A manned flying saucer attitude control method and system

Abstract

The present application belongs to the field of aircraft control technology, and discloses a manned flying saucer attitude control method and system. Through the steps of detection, identification, control mode determination and optimization solution, it effectively copes with the response hysteresis or jamming failure of the power unit and maintains high-precision attitude control of the manned flying saucer. Thus, it has the advantages of being able to effectively cope with the response hysteresis or jamming failure of the power unit and maintain high-precision attitude control of the manned flying saucer.

Description

Technical Field

[0001] The present application relates to the field of aircraft control technology, and in particular to a method and system for controlling the attitude of a manned flying saucer. Background Art

[0002] As an advanced aircraft, manned flying saucers (UFOs) demonstrate unique advantages in performing close-range, precision work tasks, such as aerial inspections of bridge structures. For example, Chinese patent CN214986060U discloses a UFO-shaped manned aircraft comprising multiple propeller-type power units. The UFO's flight control relies on its attitude control system, which coordinates the thrust output of multiple distributed power units to generate the required net force and torque, thereby achieving precise control of the UFO's position and attitude. Based on the UFO's current state data acquired by sensors and preset mission instructions, the control system calculates and sends thrust commands to each power unit in real time. Precision work tasks place extremely high demands on the UFO's control accuracy, placing stringent demands on the thrust control precision and response speed of the power units. To meet these high-precision and fast-response requirements, UFOs typically utilize power units with high power density and fast response characteristics. However, in actual operation, these high-performance power units often exhibit significant nonlinearities in their thrust output characteristics, such as thrust dead zones, hysteresis, saturation, or efficiency degradation. In addition, there is a certain response delay between the power unit receiving the thrust command and the actual generation of stable thrust, and the rate of thrust change is subject to physical limitations. These complex nonlinear characteristics and dynamic limitations make the precise thrust control of a single power unit complicated, thereby increasing the difficulty of controlling the attitude and position of the entire flying saucer.

[0003] During the long, high-load, and delicate operations of a UFO, a power unit may fail due to environmental factors, component fatigue, or aging. A common failure mode manifests as a delayed or stuck response of the unit's control actuator. This failure mode differs from complete thrust loss or reduced thrust in that the thrust output becomes unreliable and difficult to predict. The actual output value may fluctuate randomly within a certain range, and the response to control commands is subject to significant uncertainty. Because the thrust output of a power unit experiencing delayed or stuck response is unreliable and subject to random deviations and delays, existing thrust allocation algorithms designed based on ideal power unit models are unable to effectively compensate for this effect. These algorithms typically assume that the power units have good linear characteristics and fast response. Continuing to use the existing algorithm to command the remaining functioning power units would not only make it difficult to maintain the UFO's stability within the high-precision spatial domain, but the unpredictable thrust output of the failed unit would also continuously disrupt the coordinated control of the remaining units.

[0004] Traditional fault reconstruction methods can usually only restore the basic controllability of the aircraft, but it is difficult to achieve the high precision level required for delicate operations. In addition, the response hysteresis or jamming of the faulty unit may cause its thrust output to vibrate or be uneven. During the thrust reconstruction process, if the remaining units make drastic thrust adjustments for rapid compensation, this unevenness or vibration may be transmitted to the entire flying saucer structure, reducing passenger comfort. Therefore, the thrust reconstruction strategy also needs to strike a balance between quickly restoring high-precision control and maintaining flight smoothness, and try to avoid introducing new discomfort. This requires the reconstruction algorithm to consider not only torque balance, but also thrust change rate and acceleration when calculating thrust distribution to achieve a smooth transition.

[0005] In view of the above problems, the existing technology is in urgent need of improvement. Summary of the Invention

[0006] The purpose of this application is to provide a manned flying saucer attitude control method and system, which can effectively deal with the response delay or jamming failure of the power unit and maintain high-precision attitude control of the manned flying saucer.

[0007] In a first aspect, the present application provides a method for controlling the attitude of a manned flying saucer, which is applied to a manned flying saucer having multiple power units. The method comprises the following steps:

[0008] A1. Monitor the thrust command and actual thrust output of each power unit to determine whether any power unit has a response delay or a stuck fault.

[0009] A2. If a faulty unit exists, the thrust output characteristics of the faulty unit are identified based on the relevant data stream of the faulty unit, and the thrust information of the faulty unit is obtained. The relevant data stream includes the data stream of the thrust command and actual thrust output of the faulty unit, as well as the data stream of the manned flying saucer's motion status data.

[0010] A3. Determine the control mode for the faulty unit based on the thrust information of the faulty unit;

[0011] A4. Extract the dynamic constraint models and operational constraint information of the remaining normal power units from the local database. The dynamic constraint models include mathematical models of the normal power unit's thrust output characteristics, response delay, and physical coupling effects. The operational constraint information includes the reference thrust range and reference thrust change rate range of the normal power unit.

[0012] A5. With the goal of minimizing the attitude tracking error, position tracking error, and thrust change rate of the manned vehicle, solve the thrust command for each normal power unit based on the vehicle's real-time attitude, real-time position, desired attitude, and desired position, as well as the thrust information and control mode of the faulty unit, the operational constraint information of the normal power unit, and the dynamic constraint model of the normal power unit.

