Large-load eVTOL flight intelligent management and control system

Abstract

The invention discloses a large-load eVTOL flight intelligent management and control system, and relates to the technical field of intelligent control of vertical take-off and landing aircrafts. Comprising a ground effect airflow sensing and modeling module, a lift force sensitive section identification module, a lift force-load coupling evaluation module, a risk driving control parameter configuration module, a thrust adjustment sequence generation module and a posture stabilization and safe landing control module, and establishing a multi-dimensional dynamic interference model of rotor downwash airflow and terrain reflection airflow, adaptively correcting the airflow state based on laser radar and sonar sensing data, and generating a ground effect intensity distribution diagram. According to the invention, precise sensing and progressive control of sudden change of lift force caused by a ground effect are realized, stall, rebound and tipping are effectively avoided through multi-module collaborative modeling, identification, evaluation and thrust adjustment and dynamic configuration of a flight control response strategy, and the safety and stability of vertical landing of a large-load eVTOL in a complex environment are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control of vertical take-off and landing aircraft, and particularly relates to a large-load eVTOL flight intelligent management and control system. BACKGROUND

[0002] The large-load eVTOL flight intelligent management and control refers to the flight management and control requirements of electric vertical take-off and landing aircraft (eVTOL) with high load capacity in complex urban airspace or multi-task operation scenarios, which integrates multi-source perception, intelligent decision-making and adaptive control technologies, and builds a dynamic scheduling and operation optimization mechanism covering the whole process of take-off, cruising and landing. The intelligent management and control not only includes real-time monitoring and fine scheduling of the attitude, flight path, energy consumption and load of the aircraft, but also involves intelligent identification and response adjustment capabilities of aerodynamic disturbance, urban building reflection flow field, load distribution stability and flight safety boundary, to ensure safety, stability and operation efficiency under high load and high frequency flight tasks. At the same time, the system can flexibly match the control strategy according to the task type (such as logistics transportation, emergency rescue, aerial operation, etc.), to realize multi-task adaptation and whole-process collaborative optimization.

[0003] The prior art has the following disadvantages: in the process of large-load eVTOL vertical landing, when the aircraft gradually approaches the ground, the downwash airflow of the rotor is blocked by the ground to generate reflection disturbance, forming a local anti-thrust air cushion, and the lift appears nonlinear enhancement phenomenon, and then enters the ground effect area. The lift change caused by the ground effect has strong dynamic coupling characteristics, and is easily affected by multiple factors such as flight height, load distribution, number of rotors, wind speed disturbance and terrain shape, resulting in high uncertainty of the actual lift growth rate in different flight states. If the flight control system fails to real-time correct the ground effect modeling parameters to adapt to the current state of the aircraft and environmental conditions, the judgment error of the lift change trend may occur in the approach to the ground stage, triggering the flight control system to execute the thrust adjustment instruction incorrectly, which is specifically manifested as: on the one hand, the system may overcut the thrust due to the misjudgment of the overstrong lift, resulting in the aircraft stalling and sinking or even hard landing during the landing process; on the other hand, the system may output overcompensation thrust due to the lag of the system in response to the ground effect lift enhancement, causing the aircraft to bounce or attitude to fluctuate violently near the ground, and in severe cases, may cause rollover, landing gear damage or ground collision and other flight safety accidents.

[0004] The above information disclosed in the background section is only used to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0005] The purpose of the present application is to provide a large-load eVTOL flight intelligent management and control system, which realizes accurate perception and gradual control of lift mutation caused by ground effect, dynamically configures flight control response strategies through multi-module collaborative modeling, identification, evaluation and thrust adjustment, effectively avoids stall, rebound and rollover, and improves the safety and stability of large-load eVTOL in complex environments. Vertical landing, to solve the problems in the above background technology.

[0006] In order to achieve the above purpose, the present application provides the following technical scheme: a large-load eVTOL flight intelligent management and control system, comprising a ground effect airflow perception and modeling module, a lift sensitive section identification module, a lift-load coupling evaluation module, a risk-driven control parameter configuration module, a thrust adjustment sequence generation module, and an attitude stabilization and safe landing control module:

[0007] The ground effect airflow perception and modeling module establishes a multi-dimensional dynamic interference model of the airflow under the rotor and the terrain reflected airflow, adaptively corrects the airflow state based on laser radar and sonar perception data, and generates a ground effect intensity distribution map;

[0008] The lift sensitive section identification module calculates the lift change rate and amplitude in the flight trajectory according to the ground effect intensity distribution map, marks the critical height interval of abnormal growth of lift, and generates a lift sensitive section list;

[0009] The lift-load coupling evaluation module fuses the lift sensitive section list and the current load state of the aircraft, calculates the load distribution deviation trend and attitude instability index in each section, and determines the coupling risk level of lift and load;

[0010] The risk-driven control parameter configuration module configures the flight control response threshold according to the coupling risk level, and constructs a real-time deviation correction curve of the lift prediction value and the actual feedback value;

[0011] The thrust adjustment sequence generation module generates thrust control instructions based on the deviation correction curve, constructs a thrust progressive adjustment sequence, and avoids attitude instability and landing rebound caused by incorrect adjustment;

[0012] The attitude stabilization and safe landing control module executes the thrust progressive adjustment sequence, feeds it back to the rotor control module, smooths the thrust change slope, stabilizes the flight attitude, and suppresses the lift mutation impact.

[0013] Preferably, the step of establishing a multi-dimensional dynamic interference model of the airflow under the rotor and the terrain reflected airflow comprises:

[0014] Based on the rotor structure parameters, flight height and ground topography, a three-dimensional airflow model is constructed;

[0015] Laser radar perception data and sonar perception data are obtained and preprocessed synchronously;

[0016] The perception data is compared with the airflow model, and the model is adaptively corrected based on residual analysis and parameter updating algorithm;

[0017] A high-resolution ground effect intensity distribution map is generated to describe the spatial distribution of airflow disturbance.

[0018] Preferably, the step of calculating the lift change and amplitude according to the ground effect intensity distribution map and generating a lift-sensitive section list includes:

[0019] The flight trajectory is matched with the ground effect intensity distribution map to form a correspondence between the flight path and the ground effect intensity;

[0020] The lift change rate and lift change amplitude are calculated based on the matching results;

[0021] The critical height interval of abnormal growth of lift is identified according to the lift change rate and amplitude;

[0022] A plurality of critical height intervals are recorded to generate a lift-sensitive section list and provide it to the flight control system for calling.

