A probability demand-based eVTOL precise landing control parameter design method

Abstract

The application provides a probability demand-based eVTOL precise landing control parameter design method, including the following steps: based on a time domain or a frequency domain method, system identification is performed on the eVTOL to obtain dynamic characteristic parameters of the eVTOL under typical working conditions of landing; an eVTOL dynamics model is constructed according to the dynamic characteristic parameters, and an equivalent disturbance suppression input model of the eVTOL is constructed by using a system identification method; landing safety and performance requirements under a city operation scenario are set according to probability constraints; a controller is selected, parameters of the controller are taken as optimization variables for unified modeling, the optimization variables are converted into a control parameter optimization problem with probability constraints according to the equivalent disturbance suppression input model; and an optimization model of the parameters of the controller is established to maximize the anti-disturbance control precision in the eVTOL landing process under the premise of meeting the probability constraints, and the control parameter optimization problem is optimized to release the potential performance of the controller.

Description

Technical Field

[0001] This invention relates to the field of electric vertical takeoff and landing (eVTOL) technology, and in particular to a method for designing precise landing control parameters for eVTOL based on probabilistic requirements. Background Technology

[0002] For electric vertical takeoff and landing (eVTOL) aircraft designed for urban operations, eVTOLs typically need vertical takeoff and landing capabilities to adapt to the limited available space in cities. However, due to the spatial constraints of the urban environment, the area of ​​takeoff and landing sites is often limited, which places high demands on the positional accuracy of the eVTOL during the landing phase. On the other hand, in recent years, a type of split-type eVTOL has emerged, which consists of two parts: the flight body and the land-based body. The land-based body is similar to the chassis structure of a ground vehicle, and the flight body needs to land accurately on the land-based body during landing. This configuration further intensifies the requirements for landing accuracy.

[0003] To improve the positional accuracy of eVTOL during landing, two approaches are needed: firstly, improving the positioning accuracy of the navigation system, and secondly, enhancing the control accuracy of the flight control system. Control accuracy largely depends on the controller's architecture and parameter configuration.

[0004] Traditional flight controller parameter design methods typically use stability parameters such as gain margin and phase margin as design constraints, adhering to relevant aviation standards such as AS94900A. After meeting these stability margin requirements, designers further optimize system performance indicators, such as the ability to maintain position during landing. However, this design approach is often conservative. In reality, the stability margin required by an aircraft is not an immutable rigid parameter, but should be related to the degree of understanding of the uncertainties in aircraft dynamics. When the modeling of uncertainties in the system is relatively accurate, the stability margin constraints can be appropriately relaxed to improve performance. However, the specific degree to which this relaxation should be implemented usually relies on engineering experience and lacks systematic quantitative basis.

[0005] One common traditional approach to considering system uncertainty is robust control. This approach involves explicitly modeling the upper and lower limits of uncertainty and designing the controller based on worst-case scenarios. While robust control can significantly improve system stability, its assumptions about extreme conditions often lead to conservative design results, limiting the realization of the system's potential performance and failing to maximize the control system's performance potential. Summary of the Invention

[0006] To address the shortcomings of the existing technology, this invention provides a method for designing eVTOL precision landing control parameters based on probabilistic requirements, thereby solving the problems in the background technology.

[0007] This invention provides a method for designing eVTOL precision landing control parameters based on probabilistic requirements, comprising:

[0008] S1. Based on time-domain or frequency-domain methods, perform system identification of eVTOL and obtain the dynamic characteristic parameters of eVTOL under typical landing conditions;

[0009] S2. Construct an eVTOL dynamic model based on the dynamic characteristic parameters, and construct an equivalent disturbance suppression input model for eVTOL using the system identification method;

[0010] S3. Based on the landing safety and performance requirements in urban operation scenarios, set at least the following probability constraints: the probability of aircraft instability, the probability of actuator output exceeding the set threshold, and performance constraints.

[0011] S4. Select a controller and model the controller parameters as variables to be optimized in a unified manner. Based on the dynamic model and the equivalent disturbance suppression input model, transform the variables to be optimized into a control parameter optimization problem with probabilistic constraints.

[0012] S5. With the goal of maximizing the anti-disturbance control accuracy during the eVTOL landing process, and under the premise of satisfying the probability constraints, an optimization model for the controller parameters is established to optimize the control parameter optimization problem and release the potential performance of the controller.

