected into the rolling optimization module of the non-linear predictive controller through a low-latency communication link in real time to update the balance relationship between the error convergence term and the energy consumption term in the cost function. To ensure the stability of the controller, the weight needs to be processed by a first-order low-pass filter before updating to eliminate high-frequency mutation interference. At the same time, adopt a double-buffer mechanism: use the old weight to perform the optimization calculation in the current cycle, and switch to the new weight in the next cycle to avoid jumps in the control quantity caused by parameter switching. The updated weight will continuously affect the optimization target within the prediction window, forming a closed-loop adaptive control link of "perception-evaluation-correction".

[0093] Step S5: Update the multi-objective optimization model according to the corrected weight parameters, adjust the control instruction allocation ratio of the actuator through the dynamic reconfiguration interface, and write the updated control parameters into the aircraft attitude database at the same time to form an adaptive closed-loop control cycle, which specifically includes the following sub-steps:

[0094] Step S51: Re-allocate the priorities of the thrust adjustment instruction and the rudder surface deflection instruction according to the updated multi-objective optimization model;

[0095] Based on the updated multi-objective optimization model, the ratio of the attitude tracking error weight to the energy consumption weight is used as the basis for dynamic allocation, and the priority ratio of the thrust adjustment command and the rudder deflection command is recalculated. The thrust allocation ratio is determined by dividing the thrust weight by the sum of the thrust weight and the rudder weight, ensuring that the thrusters are preferentially called to achieve rapid correction under high error weights, while increasing the contribution of the rudder surface to save fuel under high energy consumption weights. At the same time, a non-linear interpolation function is introduced:

[0096] To avoid ratio jumps, a hyperbolic tangent smooth transition based on the weight change rate is introduced:

[0097]

[0098] where γ smooth (t) represents the smoothed weight change rate; γ(t) represents the original weight change rate; represents the derivative of the weight change rate γ(t); Δt represents the time interval, representing the time difference between two adjacent moments in the discretized calculation; σ represents the smoothing parameter, controlling the change gradient of the hyperbolic tangent function tanh. The smaller σ is, the steeper the transition process; the larger σ is, the smoother the transition and the more significant the smoothing effect; γ(t) - γ(t - Δt) represents the difference in the weight change rate between the current moment t and the previous moment t - Δt, reflecting the change amount of γ(t) within the time interval Δt, and is used to adjust the amplitude of the smooth transition.

[0099] According to the smooth transition of the weight change rate to allocate the ratio, avoid command jumps, and finally generate an instruction coordination strategy that matches the current control target.

[0100] Step S52: When a thruster fails or the rudder surface saturates, activate the command compensation logic of the redundant actuator;

[0101] When the deviation between the actual thrust of the thruster and the command exceeds the preset fault threshold, or the rudder deflection angle reaches the mechanical limit, trigger the redundant actuator compensation logic. Re-solve the thrust allocation of the remaining available actuators through the generalized inverse of the thruster response matrix to minimize the total adjustment amount and the residual of the original target; when the rudder surface saturates, dynamically scale the deflection angle of the unsaturated rudder surface according to the aerodynamic efficiency coefficient, and superimpose the thrust compensation term to fill the remaining error. During the compensation process, limit the overshoot amplitude of a single actuator to prevent cascading overloads and ensure that the system still maintains stable control under fault or saturation conditions.

[0102] Step S53: Feed back the adjusted command allocation ratio to the non-linear predictive controller to update the optimization model parameters for the next control cycle;

[0103] The adjusted instruction allocation ratio is fed back to the non-linear predictive controller in real time through the dynamic reconfiguration interface to update the weight parameters of its internal optimization model and the actuator constraint conditions. Before the update, first-order lag filtering is performed on the parameters to eliminate high-frequency disturbances, and the double-buffer mechanism is used to ensure that the parameter switching is synchronized with the controller iteration period to avoid instantaneous conflicts. The updated parameters are simultaneously written into the historical log of the aircraft attitude database for long-term trend analysis and offline policy optimization, forming an adaptive closed-loop across control cycles to continuously improve the robustness of the system under complex working conditions.