[0023] Preferably, the step of fusing the lift-sensitive section list with the current load state of the aircraft and constructing a lift and load coupling risk level includes:

[0024] Obtain the current load state data of the aircraft to establish a three-dimensional load distribution model;

[0025] Fuse the load state data with the lift-sensitive section list in time and position synchronization;

[0026] Calculate the load distribution deviation trend and attitude instability index in each section;

[0027] Construct a lift and load coupling risk level based on the attitude instability index and output it to the flight control system.

[0028] Preferably, the step of constructing a real-time deviation correction curve of the lift prediction value and the actual feedback value includes:

[0029] Set the response threshold in the flight control parameters according to the lift and load coupling risk level;

[0030] Establish a lift prediction model and generate a predicted lift value based on the current flight state;

[0031] Obtain the actual lift feedback value of the aircraft at present and calculate the deviation amount between the predicted lift value and the actual lift feedback value;

[0032] Construct a lift deviation correction curve based on the deviation amount and output a correction control instruction to the flight controller.

[0033] Preferably, the step of generating thrust control commands based on the deviation correction curve and constructing a thrust gradual adjustment sequence comprises:

[0034] Analyzing the lift deviation correction curve to extract thrust adjustment target parameters;

[0035] Generating thrust control output commands based on the thrust adjustment target parameters;

[0036] Decomposing the thrust control output commands into continuous thrust adjustment values within multiple time steps to form a thrust gradual adjustment sequence;

[0037] Outputting the thrust gradual adjustment sequence to the flight control module and executing multi-rotor cooperative control.

[0038] Preferably, the specific steps of executing the thrust gradual adjustment sequence and feeding it back to the rotor control module to smooth the thrust change slope, stabilize the flight attitude, and suppress lift sudden impact are as follows:

[0039] According to the lift deviation correction curve, the total change amount required for thrust adjustment is extracted, denoted as ΔT, and combined with the current attitude stability state and adjustment demand, a nonlinear gradual thrust output command is calculated, and the calculation expression is as follows:

[0040]

[0041] , wherein T(t i ) is the thrust output value at the i-th time point t, T0 is the initial thrust value, α is the thrust adjustment amplitude factor, β is the thrust adjustment slope control parameter, N is the total number of time points, i is the time point index, ΔT is the thrust adjustment amplitude, and tanh is the hyperbolic tangent function;

[0042] The generated thrust output value T(t i ) is distributed to each rotor, and the output proportion of each rotor is adjusted according to the current attitude state of the aircraft to achieve attitude stability control. First, define the current attitude state vector of the aircraft as Θ(t), Θ(t) = [φ(t), θ(t), ψ(t)], where φ(t) is the roll angle at time point t, θ(t) is the pitch angle at time point t, and ψ(t) is the yaw angle at time point t;

[0043] In order to evaluate the stability of the aircraft attitude, the attitude deviation cost index is calculated:

[0044] C(Θ) = γ1·|φ(t)| + γ2·|θ(t)| + γ3|ψ(t)|

[0045] , wherein C(Θ) is an attitude deviation cost index, representing the degree of deviation of the current overall attitude from the ideal horizontal attitude, and γ1, γ2 and γ3 are weight coefficients of the roll angle φ(t), the pitch angle θ(t) and the yaw angle ψ(t) respectively;

[0046] To realize the attitude-thrust linkage control, the distributed thrust correction is performed on each rotor based on the attitude deviation cost index, and the adjustment formula is as follows:

[0047]

[0048] , wherein T j (t i ) is the thrust output value of the jth rotor at the time point t, T j is the thrust of the jth rotor, is the sensitivity of the attitude deviation cost to the thrust of the jth rotor.

[0049] In the above technical solution, the technical effects and advantages provided by the present application are as follows:

[0050] The present application realizes the whole-process closed-loop management of accurate perception, dynamic identification, risk modeling and gradual control response of the nonlinear sudden change problem of lift caused by ground effect in the vertical landing stage of the aircraft. The system modules work collaboratively, from airflow modeling to thrust output, and can construct a refined ground effect distribution map according to real-time perception data, identify abnormal lift areas in advance, and quantify attitude instability risks combined with the current load, dynamically configure the flight control threshold and thrust adjustment path, and finally ensure the consistency of flight attitude stability and controller response through multi-rotor distributed execution. This scheme effectively avoids the stall sinking, rebound oscillation or rollover accidents caused by misjudgment of ground effect in the flight control system, significantly improves the safety, stability and control accuracy of large-load eVTOL aircraft in complex urban airspace or platform environment in low-altitude vertical landing, and has good engineering adaptability and popularization value. BRIEF DESCRIPTION OF DRAWINGS

[0051] To more clearly illustrate the technical solutions in the embodiments or prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art based on these drawings.

[0052] Figure 1 A module schematic diagram of the large-load eVTOL flight intelligent management and control system of the present application. DETAILED DESCRIPTION

[0053] Example implementations are now described with reference to the drawings. Example implementations can, however, be implemented in many different forms and should not be construed as limited to the examples set forth herein; rather, these example implementations are provided so that this disclosure will be thorough and complete, and will fully convey the inventive gist to those skilled in the art.

[0054] The present application provides a large-load eVTOL flight intelligent management and control system as shown in Figure 1 The large-load eVTOL flight intelligent management and control system includes a ground effect airflow perception and modeling module, a lift-sensitive section identification module, a lift-load coupling evaluation module, a risk-driven control parameter configuration module, a thrust adjustment sequence generation module, and an attitude stabilization and safe landing control module.

[0055] The ground effect airflow perception and modeling module establishes a multi-dimensional dynamic interference model of the airflow under the rotor and the terrain reflected airflow, acquires laser radar perception data and sonar perception data, performs adaptive correction on the airflow model based on the perception data, and generates a ground effect intensity distribution map.