[0013] Compared with the prior art, the beneficial effects of this invention are as follows:

[0014] This invention provides a probabilistic approach to designing precise landing control parameters for eVTOLs. This method no longer relies on frequency domain indicators (such as gain margin and phase margin), but instead sets at least the following probabilistic constraints based on the landing safety and performance requirements of urban operation scenarios: the probability of aircraft instability, the probability of actuator output exceeding a set threshold, and performance constraints. The optimization objective is to maximize the anti-disturbance control accuracy during eVTOL landing. Under the premise of satisfying these probabilistic constraints, an optimization model of the controller parameters is established, directly constraining the probability of eVTOL instability. This allows for maximizing the performance potential of the control system while ensuring flight safety. Furthermore, regarding the modeling problem of control parameter optimization with probabilistic constraints, this invention combines the system identification results of actual eVTOLs to probabilistically model parameter uncertainties, making the design method closer to the characteristics of real systems. This method not only improves the rationality and effectiveness of control parameter design but also significantly enhances the position control accuracy of eVTOLs during landing, making it particularly suitable for urban operation scenarios with extremely high landing accuracy requirements. Attached Figure Description

[0015] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. Some specific embodiments of the invention will be described in detail below with reference to the accompanying drawings in an exemplary and non-limiting manner. The same reference numerals in the drawings designate the same or similar parts or components. It should be understood by those skilled in the art that these drawings are not necessarily drawn to scale. In the drawings:

[0016] Figure 1 This is a flowchart illustrating a method for designing eVTOL precision landing control parameters based on probabilistic requirements according to an embodiment of the present invention.

[0017] Figure 2 yes Figure 1 This is a flowchart illustrating step S6 of a method for designing eVTOL precision landing control parameters based on probabilistic requirements according to an embodiment of the present invention.

[0018] Figure 3 This is a flowchart illustrating step S7 of a method for designing eVTOL precision landing control parameters based on probability requirements according to an embodiment of the present invention.

[0019] Figure 4 This is a time-domain diagram of an eVTOL precision landing control parameter design method based on probability requirements according to an embodiment of the present invention;

[0020] Figure 5 This is an optimized time-domain comparison diagram of an eVTOL precision landing control parameter design method based on probability requirements according to an embodiment of the present invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of this embodiment will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0022] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of this application are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0023] Furthermore, the use of terms such as "first," "second," etc., in this application is for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0024] See Figure 1 This embodiment provides a method for designing eVTOL precision landing control parameters based on probabilistic requirements, including the following steps:

[0025] S1. Based on time-domain or frequency-domain methods, perform system identification of eVTOL and obtain the dynamic characteristic parameters of eVTOL under typical landing conditions;

[0026] S2. Construct an eVTOL dynamic model based on the dynamic characteristic parameters, and construct an equivalent disturbance suppression input model for eVTOL using the system identification method;

[0027] S3. Based on the landing safety and performance requirements in urban operation scenarios, set at least the following probability constraints: the probability of eVTOL instability, the probability of actuator output exceeding the set threshold, and performance constraints.

[0028] S4. Select a controller and model the controller parameters as variables to be optimized in a unified manner. Based on the dynamic model and the equivalent disturbance suppression input model, transform the variables to be optimized into a control parameter optimization problem with probabilistic constraints.

[0029] S5. With the goal of maximizing the anti-disturbance control accuracy during the eVTOL landing process, and under the premise of satisfying the probability constraints, an optimization model for the controller parameters is established to optimize the control parameter optimization problem and release the potential performance of the controller.

[0030] It should be noted that, unless otherwise specified, the designations such as S1, S2, and S3 mentioned above are only for the purpose of making it easier to understand the relationship between the method steps and for the convenience of marking in the diagram, and do not uniquely limit the order of the steps.

[0031] It should be noted that various uncertainties and disturbances are unavoidable in actual eVTOL flight, which may jeopardize eVTOL safety. To ensure flight safety, the eVTOL controller must meet safety requirements, even in the presence of uncertainties and disturbances. Some safety requirements are defined as failure probability thresholds that the system must not exceed. For example, during automatic landing, the probability of exceeding safety limits must not exceed a specified threshold.