[0013] A6. Send the calculated thrust instructions of each normal power unit to the corresponding power unit actuator.

[0014] In a second aspect, the present application provides a manned flying saucer attitude control system, which is applied to a manned flying saucer having multiple power units, and the system includes:

[0015] A fault detection module is used to monitor the thrust command and actual thrust output of each power unit and determine whether there is a faulty unit with response hysteresis or stuck fault in each power unit;

[0016] A thrust characteristic identification module is used to identify the thrust output characteristics of a faulty unit based on the relevant data stream of the faulty unit when a faulty unit exists, and obtain thrust information of the faulty unit; the relevant data stream includes the data stream of the thrust command and actual thrust output of the faulty unit, and the data stream of the motion state data of the manned flying saucer;

[0017] A control mode determination module, configured to determine a control mode for the faulty unit based on thrust information of the faulty unit;

[0018] a data extraction module, configured to extract the dynamic constraint model and operational constraint information of the remaining normal power units from a local database; the dynamic constraint model includes a mathematical model of the thrust output characteristics, response delay, and physical coupling effects of the normal power units; and the operational constraint information includes a reference thrust range and a reference thrust change rate range of the normal power units;

[0019] A thrust optimization module is used to minimize the attitude tracking error, position tracking error and thrust change rate of the manned flying saucer, and solve the thrust command of each normal power unit based on the real-time attitude, real-time position, expected attitude and expected position of the manned flying saucer, as well as the thrust information of the faulty unit, the control mode of the faulty unit, the operation constraint information of the normal power unit and the dynamic constraint model of the normal power unit;

[0020] The thrust execution module is used to send the calculated thrust instructions of each normal power unit to the corresponding power unit execution mechanism.

[0021] Beneficial effects: The present application provides a manned flying saucer attitude control method and system, which effectively copes with the response hysteresis or jamming failure of the power unit through the steps of detection, identification, control mode determination and optimization solution, and maintains high-precision attitude control of the manned flying saucer. Thus, it has the advantage of being able to effectively cope with the response hysteresis or jamming failure of the power unit and maintain high-precision attitude control of the manned flying saucer. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a flowchart of the manned flying saucer attitude control method provided in an embodiment of the present application.

[0023] Figure 2 This is a schematic structural diagram of the manned flying saucer attitude control system provided in an embodiment of the present application.

[0024] Explanation of reference numerals: 1. Fault detection module; 2. Thrust characteristic identification module; 3. Control mode determination module; 4. Data extraction module; 5. Thrust optimization module; 6. Thrust execution module. DETAILED DESCRIPTION

[0025] The technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. The components of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for protection, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work fall within the scope of protection of this application.

[0026] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0027] refer to Figure 1 This application proposes a manned flying saucer attitude control method, which is applied to a manned flying saucer with multiple power units. The method comprises the following steps:

[0028] A1. Monitor the thrust command and actual thrust output of each power unit to determine whether any power unit has a response delay or a stuck fault.

[0029] A2. If a faulty unit exists, the thrust output characteristics of the faulty unit are identified based on the relevant data stream of the faulty unit, and the thrust information of the faulty unit is obtained. The relevant data stream includes the data stream of the thrust command and actual thrust output of the faulty unit, as well as the data stream of the manned flying saucer's motion status data.

[0030] A3. Determine the control mode for the faulty unit based on the thrust information of the faulty unit;

[0031] A4. Extract the dynamic constraint models and operational constraint information of the remaining normal power units from the local database. The dynamic constraint models include mathematical models of the normal power unit's thrust output characteristics, response delay, and physical coupling effects. The operational constraint information includes the reference thrust range and reference thrust change rate range of the normal power unit.

[0032] A5. With the goal of minimizing the attitude tracking error, position tracking error, and thrust change rate of the manned vehicle, solve the thrust command for each normal power unit based on the vehicle's real-time attitude, real-time position, desired attitude, and desired position, as well as the thrust information and control mode of the faulty unit, the operational constraint information of the normal power unit, and the dynamic constraint model of the normal power unit.

[0033] A6. Send the calculated thrust instructions of each normal power unit to the corresponding power unit actuator.

[0034] Among them, the power unit refers to the device that provides thrust for the manned flying saucer, which can be implemented in the form of propellers, jet engines, etc. Its main purpose is to generate the force required to make the flying saucer move or maintain its posture.

[0035] Among them, response hysteresis or stuck failure refers to an abnormal state in which the actual thrust output of the power unit is delayed relative to the thrust command or cannot change with the command. It is mainly used to describe a specific failure mode of the power unit.

[0036] Among them, the relevant data streams include the data streams of the thrust instructions and actual thrust output of the power unit, as well as the data streams of the motion status data of the manned flying saucer, which are mainly used to provide real-time information for fault diagnosis and characteristic identification.

[0037] Among them, the thrust output characteristic refers to the relationship between the thrust command of the power unit and the actual thrust output, including dynamic characteristics such as nonlinearity and delay, which is mainly used to describe the actual working performance of the power unit.