[0056] In order to realize accurate modeling and dynamic response control of the nonlinear lift change caused by ground effect during the flight process of a large-load electric vertical take-off and landing aircraft, especially in the vertical landing stage, a ground effect interference modeling method based on multi-source perception and adaptive modeling fusion is proposed, which includes the following four steps:

[0057] First, a multi-dimensional dynamic interference basic model of the airflow under the rotor and the ground reflected airflow is constructed. The basic model takes the parameters of the aircraft rotor arrangement, rotor diameter, speed, angle of elevation, ground material, ground slope, and distance from the ground height as modeling inputs, pre-establishes airflow field models under multiple typical working conditions using a three-dimensional fluid dynamics simulation software (such as CFD software), and simulates the disturbance behavior of irregular terrain reflected airflow by introducing a disturbance factor. In this model, a parameterized grid structure is introduced to express the airflow velocity field, pressure field, and vorticity distribution, to support subsequent fast matching and adaptive correction with measured perception data. This model has certain initial generalization ability and can cover common ground boundary types in urban airspace, such as building roofs, platforms, road surfaces, etc., providing model prior support for subsequent perception data matching.

[0058] Secondly, real-time sensing data is acquired to capture the downwash airflow and terrain reflection disturbance in the actual flight environment. The application arranges multiple sets of high-precision laser radar and ultrasonic sonar sensors on the bottom and periphery of the aircraft. The laser radar is used to acquire ground height changes, local building edges, slopes, and ground topography information. The sonar sensor is used to detect airflow disturbance echo characteristics at different height layers below the aircraft. After time synchronization mechanism and flight state parameters (such as flight height, attitude angular velocity, and load position) are spatially calibrated, the sensing data is input into the data processing module for preprocessing. The preprocessing process includes noise filtering, wavelet decomposition, outlier rejection, and scale normalization to ensure the stability and matching of the sensing data. In addition, a time window sliding processing mechanism is introduced in this embodiment to enable the system to continuously perceive airflow disturbance patterns at different time points and capture the dynamic trend of interference changes.

[0059] Thirdly, the downwash airflow disturbance model is adaptively corrected based on the above sensing data. The system uses an adaptive modeling method based on the least square method and the Bayesian parameter updating strategy. First, the spatial comparison of the preset airflow model output and the point cloud reconstruction of the sensing data is performed to calculate the difference residual of the local area air pressure gradient and the velocity vector. Then, the key parameters in the model are dynamically corrected based on the Bayesian posterior probability inference method, such as the iterative update of the rotor induced velocity decay coefficient, the reflection disturbance diffusion factor, and the ground effect enhancement factor, to gradually approximate the current actual state. In this process, a confidence weighting mechanism is introduced to weight and distribute the contribution of different sensing channels (such as laser radar and sonar) in different environmental scenarios, improving the stability and robustness of model correction. In this way, the system can adaptively adjust the airflow model structure in real time, making it more consistent with the real aerodynamic disturbance characteristics of the current flight state and ground environment.

[0060] Finally, a high-resolution ground effect intensity distribution map is generated based on the corrected multi-dimensional airflow model. The distribution map takes the bottom area of the aircraft as the reference surface and models according to the spatial grid. Each cell in the map is labeled with the ground effect intensity coefficient at the corresponding position, which is a function of the lift enhancement ratio and disturbance degree. To realize the visualization of the ground effect changes, the distribution map uses the heat map superimposed with the airflow vector field to display, and the system updates the ground effect map every predetermined time interval (such as 50 milliseconds) for subsequent lift change analysis and dynamic scheduling of flight control strategies. The distribution map can not only be used for early identification of sensitive areas in the flight path, but also as training data to input into the learning algorithm in the flight control system, enhancing its cognitive ability and response ability to ground effect disturbance characteristics.

[0061] Through the cooperative execution of the above four steps, the ground effect sensing and modeling mechanism provided by the embodiment can effectively capture the aerodynamic disturbance characteristics of a large-load electric vertical take-off and landing aircraft in the approach-to-ground phase, form a high-precision, real-time updateable ground effect intensity distribution map, and provide an accurate lift trend prediction basis for the flight control system, thereby significantly improving the safety and control accuracy of the aircraft in the landing phase.

[0062] The lift-sensitive section identification module continuously calculates the lift change rate and lift change amplitude in the flight trajectory according to the ground effect intensity distribution map, identifies the critical height interval of abnormal growth of lift, and generates a lift-sensitive section list;

[0063] In order to realize the early identification and dynamic management of the possible nonlinear lift mutation of a large-load electric vertical take-off and landing aircraft in the vertical landing phase, a lift change analysis method based on the ground effect intensity distribution map is proposed to generate a lift-sensitive section list in the flight trajectory. The method includes the following four steps:

[0064] First, a flight trajectory matching relationship is established based on the ground effect intensity distribution map. In this embodiment, the system continuously records the current and historical flight trajectory data of the aircraft, including vertical height, flight attitude angle, descent speed, and time stamp corresponding to each height layer. At the same time, the ground effect intensity distribution map generated in the previous module exists as a three-dimensional space vector grid, and the system maps each position point passed by the flight trajectory to the corresponding ground effect intensity grid, thereby forming a "flight path—ground effect intensity" matching data set. The mapping process uses interpolation and fusion algorithms for position accuracy correction, and constructs a corresponding matrix between flight state and ground effect intensity, providing a data basis for subsequent lift change trend calculation.

[0065] Second, based on the flight path matching ground effect intensity data, the lift change rate and lift change amplitude are calculated. In this step, the system samples and analyzes the lift gain change experienced by the aircraft at different height points in time series. The real-time thrust output value and acceleration data fed back by the aircraft power system are used to calculate the actual lift level, and the intensity distribution value of the ground effect area is compared to deduce the unit height lift enhancement caused by the ground effect. The lift change rate is defined as the change rate of lift increment in the unit height descent process, and the lift change amplitude is defined as the cumulative lift enhancement value in a certain height interval. This process uses a sliding window difference algorithm and a multi-order derivative detection method to ensure that the inflection points, transition points and platform regions in the lift curve can be captured, thereby fully reflecting the nonlinear characteristics of the lift evolution trend.