[0104] Embodiment 2

[0105] As Figure 2 shown, Embodiment 2 of the present application provides an aircraft attitude control system based on a semiconductor microcomputer system, including:

[0106] Data acquisition and processing module 21: The semiconductor microcomputer system collects multi-source sensor data of the aircraft in real time, calculates the attitude error vector through a multi-modal data fusion algorithm, and constructs a three-dimensional error space coordinate system, including the following sub-modules:

[0107] Multi-source data preprocessing sub-module 211: Perform noise suppression and time-domain synchronization processing on the multi-source sensor data, and extract the error components of the attitude angular velocity, acceleration, and angular displacement;

[0108] First, perform noise suppression processing on the original multi-source sensor data. Using a low-pass filter algorithm, calculate the smoothed output value after filtering by recursively weighting and averaging the data at the current moment and the previous moment, effectively weakening the high-frequency noise interference; subsequently, perform time-domain synchronization on the sensor data with mismatched timestamps. Using the linear interpolation method, calculate the synchronized value at the missing intermediate time point according to the proportional relationship between the numerical difference and the time difference of the data at adjacent moments, eliminating the timing deviation of the multi-source data, which is represented by the following formula:

[0109]

[0110] where, y(t n ) represents the synchronized value of the data of sensor s at time t n in the unified time series {t n}; M represents the number of sensor data points participating in the synchronization; β m represents the weight coefficient, which is used to measure the contribution weight of the m-th sensor data point in the synchronization calculation; y s (m) represents the value of the m-th original data point of sensor s; φ represents the cubic spline interpolation basis function, which is used to process the interpolation calculation of the time difference; t s,mrepresents the original time corresponding to the m-th data point of sensor s; Δt represents the time interval parameter; λ represents the weight coefficient used to balance the optimization weights of the cubic spline interpolation term and the dynamic time warping term; y s (i) represents the value of the i-th data point in the original data sequence of sensor s; represents the value of the j-th data point in the reference data of sensor s.

[0111] Finally, the error components of the attitude angular velocity, acceleration, and angular displacement are calculated respectively. Specifically, they are the differences between the real-time measurement values of each sensor and the expected reference values corresponding to the current attitude control target of the aircraft. The difference for the angular velocity error is obtained by subtracting the reference angular velocity from the angular velocity measured by the sensor. The acceleration and angular displacement errors are calculated using the same logic, forming a basic data set of error components.

[0112] Data fusion calculation sub-module 212: Map the error components to a unified error space through a weighted fusion strategy to generate an attitude error vector;

[0113] Design a weighted fusion strategy based on the noise characteristics of each sensor. By calculating the weight values of each error component and normalizing all the weights so that the sum of the total weights is 1. Subsequently, multiply the normalized weights by the corresponding error components respectively to calculate the weighted standardized error values, eliminating the influence of the dimensional differences of different sensors. Finally, combine the weighted angular velocity error, acceleration error, and angular displacement error according to the three-dimensional space dimensions, and weightedly superimpose them according to the weight coefficient to generate a comprehensive attitude error vector:

[0114]

[0115] Among them, E represents the comprehensive attitude error vector; S represents the number of sensors participating in the fusion; are the weight coefficients corresponding to the angular velocity error, acceleration error, and angle projection error of the s-th sensor respectively, used to adjust the contribution ratio of different error terms; are the angular velocity error vector and acceleration error vector of the s-th sensor respectively; represents the angle projection error vector of the s-th sensor; λ represents the weight coefficient; ∑ s represents the covariance matrix of the s-th sensor; W s represents the weight matrix of the s-th sensor.

[0116] Realize the deep fusion and spatial alignment of multi-source heterogeneous data.