[0038] The thrust information refers to the thrust output characteristic data of the identified faulty unit, which is mainly used to quantify the current behavior of the faulty unit.

[0039] The control mode refers to the processing strategy adopted for the faulty unit, such as deactivation or compensation, which is mainly to deal with different states of the faulty unit.

[0040] Among them, the local database is used to store the dynamic constraint model and operational constraint information of the normal power unit, which is mainly used to provide the system parameters and restrictions required for thrust redistribution. The dynamic constraint model is a mathematical model that describes the thrust output characteristics, response delay and physical coupling effects of the normal power unit. It is mainly used to accurately characterize the dynamic response of the normal unit. The operational constraint information includes the reference thrust range and reference thrust change rate range of the normal power unit, which is mainly used to limit the safe working boundary of the normal unit. The dynamic constraint model and operational constraint information of each power unit can be determined in advance through experiments and recorded in the local database.

[0041] Among them, attitude tracking error and position tracking error are the deviations between the actual attitude and position of the manned aircraft and the desired attitude and position, and are primarily used to measure the performance of the control system. Thrust rate of change is the rate of change of the power unit thrust command over time, and is primarily used to reflect the smoothness of thrust adjustment. The comprehensive objective function is a mathematical combination of control objectives such as attitude tracking error, position tracking error, and thrust rate of change. It is primarily used to balance multiple control objectives during the optimization process.

[0042] Among them, the quadratic programming method is an optimization algorithm used to solve optimization problems where the objective function is a quadratic function and the constraints are linear functions. Its main purpose is to calculate the optimal thrust command that meets the constraints.

[0043] The core innovation of this application lies in monitoring the power unit status in real time and identifying response delays or stuck faults, determining its control mode according to the real-time thrust output characteristics of the faulty unit, and combining the detailed dynamics and operational constraint model of the normal power unit, using optimization methods to solve the thrust instructions of the normal unit, so that when a specific fault occurs in the power unit, the attitude and position control capabilities of the flying saucer can be restored quickly and accurately.

[0044] Specifically, this method first continuously monitors the thrust commands received by each power unit and the actual thrust output generated. By comparing the commands and outputs, it determines whether any faulty units are experiencing response delays or stuck conditions. Once a faulty unit is detected, the system uses relevant data streams, including the faulty unit's thrust commands, actual thrust output, and the UFO's motion state data, to identify the faulty unit's current thrust output characteristics, such as thrust scaling coefficient, response delay, and stuck threshold, in real time. This information is then used to determine the appropriate control mode for the faulty unit, such as complete deactivation or hysteresis compensation. Simultaneously, the dynamic constraint models and operational constraint information for the remaining functioning power units are extracted from a pre-stored local database. This information details the thrust output characteristics, response delays, physical coupling effects, and thrust range and rate of change limitations of these units. An optimization problem is then constructed to minimize the UFO's attitude tracking error, position tracking error, and thrust rate of change. This optimization problem considers the current real-time attitude and position of the flying saucer, as well as the desired attitude and position. It also applies the thrust effect of the faulty unit (determined by its thrust information and control mode) as an external disturbance force or torque to the flying saucer's dynamic model. By combining the dynamic constraint model and operational constraint information of the normal power units as constraints, the problem is solved using optimization methods such as quadratic programming, resulting in thrust commands for each normal power unit that minimize the objective function. Finally, the calculated thrust commands are sent to the corresponding normal power unit actuators, driving them to generate thrust, thereby achieving precise control of the flying saucer's attitude and position and compensating for the effects of the faulty unit.

[0045] Through the above scheme, the present application can quickly detect and accurately identify the response hysteresis or stuck faults of the power unit and their characteristics, and adaptively adjust the control strategy according to the fault characteristics, incorporating the impact of the faulty unit into the overall control optimization. This method utilizes the detailed dynamic model and operating constraints of the normal power unit, and through optimization solution, it can fully utilize the capabilities of the remaining power unit, while compensating for the impact of the fault, achieving accurate tracking of the attitude and position of the flying saucer, and taking into account the smoothness of thrust adjustment. This solves the problem of the difficulty of quickly and accurately restoring the attitude and position control capability of the manned flying saucer power unit when the response hysteresis or stuck fault occurs.

[0046] In some embodiments, step A1 comprises:

[0047] A101. Periodically obtain the thrust command and actual thrust output of each power unit at a preset frequency;

[0048] A102. For each power unit, calculate the absolute value of the deviation between the actual thrust output and the thrust command in the current cycle, recording this as the first deviation.

[0049] A103. If the first deviation exceeds the preset deviation threshold, the cumulative duration of the first deviation exceeding the preset deviation threshold;

[0050] A104. If the duration exceeds a preset delay time threshold, the corresponding power unit is determined to be a faulty unit with a response delay fault;

[0051] A105. For each power unit, monitor whether the actual thrust output of the power unit changes within a preset observation time after receiving a new thrust command. If no change occurs, determine that the corresponding power unit is a faulty unit with a stuck fault.