[0066] Third, identify the critical height interval of abnormal lift growth. In this step, the system constructs a lift evolution curve according to the calculated lift change rate and lift change amplitude, and establishes a lift safety change envelope combining the maximum thrust adjustment rate allowed by the aircraft, the maximum attitude transformation limit and the maximum longitudinal load margin. If the lift change rate in a certain height interval exceeds the envelope safety threshold, or the change amplitude jumps significantly within a unit time, it is determined that the interval is the critical area of abnormal lift growth. The system automatically marks this height interval and records it in the form of a three-tuple of height start value, height end value and corresponding lift gain mutation, providing data support for the subsequent flight control system adjustment response strategy. To enhance the robustness of identification, the system also introduces a multi-point joint criterion to avoid misidentification caused by a single abnormal point, and improves the discrimination accuracy through repeated multiple rounds of sample comparison.

[0067] Finally, generate a lift sensitive section list and construct a structured data set for real-time call. The lift sensitive section list is composed of multiple critical height intervals, each recording the corresponding height range, lift change intensity, attitude disturbance trend, ground effect level and other information. This list is resident in memory as a high-priority dynamic data structure in the flight control system, and the flight control strategy call module can quickly match the current aircraft position with the list to activate attitude stabilization strategies, preset thrust smoothing curves and other control measures in advance. This list can also be used as historical flight data archive to train ground effect prediction models, further improving the adaptive control ability of the system in different urban airspace.

[0068] Through the above four steps, the embodiment of the application realizes the accurate identification and list processing of the abnormal lift section of the large-load electric vertical take-off and landing aircraft during the landing process, provides high real-time, clear structure and responsive flight safety information for the flight control system, effectively reduces the risk of thrust misjudgment, attitude instability or landing rebound caused by ground effect, and significantly improves the stability and safety of the whole machine system.

[0069] The lift-load coupling evaluation module obtains the current load state data of the aircraft, fuses the lift sensitive section list with the current load state data, calculates the load distribution deviation trend and attitude instability index in each lift sensitive section, and constructs the lift-load coupling risk level;

[0070] To improve the pre-judgment ability and control accuracy of the flight control system of the large-load electric vertical take-off and landing aircraft when facing ground effect disturbance near the ground, a posture stability evaluation method is proposed, which fuses the load state and lift change characteristics. This method is based on the dynamic fusion of the lift sensitive section list and the load state data of the aircraft, identifies the potential attitude instability during flight, and outputs the risk level of lift-load coupling. The method includes the following four steps:

[0071] First, the current load state data of the aircraft is acquired and a dynamic load distribution model is established. In this embodiment, a plurality of high-precision stress sensors, gravity distribution detection units and three-dimensional inertial measurement units are deployed inside the aircraft body to obtain real-time data such as the spatial position, distribution pattern and center of mass drift trend of the current load. The system reconstructs a three-dimensional load distribution map by fusing these sensor data, and matches it with the aircraft structure model to generate a load distribution offset function. This function is used to dynamically express the vector deviation between the current center of gravity of the aircraft and the design balance point, and can reflect changes in the load state caused by changes in tasks, shifts in goods or sudden disturbances. In order to improve the modeling accuracy, the system also introduces flight task information (such as the type of logistics goods and the position of the cargo container) as a priori supplementary parameter to adapt to the load configuration differences in different task scenarios.

[0072] Second, the real-time load state data described above is fused with the list of lift-sensitive sections. In this step, the system jointly analyzes the load distribution state data within the height interval corresponding to each lift-sensitive section with the flight height interval. The fusion process is based on timestamp alignment and position synchronization technology to ensure that the lift change information and load state at the same time and spatial position have a one-to-one correspondence. The system establishes a "lift disturbance-load distribution" correlation matrix to map the structural stress distribution and center of mass change amplitude of the aircraft when the lift suddenly changes. Through this fusion mechanism, the originally static lift-sensitive section can be expanded into a risk section with load characteristics, providing accurate input for subsequent stability evaluation.

[0073] Third, the attitude instability index is calculated and the lift-load coupling risk model is constructed. In this step, the system introduces a composite attitude instability evaluation function that considers the following three key factors: 1) the instantaneous change rate of the load center of gravity offset vector; 2) the nonlinear strength of the lift mutation within the corresponding lift-sensitive section; 3) the current attitude adjustment capability of the aircraft (including available residual thrust margin, control surface deflection capability, etc.). The system fuses the three factors through a weighted model and outputs a standardized attitude instability index. At the same time, this index is used in conjunction with the ground effect model to determine whether the load change will exacerbate the negative impact of the ground effect lift mutation on the attitude. Finally, the system defines the lift and load coupling risk levels according to the numerical range of the attitude instability index, combined with the task fault tolerance level and the flight control adjustment response capability, including three types of low risk, medium risk and high risk, each corresponding to a different flight control strategy response mechanism.

[0074] Finally, the lift-load coupling risk level is output to the flight control system, and a risk level dataset is established for strategy calling. The system binds each lift-sensitive section with its corresponding coupling risk level to form a structured "risk section index table", each item in the table containing: flight altitude interval, load offset trend, attitude instability index, coupling risk level. This index table is continuously updated during flight and is read in real time by the flight control strategy configuration module to dynamically adjust the thrust output slope, controller sensitivity threshold, and emergency attitude compensation instructions. This mechanism enables the aircraft to actively adjust the attitude control curve in advance when the high-risk attitude section is identified, effectively mitigating the stability threat caused by the coupling of load offset and lift enhancement. At the same time, this risk index table can also be used as a prediction module in the flight control system for flight path optimization and control instruction feedforward compensation model construction.

[0075] In summary, through the above four steps, the embodiment of the present application can realize early identification and quantitative evaluation of the attitude instability trend of large-load aircraft under complex ground effect disturbances, forming a lift-load coupling risk level data with high adaptability and high resolution, providing accurate support for subsequent flight control decision-making and attitude stability strategy formulation, significantly improving flight safety and control system robustness during landing.