[0117] Dynamic space modeling sub-module 213: Construct a three-dimensional error space coordinate system with the error amplitude, direction, and change rate as the coordinate axes according to the dynamic change trend of the error vector;

[0118] According to the attitude error vector updated in real time, a three-dimensional error space coordinate system is dynamically constructed: the first axis is the error amplitude, which characterizes the overall error intensity by calculating the modulus of the error vector in the three-dimensional space; the second axis is the error direction, which identifies the dominant direction of error distribution by calculating the projection azimuth angle of the error vector on the horizontal plane; the third axis is the error change rate, which quantifies the instantaneous dynamic evolution trend of the error by calculating the ratio of the difference value of the error amplitude at adjacent moments to the time interval.

[0119] Predictive control algorithm module 22: Based on the three-dimensional error space coordinate system, a real-time attitude adjustment model is designed using a non-linear predictive controller, and the optimal attitude adjustment amount is solved within the prediction time window through the rolling time domain optimization method, including the following sub-modules:

[0120] System modeling sub-module 221: Establish an attitude dynamics state equation based on the three-dimensional error space coordinate system;

[0121] Based on the three state variables of the error amplitude, direction angle and change rate in the three-dimensional error space coordinate system, combined with the rigid body dynamics parameters of the aircraft and the external environmental torque, a non-linear attitude dynamics state equation is established. By modeling the coupling effect between the error direction angle and the aircraft angular velocity, and the energy transfer relationship between the error amplitude change rate and the thruster torque, a functional relationship between the derivative of the state variable and the control input is constructed, forming a state evolution equation in the continuous domain to describe the dynamic process of attitude adjustment in the error space, providing a mathematical model basis for predictive control.

[0122] Prediction evaluation sub-module 222: Discretely sample the attitude adjustment amount within the prediction time window, and evaluate the error convergence and energy consumption cost of each sampling point through a cost function;

[0123] Within the set prediction time window, the continuous time is discretized into a time sequence grid with a fixed step size, and the attitude adjustment amount at each discrete time point is sampled to generate a candidate control sequence. The performance of each control sequence is quantified through a cost function:

[0124]

[0125] where, J represents the comprehensive cost function; N represents the length of the prediction window, and n = 0, 1,..., N - 1 represents the discrete moments within the prediction window; represents the square of the error amplitude at the nth moment, quantifying the magnitude of the attitude error; ω φ represents the error direction term is the weight coefficient of; represents the square of the attitude error in a certain direction at the nth moment; λ represents the weight coefficient of the energy penalty term, dynamically balancing the priority of error convergence and energy consumption; represents the weighted sum of squares of the control quantities, where \(R\) is the weight matrix used to constrain the magnitudes of the control quantities and avoid actuator overload; \(u\) n represents the control input vector at the \(n\)-th moment; \(\rho\) represents the weight coefficient of the terminal cost term, strengthening the constraint on the convergence effect of the final attitude error; \(X\) N represents the system state vector at the end of the prediction window; \(X\) ref represents the desired reference state vector, representing the ideal attitude error convergence target.

[0126] The error convergence is calculated by accumulating the squared values of the error magnitudes at each moment within the prediction window, reflecting the control effect; the energy consumption is calculated by accumulating the sum of squares of the control quantities at each moment, characterizing the system energy consumption. The two are linearly weighted by a preset weight coefficient to form a comprehensive cost index, achieving the balance optimization goal of accuracy and energy consumption, and the prediction trajectory is corrected in real time by combining the latest sensor data during rolling update.

[0127] Optimization and solution sub-module 223: Use the gradient descent method to iteratively solve the minimum value of the cost function and determine the optimal attitude adjustment amount;

[0128] Use the gradient descent method to iteratively search for the minimum value of the cost function. After initially setting the control sequence, predict the evolution path of the error over time through the forward integration state equation, and calculate the gradient of the cost function with respect to the control quantity in reverse. In each iteration, adjust the control quantity along the opposite direction of the gradient by the learning rate step size, gradually approaching the optimal solution. The iteration termination condition is set to the gradient magnitude being lower than the convergence threshold or reaching the maximum number of iterations. Finally, extract the control quantity at the first moment from the optimized control sequence as the real-time optimal adjustment instruction output to ensure that the predictive control has both fast response and global optimization characteristics.