[0052] The preset frequency can be set based on actual needs. The thrust command is the signal sent by the control system to the power unit actuator indicating the desired thrust level. The actual thrust output is the actual thrust generated by the power unit, typically measured by a sensor. The absolute value of the deviation is the non-negative difference between the actual thrust output and the thrust command, quantifying the degree to which the actual output deviates from the desired command.

[0053] The preset hysteresis time threshold can be set according to actual needs. The preset observation time refers to a preset time window during which the system waits and observes changes in the actual thrust output of the power unit after receiving a new thrust command. The value of the preset observation time can be set according to actual needs.

[0054] The fault detection scheme of this application effectively detects power unit response delays and stuck faults through a series of coordinated steps. First, the system periodically obtains the thrust command and actual thrust output of each power unit at a fixed, preset frequency. This ensures that the fault detection process can continuously and in real time obtain the latest operating data. Based on this data, the absolute value of the deviation between the actual thrust output and the thrust command for each power unit in the current cycle is calculated. This deviation directly reflects the degree of deviation between the power unit's output and the expected command. To distinguish between transient fluctuations and persistent anomalies, the system further determines whether the deviation exceeds a preset deviation threshold and accumulates the duration of the deviation exceeding the threshold. Only when this significant deviation persists for more than a preset delay time threshold is the power unit considered to have a response delay fault. This duration-based judgment logic effectively avoids misjudgments caused by short-term deviations due to normal system noise, transient response, or external disturbances. Furthermore, to detect stuck faults, the system specifically monitors whether the power unit's actual thrust output changes within a preset observation time after receiving a new thrust command. If the actual thrust output does not change as expected within a period of time after receiving the new command, the power unit is directly judged to have a stuck fault. This detection method, which targets the unique manifestations of stuckness, can quickly and accurately identify power units that have lost their ability to respond. By combining hysteresis detection based on continuous deviation with stuckness detection based on response changes, this solution can more comprehensively and accurately identify the two main failure modes of the power unit. This precise fault judgment provides a reliable premise for subsequent fault handling steps, enabling the entire attitude control method to more effectively respond to power unit failures and maintain the stability and control accuracy of the flying saucer.

[0055] In some embodiments, step A2 comprises:

[0056] A201. Extract the thrust command sequence of the faulty unit, the actual thrust output sequence, and the motion state data sequence of the manned flying saucer; the motion state data sequence includes an attitude angular velocity sequence and a moving speed sequence;

[0057] A202. Call a thrust output characteristic identification model pre-established based on the least squares method; the thrust output characteristic identification model takes thrust command, attitude angular velocity, and position velocity as input, and actual thrust output as output, and includes three parameters to be identified: thrust proportional coefficient, hysteresis time, and hysteresis threshold;

[0058] A203. Using the recursive least squares method, based on the extracted thrust command sequence, actual thrust output sequence, attitude angular velocity sequence, and movement velocity sequence, the parameters to be identified in the thrust output characteristic identification model are updated in real time to obtain the real-time thrust proportional coefficient, hysteresis time, and stuck threshold of the faulty unit.

[0059] A204. Calculate the nominal thrust value of the faulty unit based on the real-time thrust proportional coefficient of the faulty unit.

[0060] A205. Determine whether the actual thrust output of the faulty unit is affected by hysteresis or sticking based on the hysteresis time and stuckness threshold of the faulty unit. If so, modify the nominal thrust value to obtain the effective thrust value of the faulty unit. Otherwise, use the nominal thrust value as the effective thrust value of the faulty unit.

[0061] A206. The real-time thrust proportional coefficient, hysteresis time, stuck threshold and effective thrust value are used as the thrust information of the faulty unit.

[0062] In step A201, the relevant data streams of the faulty unit are extracted. The thrust command sequence refers to the time-varying record of the thrust commands sent by the control system to the faulty unit. The actual thrust output sequence refers to the measured value of the thrust actually generated by the faulty unit over time, which can be obtained using a force sensor installed on the power unit or by estimating parameters such as motor current and voltage. The motion state data sequence of the manned flying saucer includes an attitude angular velocity sequence and a movement velocity sequence. The attitude angular velocity can be measured using a gyroscope, while the movement velocity can be measured using an inertial measurement unit or a navigation system such as GPS. These data sequences serve as the basic input for thrust characteristic identification.

[0063] Among them, the thrust output characteristic identification model is pre-established based on the least squares method. Its structure describes the relationship between the thrust command, attitude angular velocity, position velocity and actual thrust output, and explicitly includes three parameters: thrust proportional coefficient, hysteresis time and stuck threshold. The model can be implemented using a linear or nonlinear regression model. For example, the actual thrust output can be modeled as a function of the thrust command, motion state variables and their nonlinear combination, which includes parameters related to the proportional coefficient, hysteresis and stuckness to be identified. Taking motion state data as input helps to capture the potential coupling effects between the thrust output of the power unit and the overall motion state of the flying saucer, such as airflow influence or structural vibration feedback.