[0076] The risk-driven control parameter configuration module configures the response threshold of the flight control system according to the lift-load coupling risk level, and constructs a real-time deviation correction curve between the lift prediction value and the actual lift feedback value;

[0077] To cope with the attitude instability risk of large-load electric vertical take-off and landing aircraft in ground effect disturbance environment, and to improve the response accuracy of the flight control system to lift mutations, a method is proposed based on dynamic configuration of response threshold and construction of real-time lift deviation correction curve according to the coupling risk level. This method combines the lift-sensitive sections and load state coupling risk levels generated in the previous steps to optimize the prediction and feedback correction mechanism of lift variation trend, which includes the following four steps:

[0078] First, the response threshold of the flight control system is configured according to the lift-load coupling risk level. The system receives the lift-sensitive section index table and its corresponding coupling risk level generated in the previous module, and dynamically sets the control parameters according to the mapping rules between the risk level and the flight control strategy. The response threshold includes: thrust adjustment upper limit, attitude adjustment maximum rate, control loop gain coefficient, sensor abnormal tolerance range and other dimensions. The response threshold configuration module adopts a hierarchical parameter integration structure, and each risk level corresponds to a complete set of parameters. Under high risk level, the control system will adopt a more compact adjustment strategy, such as reducing the attitude mutation threshold, increasing the feedback frequency, and shortening the control cycle delay; while under low risk level, the adjustment boundary is appropriately relaxed to reduce the system calculation load, realizing the adaptive balance of risk and resources.

[0079] Second, the lift prediction model is established and the predicted lift value is continuously output. In this step, the system uses the current flight height, ground effect intensity distribution, aircraft load state and historical air flow disturbance data at each flight time point to generate the predicted lift value under this state through the regression prediction algorithm. To improve the prediction accuracy, a time series model and a recurrent neural network structure are used in this embodiment to model the lift evolution trend under different flight states. The model has online learning function and can continuously update parameters according to the actual lift data fed back in each flight task, so as to gradually approach the real curve of lift change. The prediction value output module updates the prediction data at fixed time intervals (such as every 50 milliseconds) and uses it as the reference baseline for subsequent deviation calculation.

[0080] Third, the real-time lift feedback value is collected and the deviation amount is calculated. The lift feedback value is calculated by the flight control system according to the real-time thrust output of each rotor of the aircraft, the acceleration data and flight attitude angle data obtained by the inertial measurement unit. The system corrects multiple parameters through the power model combined with the center of gravity position and air density to improve the physical accuracy of the feedback lift. On this basis, the system calculates the difference between the predicted lift value and the actual feedback lift value in real time to form a time series of deviation amount. To further enhance stability, a sliding weighted filtering mechanism is also introduced in this embodiment to weaken the error noise caused by sudden disturbance or sensor jitter, ensuring that the deviation calculation curve has high fidelity and controllability.

[0081] Finally, a real-time deviation correction curve is constructed based on the time series of the deviation, and an adjustment instruction is output to the flight controller. The system uses a fitting algorithm to model the deviation data as a curve, forming a continuous deviation trend graph between lift prediction and feedback. This curve is input to the adjustment module of the flight controller, which is used to real-time correct the lift estimation reference of the controller. When the deviation trend continues to expand, the controller will actively adjust the thrust distribution strategy, such as speeding up the thrust response speed or redistributing the multi-rotor output proportion; when the deviation is small, the control response can be appropriately delayed to reduce the system shock caused by frequent action of the controller. In addition, the deviation correction curve is also used to update the feedforward control path in the flight controller, realizing the transition from "feedback compensation" to "prediction preposition" strategy, so as to actively control the lift disturbance in advance, further enhancing the landing stability and control robustness of the aircraft.

[0082] Through the above four steps, the embodiment provided by the present application not only realizes dynamic tracking and high-precision correction of the lift change trend, but also establishes a flight control response mechanism based on risk level, effectively reduces the problems of attitude instability, thrust misadjustment or system lag caused by lift prediction error, and provides a solid control foundation for safe landing of large-load eVTOL in complex airspace.

[0083] The thrust adjustment sequence generation module generates thrust control output instructions according to the real-time deviation correction curve, and constructs a thrust progressive adjustment sequence to reduce the attitude instability and landing rebound caused by misadjustment;

[0084] In order to improve the stability of large-load electric vertical take-off and landing aircraft in the process of vertical landing in the face of complex ground effect interference environment, a thrust control output method driven by real-time deviation correction curve is proposed. This method analyzes the deviation trend between the lift prediction value and the feedback value, dynamically adjusts the thrust output control instruction, and constructs a thrust adjustment sequence in a progressive manner, so as to effectively avoid the problems of aircraft attitude mutation, instability or ground rebound caused by misadjustment. The method includes the following four steps:

[0085] Firstly, the lift deviation correction curve is analyzed to extract the thrust adjustment target information. The system receives the deviation correction curve formed between the lift prediction value and the actual feedback value, and based on the characteristic parameters such as the slope, the change rate, the inflection point position of the curve, identifies the lift mutation trend and the urgency of the correction demand. For example, when the slope of the deviation curve increases rapidly, it means that the lift prediction error is rapidly expanding, and a quick response is needed for thrust compensation; when the deviation curve enters the platform period, it means that the lift state has stabilized, and the adjustment rate can be slowed down to avoid system overshoot. This process uses a set of multivariate functions to calculate, including curvature fitting analysis, gradient dynamic detection and peak trend regression, to generate a set of target thrust adjustment vectors, which clearly define the adjustment direction, intensity and time window, as the decision basis for subsequent thrust control output.

[0086] Secondly, the thrust control output instruction is generated according to the adjustment target vector. In this step, the system combines the current thrust distribution state of the aircraft, the rotor output capacity, the remaining power redundancy and the attitude response time delay, and calculates the multi-rotor joint thrust output strategy by using a nonlinear optimization algorithm to form a set of thrust adjustment instructions with constraint boundaries. Specifically, the system needs to meet the following three constraint conditions in the control instruction generation process:

[0087] (1) The adjustment rate of each rotor cannot exceed its electric control response capacity;

[0088] (2) The overall thrust change amplitude cannot cause the attitude instability caused by the center of gravity moving outside the aircraft;

[0089] (3) The control system execution time delay and the aerodynamic response need to be time-matched to prevent control oscillation caused by feedback lag. Based on these constraints, the system preferentially selects a "small amplitude and multi-frequency" adjustment strategy, that is, the large amplitude thrust adjustment requirement is decomposed into a number of small amplitude continuous adjustment operations to realize the smoothing of the output path.