[0129] Actuator allocation module 23: Input the optimal attitude adjustment amount into the actuator allocation module. Combine the current thruster response matrix and the rudder surface effectiveness coefficient of the aircraft, and generate a thrust adjustment instruction sequence for each vector thruster and a rudder surface deflection angle instruction queue through calculation, including the following sub-modules:

[0130] Thrust allocation strategy sub-module 231: Calculate the thrust allocation priorities of each vector thruster according to the thruster response matrix and generate a thrust adjustment instruction sequence;

[0131] The response matrix constructed based on the layout of the aircraft thrusters and the characteristics of the thrust direction decomposes the optimal attitude adjustment amount into specific thrust commands for each vector thruster. By calculating the pseudo-inverse of the response matrix or using the constrained least squares optimization method, a thrust allocation scheme that meets the total adjustment amount requirement and minimizes the total energy consumption is solved, where the weight coefficient of each thruster is dynamically adjusted by its response efficiency and the remaining fuel ratio. For redundant propulsion systems, thrusters with high response efficiency and good health status are preferentially selected to generate high-priority commands. At the same time, the amplitude of thrust change between adjacent moments is limited to ensure that the command sequence is smooth and conforms to the mechanical response delay characteristics of the thrusters, and finally a temporally coherent thrust adjustment command sequence is output.

[0132] Rudder surface effectiveness management sub-module 232: Based on the rudder surface effectiveness coefficient, perform saturation limit verification on the rudder surface deflection angle to generate a queue of rudder surface deflection angle commands;

[0133] Combined with the current flight speed, angle of attack, and rudder surface aerodynamic effectiveness coefficient, calculate the theoretical contribution value of the rudder surface deflection angle to the correction of the error space direction. By querying the rudder surface deflection angle - effectiveness non-linear curve, inversely deduce the minimum deflection angle required to meet the target correction amount, and superimpose the dynamic safety margin threshold to generate the initial command to inversely deduce the minimum deflection angle required to meet the target correction amount, where the target direction correction amount is Δφ cmd The corresponding rudder surface command calculation is as follows:

[0134]

[0135] Among them, δ represents the optimal deflection command of the rudder surface obtained by optimization; represents finding the solution that minimizes the subsequent expression among the values of δ; j = 1 represents the summation index range, where j represents the serial number of the rudder surface, indicating that the calculation object is from the 1st to the nth rudder surface; ΔM j represents the change in torque generated when the jth rudder surface deflects; Δφ cmd represents the target direction correction amount; e φ represents the unit vector of the target correction direction; λ represents the weight coefficient, which is used to dynamically balance the priorities of the "deviation constraint term" and the "energy constraint term"; represents the sum of the squares of the deflection amounts of the first n rudder surfaces.

[0136] If the initial command exceeds the mechanical limit or the aerodynamic stall boundary of the rudder surface, it is scaled proportionally to the maximum allowable deflection angle, and the remaining uncompensated adjustment amount is calculated and transmitted to the thrust allocation module for collaborative compensation. At the same time, monitor the real-time load status of the rudder surface actuator, and reduce the weight of its effectiveness coefficient in case of overheating or wear warning, and dynamically adjust the command generation strategy to extend the equipment life.

[0137] Instruction Synchronous Sending Sub-module 233: Send the thrust adjustment instruction and the rudder surface deflection instruction to the corresponding actuators synchronously according to the time stamp;

[0138] Use the unified clock source of the aircraft to add accurate time stamps to the thrust instruction sequence and the rudder surface instruction queue, align the instruction timings of different actuators through the interpolation algorithm, and eliminate the asynchronous error caused by communication delay. Before the instruction is sent, a conflict detection algorithm is used to check the spatio-temporal overlapping area of the actions of multiple actuators: for example, when the propeller plume is ejected in a specific direction, if there is a risk of physical interference with the rudder surface deflection, a small time offset is dynamically inserted or the thrust direction is adjusted according to the priority. The instruction is finally distributed to each actuator controller through a high-reliability bus, and a cyclic redundancy check code is attached to ensure the transmission integrity, realizing the high-precision cooperative control of multiple actuators.