[0064] For example, the model is: F_a(t)=p1*F_c(t-T_d)+p2*w_x(t)+p3*w_y(t)+p4*w_z(t)+p5*v_x(t)+p6*v_y(t)+p7*v_z(t)+p8+e(t); where t is time, F_a is the actual thrust, F_c is the thrust command, w is the attitude angular velocity vector, w_x, w_y and w_z are the components of w in the three axes, v is the moving velocity vector, v_x, v_y and v_z are the components of v in the three axes, p1-p8 are parameters to be identified, e is the error term, and T_d is the hysteresis time. The thrust proportional coefficient can be represented by p1, and the hysteresis time T_d and the stuck threshold can be obtained by adjusting the model structure or introducing additional identification steps. For example, the hysteresis time can be estimated by identifying the optimal time alignment between the command and the response, and the stuck threshold can be estimated by analyzing the residual or the actual thrust change rate, or T_d and the stuck threshold can be identified as nonlinear parameters in the model.

[0065] In step A203, the recursive least squares method (RLSM) is an online identification algorithm that updates the model's parameter estimates in real time based on newly acquired data. Specifically, the algorithm maintains a parameter estimate vector and a covariance matrix. Whenever a new data point (thrust command, actual thrust output, attitude angular velocity, and movement speed) is input, the algorithm adjusts the parameter estimate vector according to a preset update formula to more closely approximate the actual parameter values ​​at the current moment. RLSM can be implemented with a forgetting factor. For example, the forgetting factor can be set between 0.95 and 1 to balance reliance on historical data with responsiveness to the latest data. This real-time update allows for dynamic tracking of the faulty unit's thrust characteristics, which may change over time, improving the real-time and adaptability of the identification results. This step outputs the faulty unit's current real-time thrust proportional coefficient, hysteresis time, and stuck threshold estimates.

[0066] For example, the recursive least squares method updates model parameters p1-p8 online based on the data collected in real time in step A201. Simultaneously, the hysteresis time T_d is estimated in real time by analyzing the time difference between the command change and the actual thrust response. The stuck threshold is estimated in real time by monitoring the magnitude of the change in actual thrust output after the command change. Thus, the system obtains the thrust proportional coefficient (e.g., the current value of p1), the hysteresis time (e.g., the estimated value T_d_est), and the stuck threshold (e.g., the estimated value Th_est) of the faulty unit in real time.

[0067] In step A204, the nominal thrust value is the ideal thrust output calculated based on the current thrust command and the identified real-time thrust scaling factor. For example, if the real-time thrust scaling factor is k and the current thrust command is u, the nominal thrust value can be simply calculated as k*u. This value represents the thrust output expected based on the current command and the scaling relationship, assuming no hysteresis or sticking.

[0068] In step A205, the identified hysteresis time and stuck threshold are used to determine whether the actual thrust output of the faulty unit at the current moment is affected by these two non-ideal characteristics. If so, the nominal thrust value is modified based on the degree of impact to obtain an effective thrust value that is closer to the actual situation. The effective thrust value reflects the actual thrust that the faulty unit can produce and is available for control in its current state.

[0069] In step A206, the thrust information includes the real-time thrust scaling factor, hysteresis time and stuck threshold, and effective thrust value. This information comprehensively describes the current thrust characteristics and actual output capacity of the faulty unit. The thrust information identified in this step provides key input for subsequently determining the control mode of the faulty unit and resolving the thrust commands for the normal power units. By providing accurate thrust information about the faulty unit, the subsequent control algorithm can more effectively utilize the capabilities of the remaining normal units to compensate for the impact of the faulty unit, thereby maintaining the attitude and position control accuracy of the flying saucer. This approach, combining the identification of faulty unit characteristics with subsequent control decisions, enables the system to achieve high-precision control even in the presence of complex faults.

[0070] Through the above technical solution, the present application improves the recognition accuracy and real-time performance of the complex thrust output characteristics of the faulty unit, making the acquired thrust information of the faulty unit more reliable, providing accurate input for subsequent fault-tolerant control, and enhancing the attitude and position control performance of the system when there is response hysteresis or stuck faults.

[0071] Preferably, step A205 may include:

[0072] Calculate the time difference between the moment the thrust command changes and the moment the actual thrust begins to respond. If the time difference is greater than the hysteresis time of the faulty unit, then determine that the faulty unit is affected by hysteresis.

[0073] Monitor the change in actual thrust output over a number of consecutive preset observation periods. If the change is less than a stuck threshold of the faulty unit, the faulty unit is determined to be affected by stuckness.

[0074] If the faulty unit is only affected by hysteresis, then based on the time difference, a pre-calibrated hysteresis compensation model is used to perform time compensation on the nominal thrust value to obtain a corrected thrust value;

[0075] If the faulty unit is affected by sticking, the average value of the actual thrust output during several consecutive preset observation periods is used as the corrected thrust value;

[0076] If the faulty unit is not affected by hysteresis or sticking, the original nominal thrust value is used as the corrected thrust value;

[0077] The corrected thrust value is taken as the effective thrust value of the faulty unit.

[0078] Among them, the time point when the new thrust command is sent to the faulty unit is recorded, and the actual thrust output of the faulty unit is monitored. When the actual thrust output begins to undergo a detectable change, the time point is recorded. The difference between the two time points is the time difference from the moment the thrust command changes to the moment the actual thrust begins to respond.