[0090] Thirdly, the thrust gradual adjustment sequence is constructed and the time step control strategy is set. The thrust gradual adjustment sequence refers to the decomposition of a single thrust adjustment requirement into a sequence of control output instructions that are continuous in time and smooth in amplitude. In this embodiment, the system divides each adjustment period into a plurality of fixed time step units (for example, each 20 milliseconds is a step), and calculates the thrust increment to be adjusted in each step. The system introduces an adaptive increment factor according to the dynamic trend of the deviation curve in each time step, so that the adjustment speed can be adjusted in time according to the system response. For example, when the attitude stability is good, the thrust adjustment amplitude of each step can be increased to speed up the control convergence; and when the attitude disturbance is significant, the adjustment step is automatically reduced and the response period is prolonged to realize system protection. During the entire sequence execution period, the system also introduces a redundancy detection mechanism to prevent the thrust execution deviation from being too large due to sensor failure or actuator response anomaly, which affects the safety of the aircraft.

[0091] Finally, the thrust progressive adjustment sequence is output to the flight control module and the multi-rotor performs cooperative control. After receiving the complete thrust adjustment sequence, the flight control module issues it to the control drive unit of each rotor in chronological order, while implementing a dynamic verification mechanism in combination with attitude angular velocity feedback to ensure that the actual output is highly consistent with the expected adjustment curve. During the adjustment process, the system continuously monitors the attitude stability indicators of the aircraft, including roll angle rate, pitch angle rate, vertical acceleration, and ground speed change rate. If any of these indicators exceeds the safety boundary, the current thrust sequence execution is immediately interrupted, and the built-in rapid attitude recovery strategy is enabled to prevent system loss of control. In addition, to achieve higher level safety redundancy, the module also supports uploading thrust adjustment history data to the upper control system for task playback and flight control model optimization, providing data basis and strategy adjustment basis for future tasks.

[0092] In summary, through the above four steps, the embodiment realizes a thrust progressive control method based on real-time lift deviation trend dynamic adjustment. This method has the characteristics of smooth control output, adaptive response rate, and attitude stability guarantee, and is particularly suitable for large-load eVTOL aircraft performing high-precision vertical landing operations in complex urban airspace or ground disturbance environments. It effectively improves the anti-disturbance ability and control accuracy of the aircraft during the critical landing stage, significantly reduces the risk of hard landing, rebound, or overturning, and other flight safety risks.

[0093] The attitude stabilization and safe landing control module executes the thrust progressive adjustment sequence and feeds back the adjustment sequence to the rotor control module to smooth the change slope of the thrust output, stabilize the flight attitude, suppress the disturbance caused by lift mutation, and ensure the safe landing of the aircraft.

[0094] The core function of this step is to apply the thrust progressive adjustment sequence that has been corrected for lift deviation and optimized by multiple factors to the aircraft rotor actuators accurately and precisely, achieving attitude stabilization control and disturbance suppression of the aircraft during the landing stage. Specifically, large-load eVTOLs are affected by ground effect, load shift, wind field disturbance, and other nonlinear factors during landing, which can easily cause lift mutation, excessive thrust response, attitude shaking, and other instability phenomena. If the control system directly issues the thrust adjustment signal without smoothing to the rotor control unit, it may cause abrupt changes in thrust output, leading to attitude oscillation, landing rebound, or even overturning of the aircraft.

[0095] Therefore, this step subdivides the thrust variation process in the time dimension by executing a thrust progressive adjustment sequence, so that the thrust output is slowly adjusted in a continuous and controllable gradient manner, thereby reducing the sensitivity of the system to disturbances. At the same time, by feeding back the adjustment sequence to the rotor control module, coordinated control of multiple rotors is realized, so that the thrust output is dynamically balanced among the rotors, further reducing the attitude deviation caused by local thrust imbalance. The system also monitors the change slope of the thrust output in real time and dynamically limits the amplitude, ensuring that the control system does not overcompensate or underrespond when lift mutation occurs. Ultimately, this mechanism can effectively suppress the sudden disturbance caused by the enhancement of ground effect, ensuring that the aircraft completes a safe landing with a stable and gentle attitude during the approach to the ground phase, significantly improving the landing safety and disturbance rejection capability of the overall system.

[0096] The specific steps of executing a thrust progressive adjustment sequence, feeding it back to the rotor control module, smoothing the thrust change slope, stabilizing the flight attitude, and suppressing the impact of lift mutation are as follows:

[0097] According to the lift deviation correction curve, the total variation required for thrust adjustment is extracted, denoted as ΔT, and combined with the current attitude stability state and adjustment demand, a nonlinear progressive thrust output command is calculated, and the calculation expression is as follows:

[0098]

[0099] , where T(t i ) is the thrust output value at the i-th time point t, which is the expected thrust value calculated at the i-th control time in the execution of the thrust progressive adjustment sequence, T is the reference thrust sent to the flight control system for control of the rotor motor output, T is the direct target of the adjustment system output, T0 is the initial thrust value, which is the reference thrust value at the start of the current adjustment sequence, and is usually equal to the actual output thrust at the last stable flight time, serving as the starting point of thrust adjustment to avoid jump control, α is the thrust adjustment amplitude factor, with a value range of α ∈ (0, 1], used to control the proportion factor actually effective in this thrust adjustment, limiting the adjustment intensity to prevent the thrust from changing too fast in a short time, β is the thrust adjustment slope control parameter, determining the "steepness" factor of the adjustment curve change speed, β > 0, the larger the value, the steeper the curve, the more aggressive the response, the smaller the value, the slower the adjustment, suitable for conservative stability strategy, N is the total number of time points, i is the time point index, i ∈ [0, N], ΔT is the thrust adjustment amplitude, which is the difference between the target and the current thrust, is the thrust range to be compensated in this adjustment, used to determine the thrust "span" of the adjustment sequence, determining the vertical range of the output curve, and tanh is the hyperbolic tangent function.