[0139] Closed-loop Verification and Dynamic Adjustment Module 24: During the execution of the instruction, the closed-loop verification module calculates the attitude control residual index and the energy consumption efficiency coefficient in real time, and dynamically corrects the rolling optimization weight parameters of the predictive controller, including the following sub-modules:

[0140] Control Precision Adaptive Sub-module 241: Judge the current control precision according to the attitude control residual index, and adjust the weight coefficient of the attitude tracking error;

[0141] The closed-loop verification module calculates the attitude control residual index, the sliding window mean of the residual square, by collecting the difference between the actual attitude parameters and the desired attitude parameters of the aircraft in real time: where, R is the sliding window mean of the residual square; N is the length of the sliding window, that is, the number of samples included in the window; k is the time index variable, used to traverse each moment in the sliding window, and the value range is from t - N + 1 to t; e(k) is the residual vector of the system at moment k. The formula is used to quantify the current control precision.

[0142] (1), When the attitude control residual index exceeds the preset threshold, increase the weight of the attitude tracking error term;

[0143] When the residual index exceeds the dynamic threshold, gradually increase the weight coefficient of the attitude tracking error in the cost function based on the proportional-integral regulation logic, and strengthen the priority of the controller for error convergence;

[0144] (2), When the residual index is lower than the threshold, gradually restore the default weight allocation;

[0145] If the residual continues to be lower than the threshold, the weight is reduced according to the exponential decay law to avoid over-control. At the same time, the historical residual trend analysis is introduced to prevent the weight from oscillating frequently and ensure the smooth and stable adjustment process.

[0146] Energy consumption optimization decision submodule 242: evaluates the energy consumption state of the actuator according to the energy consumption efficiency coefficient, and optimizes the weight distribution of energy consumption items;

[0147] The energy consumption efficiency coefficient is calculated by the ratio of the total energy consumption of the actuator per unit time to the convergence rate of the attitude error, which characterizes the effectiveness of energy utilization. When the efficiency coefficient is lower than the dynamic benchmark, the piecewise linear interpolation method is used to reduce the energy consumption weight: the difference between the efficiency coefficient and the benchmark is mapped to the weight reduction according to the preset ratio, and the reduction increases as the difference increases, allowing temporary energy consumption to increase to maintain control accuracy; if the efficiency coefficient is higher than the benchmark, the energy consumption weight is limited to the minimum allowable value through the saturation function to avoid excessive pursuit of energy saving leading to a decrease in control performance. At the same time, the benchmark range is scaled in real time in combination with the remaining fuel of the thruster. When the fuel is insufficient, the benchmark is increased to force energy saving and extend the mission endurance.

[0148] Parameter real-time synchronization submodule 243: feeds back the adjusted weight coefficient to the nonlinear prediction controller in real time;

[0149] The adjusted weight coefficient is smoothed by a first-order low-pass filter and stored in the waiting-to-update area of ​​the double buffer. The current control cycle continues to use the original weight calculation instruction, and switches to the new weight after filtering at the beginning of the next cycle to ensure that there is no instantaneous jump in the parameter switching. During the synchronization process, the rate of change of the cost function is monitored in real time. If the rate of change exceeds the safety threshold or an oscillation trend is detected, the exception handling is triggered: the weight update is suspended and rolled back to the last stable parameter, and the fault code is recorded. The successfully updated weight parameters are associated with the timestamp and environmental status and stored in the database for offline analysis of the effectiveness of the control strategy and long-term optimization of energy consumption.