[0079] The difference between the maximum and minimum values ​​of the actual thrust output data collected continuously over several preset observation periods (the specific number can be set according to actual needs, for example, 3) can be used as the variation amplitude.

[0080] Among them, the pre-calibrated hysteresis compensation model refers to a mathematical model of the delay relationship between the thrust command change and the actual thrust response established through experiments or simulations. The model can be a function, a lookup table or a dynamic system model. Specifically, the system identification method can be used for offline calibration. The model is used to correct the nominal thrust value according to the measured time difference to compensate for the impact of response hysteresis.

[0081] This scheme is combined with real-time updating of the thrust proportional coefficient, hysteresis time and stuck threshold of the faulty unit through identification. The parameters obtained by identification are used to accurately judge the fault type, and different correction strategies are adopted according to these parameters. This makes the assessment of the thrust output characteristics of the faulty unit more accurate, providing a more reliable basis for subsequent control decisions.

[0082] Specifically, when a faulty unit exhibits an anomaly, the first step is to determine the specific impact. The time difference between the issuance of a thrust command and the onset of actual thrust change is calculated and compared with the faulty unit's hysteresis time, as determined by real-time identification. If the time difference exceeds the hysteresis time, the faulty unit has a significant response delay, indicating that it is affected by hysteresis. Furthermore, the magnitude of the faulty unit's actual thrust output change over a period of time is continuously monitored and compared with the faulty unit's stuck threshold (which can be set as needed) as determined by real-time identification. If the magnitude of the change is less than the stuck threshold, the faulty unit's thrust output change is minimal or nonexistent, indicating that it is affected by stuckness. After determining the type of impact on the faulty unit, the nominal thrust value calculated based on the real-time thrust scaling factor is corrected according to the specific situation. If only hysteresis is present, the measured time difference is used to compensate the nominal thrust value using a pre-established hysteresis compensation model to predict the likely actual thrust value after accounting for the delay. If the faulty unit is affected by hysteresis, the thrust output may be unstable or fixed at a certain value during this state. In this case, the average of the actual thrust output over a continuous period of time is used as the corrected thrust value, which better reflects the faulty unit's current actual thrust level. If the faulty unit is neither affected by hysteresis nor by hysteresis, its thrust output is considered to be basically normal, and the nominal thrust value is directly used as the corrected thrust value. Ultimately, the value obtained through the above different correction processes is determined as the effective thrust value of the faulty unit. This effective thrust value comprehensively considers the faulty unit's real-time thrust proportional coefficient, response hysteresis, and hysteresis characteristics. Compared to using only the nominal thrust value or the original actual thrust output, it can more accurately reflect the faulty unit's current thrust output capacity or status. This provides more precise input information for the subsequent determination of the faulty unit's control mode and the thrust optimization of the normal power unit, helping to maintain high-precision control of the UFO during fault conditions.

[0083] In some embodiments, step A3 comprises:

[0084] A301. According to the thrust information of the faulty unit, determine the state type of the faulty unit; the state type includes a complete failure state, a response hysteresis state, and a stuck state;

[0085] A302. If the status type of the faulty unit is a complete failure state, the control mode of the faulty unit is set to the disabled mode and the thrust output of the faulty unit is turned off;

[0086] A303. If the faulty unit is in a delayed response state, evaluate whether the faulty unit has any remaining thrust capability based on the real-time thrust proportional coefficient in the thrust information.

[0087] A304. If the faulty unit has residual thrust capacity, set the control mode of the faulty unit to hysteresis compensation mode. Based on the hysteresis time in the thrust information, use a pre-calibrated hysteresis compensation model to compensate the thrust command of the faulty unit and send the compensated thrust command to the actuator of the faulty unit.

[0088] A305. If the faulty unit has no remaining thrust capacity, set the control mode of the faulty unit to the disabled mode and shut down the thrust output of the faulty unit;

[0089] A306. If the state type of the faulty unit is a stuck state, set the control mode of the faulty unit to a fixed thrust mode, and fix the thrust output of the faulty unit to the effective thrust value in the thrust information.

[0090] Among them, the deactivation mode refers to removing the faulty unit from the control system, no longer sending thrust commands to it, and physically or logically shutting down its thrust output. The hysteresis compensation mode refers to pre-processing the thrust commands sent to the faulty unit in the response hysteresis state to offset or reduce the impact of its response delay, so that its actual thrust output can be closer to the expected value. The fixed thrust mode refers to treating the thrust output of the faulty unit in the stuck state as a constant, known external force or torque. In the subsequent control calculation, no attempt is made to control the unit. Instead, its fixed thrust is incorporated into the overall dynamic model as a known disturbance.

[0091] Among them, the residual thrust capacity refers to whether the faulty unit in the response hysteresis state can still generate a sufficiently large thrust output that can be used for control after considering its hysteresis characteristics.