[0100] In the process of thrust gradual adjustment, the main role of introducing the hyperbolic tangent function tanh is to construct a continuous, derivable, and nonlinear thrust variation path with "S-shaped curve" characteristics to achieve the smooth transition of thrust from the initial value to the target value. Specifically, the tanh function has the characteristics of rapid change in the middle and gradual change at both ends in mathematics, with a value range of (-1, 1), a continuous derivative, and a controlled change, which makes it very suitable for "smoothing" the sudden thrust adjustment process into a natural adjustment curve of "slow start-middle acceleration-end slow stop". By introducing the slope control parameter β, the fast and slow degree of thrust response can be flexibly adjusted, thereby effectively avoiding the attitude disturbance or system shock of the aircraft caused by the rapid change of the thrust command. In this system, the tanh function not only enhances the flexibility and stability of thrust adjustment, but also improves the control robustness of ground effect interference and load coupling changes, and is a key mathematical tool for realizing nonlinear output adjustment of thrust.

[0101] The role of this step is to discretize the target thrust adjustment task into continuous gradual changes in multiple time segments through mathematical modeling, thereby avoiding phenomena such as increased lift disturbance, attitude oscillation, or control system instability of the aircraft during the close-to-ground phase caused by single sudden output.

[0102] The generated thrust output value T(ti) is assigned to each rotor, and the output proportion of each rotor is adjusted according to the current attitude state of the aircraft to achieve attitude stability control. First, define the current attitude state vector of the aircraft as Θ(t), Θ(t) = [φ(t), θ(t), ψ(t)], where φ(t) is the roll angle at time point t, representing the angle of rotation of the aircraft around its forward direction (longitudinal axis), theoretically [-180°, +180°], but commonly used for control within ±45° range, θ(t) is the left-right tilt of the aircraft, affecting lateral stability, θ(t) is the pitch angle at time point t, which is the angle of rotation of the aircraft around its lateral axis, controlled within [-30°, +30°] interval to prevent excessive lifting or diving, and ψ(t) is the yaw angle at time point t, which is the angle of rotation of the aircraft around its vertical axis, theoretically [0°, 360°] periodically, but the instantaneous change is usually less than ±20°, describing whether the aircraft heading is deviated.

[0103] To evaluate the stability of the aircraft attitude, calculate the attitude deviation cost index:

[0104] C(Θ) = γ1·|φ(t)| + γ2·|θ(t)| + γ3·|φ(t)

[0105] , wherein C(Θ) is an attitude deviation cost index representing the degree of deviation of the current overall attitude from the ideal horizontal attitude (zero attitude), used to dynamically evaluate the current attitude stability state and provide real-time basis for the thrust redistribution strategy, γ1, γ2, and γ3 are weight coefficients of the roll angle φ(t), the pitch angle θ(t), and the yaw angle ψ(t) respectively, used for weighted calculation of the overall attitude deviation severity, which are preset by the system and can be configured according to the task type, the aircraft structure (such as longitudinal / lateral symmetry), and the load position;

[0106] To realize the attitude-thrust coupled control, distributed thrust correction is performed on each rotor based on the attitude deviation cost index, and the adjustment formula is as follows:

[0107]

[0108] , wherein T j (t i ) is the thrust output value of the jth rotor at the time point t, is the control instruction finally acting on the electronic governor and the actuator of the jth rotor, directly affecting the lift generated by the rotor, according to the specific direction and degree of the attitude imbalance of the aircraft, the thrust value is dynamically weighted and adjusted to offset or balance the attitude deviation trend, T j is the thrust of the jth rotor, is the sensitivity of the attitude deviation cost to the thrust of the jth rotor, representing the influence intensity of the change of the thrust value of the jth rotor on the overall attitude instability index C(Θ), if the thrust of a certain rotor rises, the attitude deviation increases, then the partial derivative is positive, the controller will reduce the current output proportion of the rotor, and vice versa, thus realizing the coupled control of thrust distribution and attitude correction, and improving the flight stability.

[0109] The thrust adjustment mechanism can dynamically maintain the total thrust while differentially adjusting among the multi-rotors, thereby suppressing the flight attitude deviation caused by ground effect, load disturbance, or incorrect instructions, and ensuring that the aircraft completes the landing operation with a stable attitude close to the ground.

[0110] Through the cooperative control of the above two steps, the embodiment provides a thrust output adjustment method based on real-time prediction and nonlinear feedback fusion, which can effectively improve the overall control stability and safe landing ability of the aircraft in the flight phase with strong ground effect interference and unstable attitude, and has high engineering practical value and popularization potential.

[0111] Through the large-load eVTOL flight intelligent management and control system proposed above, the whole-process closed-loop management of accurate perception, dynamic identification, risk modeling and gradual control response of the lift nonlinear mutation problem caused by ground effect in the vertical landing stage of the aircraft can be realized. The modules of the system work collaboratively, from airflow modeling to thrust output, can construct a refined ground effect distribution map according to real-time perception data, identify abnormal lift areas in advance, and combine the current load to quantify the attitude instability risk, dynamically configure the flight control threshold and thrust adjustment path, and finally ensure the flight attitude stability and controller response consistency through multi-rotor distributed execution. This scheme effectively avoids the stall sinking, rebound oscillation or overturning accidents caused by misjudgment of ground effect in the flight control system, significantly improves the safety, stability and control accuracy of large-load eVTOL aircraft in complex urban airspace or platform environment during low-altitude vertical landing, and has good engineering adaptability and popularization value.

[0112] The above formulas are dimensionless numerical calculations, and the formulas are obtained by software simulation of a large amount of data to obtain a formula of the latest real situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0113] The above only describes some exemplary embodiments of the present application by way of illustration, and it is self-evident that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present application. Therefore, the above figures and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present application.

[0114] It should be noted that in this paper, if there are relationship terms such as first and second, they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment. Without more limitations, the element defined by the statement "including a" does not exclude the presence of another identical element in the process, method, article or equipment including the element.

[0115] It should be understood that in various embodiments of the present application, the size of the sequence number of each process described above does not mean the order of execution, and the execution order of each process should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0116] Those skilled in the art can clearly understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0117] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0118] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e. they can be located in one place or distributed on multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiment according to actual needs.

[0119] In addition, the functional units in each embodiment of the present application can be integrated into one processing unit, or each unit can be physically present separately, or two or more units can be integrated into one unit.

[0120] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any skilled person in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0121] The above only describes some exemplary embodiments of the present application by way of illustration, and it is needless to say that those skilled in the art can modify the described embodiments in various ways without deviating from the spirit and scope of the present application. Therefore, the above figures and descriptions are illustrative in nature and should not be understood as limiting the scope of protection of the claims of the present application.