[0032] It should be noted that the method provided in this embodiment addresses the technical problem in existing technologies that cannot maximize the performance of eVTOL controllers during landing. This invention provides a probabilistic approach to designing precise landing control parameters for eVTOLs. This method no longer relies on frequency domain indicators (such as gain margin and phase margin), but instead sets at least the following probabilistic constraints based on the landing safety and performance requirements of urban operation scenarios: the probability of aircraft instability, the probability of actuator output exceeding a set threshold, and performance constraints. The optimization objective is to maximize the anti-disturbance control accuracy during eVTOL landing. Under the premise of satisfying the probabilistic constraints, an optimization model of the controller parameters is established, directly constraining the probability of eVTOL instability. This allows for maximizing the performance potential of the control system while ensuring flight safety. Furthermore, regarding the modeling problem of control parameter optimization with probabilistic constraints, this invention combines the system identification results of actual eVTOLs to probabilistically model parameter uncertainties, making the design method closer to the characteristics of real systems. This method not only improves the rationality and effectiveness of control parameter design but also significantly enhances the position control accuracy of eVTOLs during landing, making it particularly suitable for urban operation scenarios with extremely high landing accuracy requirements.

[0033] Preferably, step S1 includes: obtaining the dynamic characteristic parameters of the eVTOL through frequency domain or time domain system identification methods, and performing probability distribution modeling on the dynamic uncertainty parameters of the eVTOL, wherein the probability distribution modeling is determined based on the actual system identification results;

[0034] The dynamic characteristic parameters include the mass, moment of inertia, and aerodynamic derivative of the eVTOL under landing conditions; the parameter uncertainties include aircraft mass deviation, aerodynamic parameter perturbations, and actuator delays; and the probability distribution types include normal distribution, uniform distribution, or empirical distribution based on measured data.

[0035] It should be noted that in traditional control parameter design, dynamic uncertainties (such as mass deviation and aerodynamic parameter errors) are often assumed to be "fixed intervals" or "standard probability distributions." However, the uncertainties of actual systems often have complex characteristics (such as skewed distributions and multimodal distributions), which differ significantly from the assumed models. By identifying the system and obtaining measured data of the dynamic characteristic parameters of the eVTOL under real operating conditions (such as different loads, wind speeds, and temperatures), such as the range of mass fluctuations and the dispersion of aerodynamic derivatives, and selecting the probability distribution type based on the statistical laws of the data, the uncertainty modeling can be made closer to the essence of objective reality.

[0036] In addition, step S3 requires calculating core constraints such as instability probability and actuator saturation probability based on the probability distribution of uncertainty parameters. If the distribution model is inaccurate, it will directly lead to the failure of the constraint calculation results, such as underestimating the risk or being overly conservative.

[0037] Traditional robust control methods design controllers based on "worst-case conditions" (such as parameters taking extreme values), which leads to excessive stabilization of control parameters and sacrifices dynamic performance. In contrast, this method uses probabilistic modeling driven by actual data to accurately quantify "low-probability extreme events" and avoids performance suppression for medium- to high-probability conditions.

[0038] The urban eVTOL operating environment is complex (e.g., wind shear from high-rise buildings, ground effect, temperature changes), and dynamic parameters may change dynamically with operating conditions, making it difficult for fixed distribution models to cover the entire scenario. This method identifies separable scenarios (e.g., hovering, low-speed descent, strong wind interference) in actual systems and establishes an uncertainty probability model under multiple operating conditions, enabling controller parameter optimization to adapt to the parameter characteristics of different scenarios.

[0039] Preferably, in an urban operating environment, the probability constraint in step S3 includes: the probability of the eVTOL becoming unstable must not exceed 10⁻ 6 The instability probability constraint, identified through a linearized model using time-domain or frequency-domain methods, ensures a safety baseline under extreme conditions; the probability of the actuator output exceeding a set threshold must not exceed 10⁻⁻⁴. 4 The actuator over-limit constraints are identified by a linearized model using time-domain or frequency-domain methods; the performance constraints include at least one of error, energy consumption, and handling stability. The error, energy consumption, and stability constraints are transformed into optimizable quantitative indicators through CETI, enabling high-precision landing and low-loss operation in urban scenarios.

[0040] It should be noted that, in an urban operating environment, the probability of the eVTOL becoming unstable can be 10⁻ 6 Depending on security requirements, it can also be set to less than 10⁻ 6 However, in non-urban operating environments, specifically in remote areas, the probability of an eVTOL experiencing instability can be as high as 10⁻⁻⁶. 4 .

[0041] It should be noted that by setting probabilistic constraints, a precise balance between high security and robustness against disturbances is achieved. This technical approach enables eVTOL to meet safety requirements while fully unleashing the system's performance potential in complex urban environments, making it particularly suitable for scenarios with stringent requirements for both accuracy and robustness, such as split-type eVTOL systems.