[0092] The core of the method proposed in this application lies in adopting targeted control modes based on the specific state of the faulty unit. This differs from simply determining the presence of a fault or applying a uniform treatment approach, and can more accurately address the diversity of faulty units. This method relies closely on the accurate identification of the faulty unit's thrust information in the previous step. First, based on the identified thrust information, the faulty unit's state is clearly classified into three types: complete failure, response delay, or stuck. This classification forms the basis for subsequent treatment strategies and avoids the use of a single, ineffective or suboptimal approach for different fault types. If a faulty unit is determined to be in a complete failure state, indicating a loss of thrust capability, its control mode is set to disabled and thrust output is turned off. This removes the unit from control, avoids sending invalid commands to it or including its thrust in calculations, simplifies the subsequent thrust allocation of healthy units, and eliminates the uncertainty it may introduce. If the faulty unit is determined to be in a response delay state, its remaining thrust capability is further assessed based on the real-time thrust proportional coefficient in the thrust information. If the evaluation results indicate that the unit still has residual thrust capacity, its control mode is set to hysteresis compensation mode. Based on the hysteresis time in the thrust information, a pre-calibrated hysteresis compensation model is used to compensate for its thrust command. This compensation method can offset the impact of response delays to a certain extent, allowing the unit's residual thrust capacity to be utilized, reducing the burden on the remaining healthy units and facilitating more precise control. If the evaluation results indicate that the faulty unit with hysteresis response does not have residual thrust capacity, its control mode is similarly set to disabled mode and its thrust output is shut down, ensuring that only units with actual thrust capacity are considered for continued use. If the faulty unit is determined to be stuck, indicating that its thrust output is fixed or varies minimally, its control mode is set to fixed thrust mode, and its thrust output is fixed to the effective thrust value in the thrust information. In this way, in subsequent thrust allocation calculations, this fixed thrust can be treated as a known, constant external disturbance, which is compensated by the remaining healthy units, rather than attempting to control an unresponsive unit. This improves the accuracy and effectiveness of thrust allocation to the healthy units.

[0093] Through the above steps, the method of the present application can adopt the most appropriate control mode according to the different state types of the faulty unit, thereby more accurately processing the faulty unit and providing accurate input and constraints for the subsequent thrust reconstruction of the normal power unit.

[0094] In one embodiment, determining the faulty unit's state type based on its thrust information can be specifically implemented. For example, this can be determined based on the real-time thrust proportionality factor and the stuck threshold. If the real-time thrust proportionality factor remains below a preset failure threshold for a prolonged period, the faulty unit can be determined to be in a completely failed state. If the faulty unit is not in a completely failed state, and the stuck threshold is below a preset stuck amplitude threshold, the faulty unit can be determined to be in a stuck state; otherwise, the faulty unit is determined to be in a response hysteresis state. For a faulty unit in a response hysteresis state, assessing whether it has residual thrust capability can be performed based on the real-time thrust proportionality factor. For example, if the real-time thrust proportionality factor is greater than a non-zero threshold, the unit is considered to have residual thrust capability. If the faulty unit has residual thrust capability and is set to hysteresis compensation mode, a pre-calibrated hysteresis compensation model can be used to compensate for the thrust command. This hysteresis compensation model can be based on a Smith Predictor, which predicts the faulty unit's future output based on the identified hysteresis time and adjusts the current thrust command accordingly. When a faulty unit is stuck and set to fixed thrust mode, the effective thrust value in the thrust information can be used as the constant thrust output of the unit. In subsequent thrust allocation calculations, this fixed thrust value and the torque it generates can be treated as a known external disturbance acting on the manned vehicle, which is compensated by the remaining normal power units.

[0095] Through the above technical solution, the present application can adopt the most appropriate control mode according to the different state types of the faulty unit, thereby more accurately handling the faulty unit and solving the problem of improper handling due to the diversity of the faulty unit state. This provides accurate input and constraints for the subsequent thrust reconstruction of the normal power unit, improves the performance and robustness of the entire attitude control system in the event of a fault, and helps to achieve high-precision attitude control of the manned flying saucer.

[0096] Preferably, step A301 may include:

[0097] If the real-time thrust proportional coefficient is lower than the preset failure threshold and the duration exceeds the preset failure time threshold, the state type of the faulty unit is determined to be a complete failure state;

[0098] When the state type of the faulty unit is not a complete failure state, if the stuck threshold is lower than the preset stuck amplitude threshold, the state type of the faulty unit is determined to be a stuck state; otherwise, the state type of the faulty unit is determined to be a response hysteresis state.

[0099] The preset failure threshold refers to the lower limit of the real-time thrust proportional coefficient used to determine whether a faulty unit has completely lost thrust capability. This can be set based on the power unit's design parameters or experimental data. The duration refers to the duration of time the real-time thrust proportional coefficient remains below the preset failure threshold. The preset failure time threshold refers to the upper limit of the duration used to determine whether a faulty unit is in a completely failed state. This can be set based on the system's tolerance for transient anomalies.

[0100] The preset stuck amplitude threshold refers to a predetermined upper limit value of the stuck threshold used to determine whether the faulty unit is in a stuck state, which can be set according to the normal thrust fluctuation range of the power unit or experimental data.