[0150] Dynamic configuration management module 25: updates the multi-objective optimization model according to the corrected weight parameters, adjusts the control instruction allocation ratio of the actuator through the dynamic reconfiguration interface, and writes the updated control parameters into the aircraft attitude database to form an adaptive closed-loop control cycle, including the following sub-modules:

[0151] Dynamic strategy reconfiguration submodule 251: reallocate the priorities of thrust adjustment instructions and rudder deflection instructions according to the updated multi-objective optimization model;

[0152] Based on the updated weight parameters of the multi-objective optimization model, the coordinated allocation ratio of the thrust adjustment command and the rudder deflection command is dynamically calculated. The priority function is constructed through the weight ratio, and the error weight ratio is mapped to the thrust command allocation intensity, while the energy consumption weight ratio corresponds to the contribution of the rudder aerodynamic correction, forming a dynamic balance relationship between thrust and rudder. A nonlinear smoothing function is used to handle the ratio switching process, and the transition speed is adaptively adjusted according to the weight change rate to avoid command jumps. The nonlinear interpolation function is expressed by the following formula:

[0153]

[0154] Among them, γ smooth (t) represents the smoothed weight change rate; γ(t) represents the original weight change rate; represents the derivative of the weight change rate γ(t); Δt represents the time interval, which represents the time difference between two adjacent moments in the discretized calculation; σ represents the smoothing parameter, which controls the change gradient of the hyperbolic tangent function tanh. The smaller σ is, the steeper the transition process is; the larger σ is, the smoother the transition is, and the more significant the smoothing effect is; γ(t) - γ(t - Δt) represents the difference in the weight change rate between the current moment t and the previous moment t - Δt, which reflects the change amount of γ(t) within the time of Δt and is used to adjust the amplitude of the smooth transition.

[0155] When the error weight is significantly improved, the high-response characteristics of the thruster are preferentially enhanced to accelerate the error convergence; when the energy consumption weight dominates, the resources are tilted to the rudder surface control to reduce fuel consumption, realizing the real-time strategy adaptation of multi-objective optimization.

[0156] Fault-tolerant execution management sub-module 252: When a thruster fails or the rudder surface is saturated, activate the instruction compensation logic of the redundant actuator;

[0157] When the deviation between the actual thrust of the thruster and the instruction exceeds the preset fault tolerance, or the deflection angle of the rudder surface reaches the mechanical limit, trigger the redundant compensation logic. For thruster faults, based on the pseudo-inverse solution of the response matrix of the remaining available thrusters, calculate the new thrust distribution to minimize the sum of the squares of the residuals between the total adjustment amount and the target value, while limiting the maximum thrust increase of a single thruster to prevent overload. When the rudder surface is saturated, scale the deflection angle of the unsaturated rudder surface according to the current aerodynamic efficiency coefficient, calculate the remaining uncompensated error amount, and convert it into an equivalent thrust instruction to be superimposed on the thruster allocation module.

[0158] Closed-loop parameter iteration sub-module 253: Feed back the adjusted instruction allocation ratio to the non-linear predictive controller to update the optimization model parameters in the next control cycle;

[0159] Feed back the dynamically adjusted thrust-rudder surface instruction allocation ratio to the predictive controller through a low-latency communication interface to update the actuator constraint weights and energy consumption calculation parameters in its optimization model. Before the update, perform a first-order lag filtering process on the proportional parameters, and the filtering time constant matches the controller iteration period to eliminate high-frequency disturbances. Use a double-buffer mechanism to achieve seamless switching: the current control cycle uses the old parameters to generate instructions, and the next cycle automatically loads the filtered new parameters. The updated parameters are associated with environmental states such as flight altitude and speed according to the time stamp, and are written into the historical record table of the attitude database for offline analysis of the long-term adaptability of the control strategy, and to optimize the weight adjustment rule, forming a self-evolving closed-loop control system across task cycles.

[0160] Corresponding to the above embodiments, an embodiment of the present invention provides a computer storage medium, including: at least one memory and at least one processor;

[0161] The memory is used to store one or more program instructions;

[0162] The processor is used to run one or more program instructions to execute an aircraft attitude control method based on a semiconductor microcomputer system;

[0163] Corresponding to the above embodiments, an embodiment of the present invention provides a computer-readable storage medium. The computer storage medium contains one or more program instructions, and the one or more program instructions are used to be executed by a processor to execute an aircraft attitude control method based on a semiconductor microcomputer system.

[0164] The disclosed embodiment of the present invention provides a computer-readable storage medium. Computer program instructions are stored in the computer-readable storage medium. When the computer program instructions run on a computer, the computer is caused to execute the above-mentioned aircraft attitude control method based on a semiconductor microcomputer system.