[0042] Preferably, step S4 includes: selecting a controller structure and using the controller's parameters as variables to be optimized. The controller structure includes at least one of a PID controller, a linear extended state observer, or a nonlinear dynamic inverse controller. It should be noted that parameter optimization ensures that probabilistic constraints (instability, saturation) are strictly satisfied, making the risk quantifiable and traceable; simultaneously, it unlocks the system potential of the controller, improves position control accuracy, and reduces energy consumption.

[0043] Preferably, step S5 includes: the disturbance rejection control accuracy is evaluated through the equivalent disturbance rejection input model, which is constructed using a system identification method to quantify the eVTOL's ability to suppress external disturbances during the landing phase. It should be noted that traditional disturbance rejection performance evaluations often rely on "time-domain response curves" (such as overshoot and settling time under step disturbances) or "frequency-domain margins" (such as disturbance rejection bandwidth), but these indicators cannot directly correlate the quantitative relationship between actual disturbance scenarios (such as crosswinds and ground effects) and control accuracy. The equivalent disturbance rejection input model (CETI) constructs a transmission relationship of "external disturbance - equivalent control input - position error" through system identification, transforming disturbance rejection capability into a calculable quantitative indicator. This avoids the traditional qualitative description of "strong / weak disturbance rejection," making disturbance rejection performance measurable and comparable (such as comparing disturbance rejection rates under different controller parameters), providing a clear quantitative target for control parameter optimization. The CETI model is not an ideal model derived from theory, but rather constructed based on real eVTOL flight data using frequency or time domain system identification methods. Therefore, it accurately reflects the disturbance characteristics of the actual system. The CETI model directly correlates controller parameters, disturbance suppression effectiveness, and position error, providing a clear objective function and constraint basis for optimization algorithms, avoiding the blindness of traditional "empirical parameter tuning." Urban eVTOLs face various external disturbances during the landing phase, and different disturbances have varying degrees of impact on control accuracy. The CETI model can quantify the contribution weight of each disturbance, enabling differentiated design that prioritizes the suppression of key disturbances. The construction of the CETI model does not depend on a specific controller type and can serve as a unified disturbance rejection performance evaluation tool, solving the problem of incomparable disturbance rejection effects between different controllers. The CETI model provides support for the evaluation of disturbance rejection control accuracy during the eVTOL landing phase.

[0044] Preferably, the anti-disturbance control accuracy of the optimization target is achieved by minimizing the variance of the position holding error during the landing process.

[0045] It should be noted that minimizing the position holding error variance as an optimization objective is essentially about improving the stability and consistency of disturbance rejection control accuracy through statistical characteristic optimization: it avoids the traditional average error from masking the risk of fluctuations, and ensures the safety boundary through collaboration with probabilistic constraints. Ultimately, in the high-precision landing scenario of urban eVTOL (especially split-type configurations), it achieves engineering value with small error fluctuations and low extreme risks.

[0046] Preferably, the eVTOL precise landing control parameter design method based on probability requirements further includes step S6: using Monte Carlo sampling, importance sampling, or surrogate model methods to efficiently estimate the constraint probability, and employing at least one solution strategy from particle swarm optimization, evolutionary algorithm, or Bayesian optimization to complete the global optimization search for the control parameter optimization problem. It should be noted that through the collaborative design of efficient probability estimation and intelligent optimization algorithms, the core problems of high computational cost, inaccurate estimation of low-probability events, and difficulty in finding the global optimum in traditional control parameter design are solved.

[0047] Preferably, the design method for eVTOL precise landing control parameters based on probabilistic requirements further includes step S7: applying the optimized controller parameters to a nonlinear eVTOL model with uncertainty, performing closed-loop control simulation, and evaluating the anti-disturbance performance of the eVTOL under typical disturbance scenarios and the satisfaction of probabilistic requirements; if the evaluation results do not meet the design requirements, the process returns to step S6 for iterative optimization.

[0048] Preferably, the surrogate model is a radial basis function neural network, which achieves efficient estimation of constraint probabilities after training with sample points, and the sample points are randomly sampled from the eVTOL dynamic model.

[0049] Preferably, the typical interference scenarios include crosswind disturbance, ground effect, and sensor noise, and the simulation verification needs to cover extreme interference conditions in urban operation scenarios.