Claims

1. A heavy-load eVTOL flight intelligent management and control system, characterized in that, The method comprises a ground effect airflow perception and modeling module, a lift sensitive section identification module, a lift-load coupling evaluation module, a risk driven control parameter configuration module, a thrust adjustment sequence generation module, and an attitude stabilization and safe landing control module. The ground effect airflow perception and modeling module establishes a multi-dimensional dynamic interference model of the airflow under the rotor and the terrain reflected airflow, adaptively corrects the airflow state based on the laser radar and sonar perception data, and generates a ground effect intensity distribution map. The lift sensitive section identification module calculates the lift change rate and amplitude in the flight trajectory according to the ground effect intensity distribution map, marks the critical height interval of abnormal growth of lift, and generates a lift sensitive section list. The lift-load coupling evaluation module fuses the lift sensitive section list and the current load state of the aircraft, calculates the load distribution deviation trend and attitude instability index in each section, and determines the coupling risk level of lift and load. The risk driven control parameter configuration module configures the response threshold of the flight control according to the coupling risk level, and constructs a real-time deviation correction curve of the lift predicted value and the actual feedback value. The thrust adjustment sequence generation module generates thrust control instructions based on the deviation correction curve and constructs a thrust progressive adjustment sequence. The attitude stabilization and safe landing control module executes the thrust progressive adjustment sequence and feeds it back to the rotor control module to smooth the thrust change slope and suppress the lift mutation impact.

2. The heavy-load eVTOL flight intelligent management and control system according to claim 1, characterized in that, The steps of establishing a multi-dimensional dynamic interference model of the airflow under the rotor and the terrain reflected airflow include: Constructing a three-dimensional airflow model based on rotor structural parameters, flight height and ground topography; Obtaining laser radar perception data and sonar perception data and performing synchronous preprocessing; Comparing the perception data with the airflow model, and adaptively correcting the model based on residual analysis and parameter updating algorithm; Generating a high-resolution ground effect intensity distribution map to describe the spatial distribution of airflow disturbance.

3. The heavy-load eVTOL flight intelligent management and control system according to claim 1, characterized in that, The steps of calculating the lift change and amplitude according to the ground effect intensity distribution map and generating a lift sensitive section list include: Matching the flight trajectory with the ground effect intensity distribution map to form a correspondence between the flight path and the ground effect intensity; Calculating the lift change rate and amplitude based on the matching results; Identifying the critical height interval of abnormal growth of lift according to the lift change rate and amplitude; Recording multiple critical height intervals, generating a lift sensitive section list and providing it for the flight control system to call.

4. The heavy-load eVTOL flight intelligent management and control system according to claim 1, characterized in that, The steps of fusing the lift sensitive section list with the current load state of the aircraft and constructing the coupling risk level of lift and load include: Obtaining the current load state data of the aircraft to establish a three-dimensional load distribution model; Synchronously fusing the load state data with the lift sensitive section list in time and position; Calculating the load distribution deviation trend and attitude instability index in each section; Based on the attitude instability index, the coupling risk level of lift and load is constructed and output to the flight control system.

5. The heavy-load eVTOL flight intelligent management and control system according to claim 1, characterized in that, The steps of constructing a real-time deviation correction curve of the lift predicted value and the actual feedback value include: Setting the response threshold in the flight control parameter according to the coupling risk level of lift and load; Establishing a lift prediction model and generating a predicted lift value based on the current flight state; Obtaining the current actual lift feedback value of the aircraft, and calculating the deviation between the predicted lift value and the actual lift feedback value; Based on the deviation, constructing a lift deviation correction curve and outputting a correction control instruction to the flight controller.

6. The heavy-load eVTOL flight intelligent management and control system according to claim 1, characterized in that, The step of generating a thrust control instruction based on the deviation correction curve and constructing a thrust gradual adjustment sequence includes: Analyzing the lift deviation correction curve to extract a thrust adjustment target parameter; Generating a thrust control output instruction based on the thrust adjustment target parameter; Decomposing the thrust control output instruction into continuous thrust adjustment values in multiple time steps to form a thrust gradual adjustment sequence; Outputting the thrust gradual adjustment sequence to the flight control module and executing multi-rotor cooperative control.

7. The heavy-load eVTOL flight intelligent management and control system according to claim 1, characterized in that, The specific steps of executing the thrust gradual adjustment sequence and feeding it back to the rotor control module to smooth the thrust change slope, stabilize the flight attitude, and suppress the lift mutation impact are as follows: According to the lift deviation correction curve, extract the total change amount required for thrust adjustment, denoted as ΔT, and combine the current attitude stability state and adjustment demand to calculate the nonlinear gradual thrust output instruction, the calculation expression is as follows: ， where T(t i ) is the thrust output value at the i-th time point t, T0 is the initial thrust value, a is the thrust adjustment amplitude factor, β is the thrust adjustment slope control parameter, N is the total number of time points, i is the time point index, ΔT is the thrust adjustment amplitude, and tanh is the hyperbolic tangent function. The generated thrust output value T(t i ) is distributed to each rotor, and the output proportion of each rotor is adjusted according to the current attitude state of the aircraft to realize attitude stabilization control. First, define the current attitude state vector of the aircraft as Θ(t), Θ(t) = [φ(t), θ(t), ψ(t)], where φ(t) is the roll angle at time point t, θ(t) is the pitch angle at time point t, and ψ(t) is the yaw angle at time point t. In order to evaluate the stability of the aircraft attitude, the attitude deviation cost index is calculated: C(Θ)=γ1·|φ(t)|+γ2·|θ(t)|+γ3·|ψ(t)|, Where C(Θ) is the attitude deviation cost index, representing the degree of deviation of the current overall attitude from the ideal horizontal attitude, γ1, γ2, γ3 are the weight coefficients of roll φ(t), pitch θ(t) and yaw ψ(t) respectively; In order to realize the attitude-thrust linkage control, based on the attitude deviation cost index, the distributed thrust correction is carried out for each rotor, and the adjustment formula is as follows: ， where T j (t i ) is the thrust output value of the jth rotor at time point t, T j is the thrust of the jth rotor, is the sensitivity of the attitude deviation cost to the jth rotor thrust.

Images