[0165] In an embodiment of the present invention, the processor may be an integrated circuit chip with signal processing capabilities. The processor may be a general-purpose processor, a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0166] It can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor. The steps of the method disclosed in combination with the embodiments of the present invention can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module may be located in a random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, register, a mature storage medium in the art. The processor reads the information in the storage medium and combines its hardware to complete the steps of the above method.

[0167] The storage medium may be a memory, for example, it may be a volatile memory or a non-volatile memory, or may include both a volatile and a non-volatile memory.

[0168] Among them, the non-volatile memory can be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically Erasable PROM (EEPROM), or a flash memory.

[0169] The volatile memory can be a Random Access Memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DRRAM).

[0170] The storage media described in the embodiments of the present invention are intended to include but not limited to these and any other suitable types of memories.

[0171] Those skilled in the art should be able to realize that in one or more of the above examples, the functions described in the present invention can be implemented by a combination of hardware and software. When applying software, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. The computer-readable medium includes computer storage media and communication media, where the communication media includes any medium that facilitates the transfer of a computer program from one place to another. The storage media can be any available medium accessible by a general-purpose or special-purpose computer.

[0172] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is only for the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, and improvements made on the basis of the technical solutions of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for controlling the attitude of an aircraft based on a semiconductor microcomputer system, characterized in that, Including: S1. The semiconductor microcomputer system collects the multi-source sensor data of the aircraft in real time, calculates the attitude error vector through the multi-modal data fusion algorithm, and constructs a three-dimensional error space coordinate system; S2. Based on the three-dimensional error space coordinate system, a real-time attitude adjustment model is designed by using a non-linear predictive controller, and the optimal attitude adjustment amount is solved within the prediction window period through the rolling horizon optimization method; S3. The optimal attitude adjustment amount is input into the actuator allocation module. Combining the current thruster response matrix and the rudder surface effectiveness coefficient of the aircraft, a thrust adjustment instruction sequence for each vector thruster and a rudder surface deflection angle instruction queue are generated through calculation; S4. During the execution of the instructions, the closed-loop verification module calculates the attitude control residual index and the energy consumption efficiency coefficient in real time, and dynamically corrects the rolling optimization weight parameters of the predictive controller; S5. Update the multi-objective optimization model according to the corrected weight parameters, adjust the control instruction allocation ratio of the actuator through the dynamic reconfiguration interface, and at the same time write the updated control parameters into the aircraft attitude database to form an adaptive closed-loop control cycle.

2. The aircraft attitude control method based on a semiconductor microcomputer system according to claim 1, wherein The semiconductor microcomputer system collects the multi-source sensor data of the aircraft in real time, calculates the attitude error vector through the multi-modal data fusion algorithm, and constructs a three-dimensional error space coordinate system, including the following sub-steps: Perform noise suppression and time-domain synchronization processing on the multi-source sensor data, and extract the error components of the attitude angular velocity, acceleration and angular displacement; Map the error components to a unified error space through a weighted fusion strategy to generate an attitude error vector; According to the dynamic change trend of the error vector, construct a three-dimensional error space coordinate system with the error amplitude, direction and change rate as the coordinate axes.

3. A method for controlling the attitude of an aircraft based on a semiconductor microcomputer system according to claim 1, characterized in that, Based on the three-dimensional error space coordinate system, a real-time attitude adjustment model is designed by using a non-linear predictive controller, and the optimal attitude adjustment amount is solved within the prediction window period through the rolling horizon optimization method, including the following sub-steps: Establish an attitude dynamics state equation based on the three-dimensional error space coordinate system; Perform discretized sampling on the attitude adjustment amount within the prediction window period, and evaluate the error convergence and energy consumption cost of each sampling point through a cost function; Use the gradient descent method to iteratively solve the minimum value of the cost function to determine the optimal attitude adjustment amount.