[0050] In some specific implementations, in step S1, time-domain identification involves acquiring time-domain response data of the eVTOL during descent, including attitude angles (such as pitch and roll angles), position (such as x / y / z axis coordinates), and angular velocity, by applying a step signal, sinusoidal sweep signal, or random noise signal. The parameters of the dynamic equations (such as mass, moment of inertia, and aerodynamic derivatives) are then fitted using least squares, Kalman filtering, or neural networks. In other specific embodiments of step S1, the time-domain data is converted into frequency-domain characteristics using Fourier transform to obtain amplitude-frequency response curves and phase-frequency response curves. The parameters of the transfer function model are estimated using nonlinear least squares fitting. Based on the time-domain or frequency-domain identification methods, the dynamic characteristics of the hovering and near-ground descent phases are identified.

[0051] In step S2, an eVTOL dynamic model is constructed based on the dynamic characteristic parameters identified in step S1. An equivalent disturbance suppression input model for eVTOL is constructed using the system identification method to quantify the impact of external disturbances (crosswinds, ground effects) on the position error.

[0052] In step S3, regarding the instability probability constraint, based on the urban eVTOL safety requirements, the instability probability P is set to ≤ 10⁻ 6 The random distribution of dynamic parameters is simulated using Monte Carlo sampling to calculate the probability that the closed-loop system poles fall into the right half-plane. For actuator saturation probability constraints, considering motor output limitations (such as maximum speed and current threshold), the probability that the actuator output exceeds the threshold is set to P ≤ 10⁻. 4 For example, if the maximum torque of the motor is 20 N·m, the probability of torque exceeding the limit is calculated through probabilistic modeling. For uncertainty probability modeling, based on the identification results of step S1, dynamic parameters (such as mass and moment of inertia) are treated as random variables, assuming that they follow a normal distribution, or an empirical distribution is adopted based on measured data.

[0053] In step S4, the parameters of the controller are modeled as variables to be optimized. Based on the equivalent disturbance suppression input model, the variables to be optimized are transformed into a control parameter optimization problem with probabilistic constraints, thus constructing a control parameter optimization problem with probabilistic constraints.

[0054] In step S5, with the optimization objective of maximizing the anti-disturbance control accuracy during the eVTOL landing process, and under the premise of satisfying the probability constraints, an optimization model for the controller parameters is established to optimize the controller parameter optimization problem and release the potential performance of the controller.

[0055] Taking a certain eVTOL aircraft as an example, its initial control parameters follow the requirements of AS94900, with sufficient stability margins: a magnitude margin of 6 dB and a phase margin of 45 degrees. Taking its roll attitude disturbance rejection as an example, its performance in suppressing roll equivalent disturbances is as follows: The corresponding time domain is as follows Figure 4 As shown in the figure. Among them, "phi Time Domain" indicates the time domain, which means that this is the curve of the phase (phi) changing in the time domain, that is, the dynamic trend of phase change over time; the horizontal axis represents time, and the vertical axis represents the phase angle, mainly reflecting the fluctuation of the phase angle over time.

[0056] If the goal is to achieve the best possible disturbance rejection capability while maintaining sufficient stability, for example, ensuring that the instability probability is no higher than 1e-6, then the optimization problem for S4 is: in To achieve the optimization objective, namely, a roll angle of 3 times the RMS value for disturbance suppression, the probability constraint that needs to be satisfied is that the instability probability is less than 10 to the power of -6. "Subject to" means subject to.

[0057] Thus, the S4 optimization problem with probabilistic constraints was constructed.

[0058] For S5, the control parameters to be optimized include the proportional term of the roll angle control, the proportional and integral terms of the angular velocity, and the zeros and poles of the lead-lag filter. The control parameters can be optimized by using constrained optimization algorithms, such as patternsearch.

[0059] The final calculation result shows that the instability probability exactly matches the constraint boundary: At this time The performance is more than 5 times better than before optimization. For example... Figure 5 As shown in the figure, there are two curves (blue and yellow-green) representing phase comparison data under different conditions or objects. The blue wavy line represents the effect before optimization, and the yellow-green wavy line represents the effect after optimization.

[0060] In step S6, the constraint probability is efficiently estimated using Monte Carlo sampling, importance sampling, or surrogate model methods. At least one solution strategy, such as particle swarm optimization, evolutionary algorithm, or Bayesian optimization, is used to complete the global optimization search for the control parameter optimization problem.

[0061] In step S7, the optimized controller parameters are applied to the nonlinear eVTOL model with uncertainty to perform closed-loop control simulation and evaluate the eVTOL's disturbance rejection performance and its ability to meet probabilistic requirements under typical disturbance scenarios. If the evaluation results do not meet the design requirements, such as the instability probability still being higher than 10⁻ 6 If the condition is not met, return to step S6 for iterative optimization until the target is met.