4. A method for controlling the attitude of an aircraft based on a semiconductor microcomputer system according to claim 1, characterized in that, The optimal attitude adjustment amount is input into the actuator allocation module. Combining the current thruster response matrix and the rudder surface effectiveness coefficient of the aircraft, a thrust adjustment instruction sequence for each vector thruster and a rudder surface deflection angle instruction queue are generated through calculation, including the following sub-steps: Calculate the thrust allocation priority of each vector thruster according to the thruster response matrix to generate a thrust adjustment instruction sequence; Perform saturation limit verification on the rudder surface deflection angle based on the rudder surface effectiveness coefficient to generate a rudder surface deflection angle instruction queue; Send the thrust adjustment instruction and the rudder surface deflection instruction to the corresponding actuator synchronously according to the time stamp.

5. A method for controlling the attitude of an aircraft based on a semiconductor microcomputer system according to claim 1, characterized in that, During the execution of the instructions, the closed-loop verification module calculates the attitude control residual index and the energy consumption efficiency coefficient in real time, and dynamically corrects the rolling optimization weight parameters of the predictive controller, including the following sub-steps: Judge the current control accuracy according to the attitude control residual index and adjust the weight coefficient of the attitude tracking error; Evaluate the energy consumption status of the actuator according to the energy consumption efficiency coefficient, and optimize the weight allocation of the energy consumption items; Feed back the adjusted weight coefficient to the non-linear predictive controller in real time.

6. A method for controlling the attitude of an aircraft based on a semiconductor microcomputer system according to claim 5, characterized in that, Judge the current control accuracy according to the attitude control residual index, and adjust the weight coefficient of the attitude tracking error, including the following sub-steps: When the attitude control residual index exceeds the preset threshold, increase the weight of the attitude tracking error term; When the residual index is lower than the threshold, gradually restore the default weight allocation.

7. A method for controlling the attitude of an aircraft based on a semiconductor microcomputer system according to claim 1, characterized in that, Update the multi-objective optimization model according to the corrected weight parameters, adjust the control instruction allocation ratio of the actuator through the dynamic reconfiguration interface, and at the same time write the updated control parameters into the aircraft attitude database to form an adaptive closed-loop control cycle, including the following sub-steps: According to the updated multi-objective optimization model, re-allocate the priorities of the thrust adjustment instruction and the rudder surface deflection instruction; When the thruster fails or the rudder surface is saturated, activate the instruction compensation logic of the redundant actuator; Feed back the adjusted instruction allocation ratio to the non-linear predictive controller to update the optimization model parameters of the next control cycle.

8. An aircraft attitude control system based on a semiconductor microcomputer system, characterized in that, Including: Data acquisition and processing module: The semiconductor microcomputer system collects multi-source sensor data of the aircraft in real time, calculates the attitude error vector through the multi-modal data fusion algorithm, and constructs a three-dimensional error space coordinate system; Predictive control algorithm module: Based on the three-dimensional error space coordinate system, use a non-linear predictive controller to design a real-time attitude adjustment model, and solve the optimal attitude adjustment amount within the prediction window period through the rolling time domain optimization method; Actuator allocation module: Input the optimal attitude adjustment amount into the actuator allocation module, combine the current thruster response matrix and the rudder surface effectiveness coefficient of the aircraft, and generate a thrust adjustment instruction sequence for each vector thruster and a rudder surface deflection angle instruction queue through calculation; Closed-loop verification and dynamic adjustment module: During the execution of the instruction, the closed-loop verification module calculates the attitude control residual index and the energy consumption efficiency coefficient in real time, and dynamically corrects the rolling optimization weight parameters of the predictive controller; Dynamic configuration management module: Update the multi-objective optimization model according to the corrected weight parameters, adjust the control instruction allocation ratio of the actuator through the dynamic reconfiguration interface, and at the same time write the updated control parameters into the aircraft attitude database to form an adaptive closed-loop control cycle.

9. A computer storage medium, characterized in that, Including: At least one memory and at least one processor; The memory is used to store one or more program instructions; The processor is used to run one or more program instructions to execute a method for controlling the attitude of an aircraft based on a semiconductor microcomputer system as described in any one of claims 1-7.

Images