[0062] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A probabilistic demand-based eVTOL precision landing control parameter design method, characterized in that, Comprise: S1, based on time domain or frequency domain method, system identification is carried out to eVTOL, and the dynamic characteristic parameters of eVTOL under the typical working condition of landing are acquired; S2, a dynamics model of eVTOL is constructed according to the dynamic characteristic parameters, and an equivalent disturbance suppression input model of eVTOL is constructed by using a system identification method; S3, according to the landing safety and performance requirements under the urban operation scene, at least the following probability constraints are set: the probability of instability of the aircraft, the probability of actuator output exceeding the set threshold and the performance constraint; S4, a controller is selected, the parameters of the controller are taken as optimization variables for unified modeling, the optimization variables are converted into a control parameter optimization problem with probability constraints according to the dynamics model and the equivalent disturbance suppression input model; S5, the optimization model of the parameters of the controller is established to maximize the anti-disturbance control accuracy in the landing process of eVTOL, and the global optimization search of the control parameter optimization problem is completed by using at least one of particle swarm optimization, evolutionary algorithm and Bayesian optimization.

2. The probabilistic demand based eVTOL precision landing control parameter design method of claim 1, wherein: The step S1 comprises The dynamic characteristic parameters of eVTOL are acquired by a frequency domain or time domain system identification method, and the probability distribution modeling of the dynamic uncertainty parameters of eVTOL is carried out, wherein the probability distribution modeling is determined based on the actual system identification result; Wherein, the dynamic characteristic parameters include the mass, the moment of inertia and the aerodynamic derivative of eVTOL under the landing working condition;The parameter uncertainty includes the mass deviation of the aircraft, the aerodynamic parameter perturbation and the actuator delay, and the probability distribution type includes normal distribution, uniform distribution or empirical distribution based on measured data.

3. The probabilistic demand based eVTOL precision landing control parameter design method of claim 1, wherein, The probability constraints in the step S3 include: The probability of the eVTOL becoming unstable must not exceed 10⁻ 6 ; The probability that the actuator output exceeds a set threshold shall not exceed 10⁻ 4 ; The performance constraints include at least one of error, energy consumption and steering stability.

4. The probabilistic demand based eVTOL precision landing control parameter design method of claim 1, wherein, The step S4 comprises: The controller structure is selected, and the parameters of the controller are taken as optimization variables, wherein the controller structure includes at least one of PID controller, linear extended state observer or nonlinear dynamic inverse controller.

5. The probabilistic demand based eVTOL precision landing control parameter design method of claim 1, wherein, The step S5 comprises: The anti-disturbance control accuracy is evaluated by the equivalent disturbance suppression input model, and the equivalent disturbance suppression input model is constructed by a system identification method, which is used to quantify the suppression ability of eVTOL to external disturbance in the landing stage.

6. The probability demand based eVTOL precise landing control parameter design method according to claim 5, wherein: The anti-disturbance control accuracy of the optimization target is realized by minimizing the position keeping error variance in the landing process.

7. The probabilistic demand based eVTOL precision landing control parameter design method of claim 6, wherein, Further comprise: S6, the constraint probability is efficiently estimated by using Monte Carlo sampling, importance sampling or proxy model method, and at least one of particle swarm optimization, evolutionary algorithm and Bayesian optimization is used as a solving strategy to complete the global optimization search of the control parameter optimization problem.

8. The probabilistic demand based eVTOL precision landing control parameter design method of claim 7, wherein, Further comprise: S7, the controller parameters obtained by optimization are applied to the nonlinear eVTOL model with uncertainty for closed-loop control simulation, the anti-disturbance performance of eVTOL under the typical interference scene and the satisfaction of the probability demand are evaluated, and if the evaluation result does not meet the design requirement, the step S6 is returned for iterative optimization.

9. The probabilistic demand based eVTOL precision landing control parameter design method of claim 8, wherein, The step S6 comprises: The proxy model is a radial basis function neural network, which realizes efficient estimation of the constraint probability after being trained by sample points derived from random sampling of the eVTOL dynamics model.

10. The probabilistic demand based eVTOL precision landing control parameter design method of claim 9, wherein, The step S7 comprises: The typical interference scenarios include side wind disturbance, ground effect and sensor noise, and the simulation verification needs to cover extreme interference conditions in the urban operation scenario.

Images