[0101] By determining a complete failure based on the real-time thrust scaling factor and duration, and then determining whether a stuck or delayed response state is present based on the stuck threshold, this method accurately distinguishes the specific state type of the faulty unit. This accurate state classification enables more targeted control mode selection, such as deactivating completely failed units, applying fixed thrust to stuck units, and compensating for delayed response units. Consequently, effective compensation measures can be implemented based on the actual manifestation of the fault, effectively restoring the attitude control accuracy of the manned flying saucer and improving the system's adaptability to different types of power unit failures.

[0102] In some embodiments, step A5 comprises:

[0103] A501. Construct a comprehensive objective function that includes an attitude tracking error term, a position tracking error term, and a thrust adjustment term. The attitude tracking error term and the position tracking error term are calculated based on the real-time attitude, real-time position, desired attitude, and desired position of the manned vehicle. The thrust adjustment term is calculated based on the thrust command change rate of each normal power unit.

[0104] A502. Based on the thrust information and control mode of the faulty unit, calculate the net force and net torque generated by the faulty unit on the manned flying vehicle. Apply this as an external disturbance force to the dynamic model of the manned flying vehicle, thereby obtaining a dynamic model of the flying vehicle that takes into account the thrust disturbance caused by the faulty unit.

[0105] A503. Determine the thrust feasible region of each normal power unit based on the dynamic constraint model and operational constraint information of the normal power unit; the thrust feasible region defines the available thrust range and available thrust change rate range of each normal power unit;

[0106] A504. Taking the comprehensive objective function as the optimization target, the UFO dynamics model taking into account the thrust interference of the faulty unit and the thrust feasible domain as constraints, the quadratic programming method is used to solve and obtain the thrust instructions of each normal power unit that meets the constraints and minimizes the comprehensive objective function.

[0107] In step A501, attitude error quaternions, position error vectors, and other forms can be used to quantify the degree of deviation in attitude and position. The thrust adjustment term can use the squared norm of the thrust command rate of change vector to measure the severity of the thrust change. By weighting and summing these terms to form a single optimization objective, the optimization process can simultaneously consider control accuracy and operational smoothness. For example, the comprehensive objective function can be: J = w_p*||p_e||^2+w_q*||q_e||^2+w_r*||ΔT_n||^2; where p_e is the position tracking error vector, q_e is the attitude tracking error quaternion (or equivalent representation), ΔT_n is the normal power unit thrust command rate of change vector, and w_p, w_q, and w_r are weight coefficients used to balance the importance of position tracking, attitude tracking, and thrust smoothness. ||*|| represents the norm of a vector or quaternion. The position tracking error vector can be calculated as the difference between the desired position and the real-time position. The attitude tracking error quaternion can be calculated as the product of the desired attitude quaternion and the real-time attitude quaternion. The i-th element of the thrust command change rate vector ΔT_n is the thrust command change of the normal power unit i in the current cycle.

[0108] In step A502, the forces and moments generated by the faulty unit are calculated based on its installation location and thrust direction. These forces and moments are considered known external disturbances acting on the flying saucer and incorporated into the six-degree-of-freedom dynamic equations of the flying saucer. This modifies the dynamic model, allowing subsequent thrust distribution calculations to directly compensate for the effects of the faulty unit.

[0109] For example, based on the thrust information of the faulty unit (e.g., the effective thrust value T_f_eff obtained in the previous step) and the control mode (e.g., fixed thrust mode), calculate the resultant force F_f and resultant torque τ_f generated by the faulty unit on the flying saucer. For example, if the faulty unit f is in fixed thrust mode, its effective thrust is T_f_eff, and its installation position vector is r_f, then the force it generates is F_f = T_f_eff * u_f (u_f is the unit vector of the thrust direction), and the torque it generates is τ_f = r_f * F_f. The forces and torques generated by all faulty units are superimposed to obtain the total fault interference force F_dist and interference torque τ_dist; the dynamic model of the manned flying saucer can be expressed as:

[0110] m*a=F_n+F_dist

[0111] I*α+ωx(I*ω)=τ_n+τ_dist

[0112] Where m is the mass of the manned flying saucer, a is the acceleration of the movement, F_n is the resultant force generated by the normal power unit, I is the flying saucer's moment of inertia matrix, α is the attitude angular acceleration, ω is the attitude angular velocity, and τ_n is the resultant torque generated by the normal power unit. F_n and τ_n are linear functions of the normal power unit thrust command T_n:

[0113] [F_n;τ_n]=B_n*T_n

[0114] Where B_n is the moment arm matrix of the normal power unit. Moving the interference term to the right side of the equation, we get the UFO dynamics model considering the thrust interference of the faulty unit:

[0115] m*a=B_n_F*T_n+F_dist

[0116] I*α=B_n_τ*T_n+τ_dist-ωx(I*ω)

[0117] Where B_n_F and B_n_τ are the corresponding parts of B_n.

[0118] In step A503, the thrust feasible region defines the thrust value that each normal power unit can output during the current control cycle and its range of variation relative to the previous cycle. This ensures that the calculated thrust command is physically achievable and does not exceed the power unit's capability.

[0119] In step A504, quadratic programming is a method for solving optimization problems involving quadratic objective functions with linear equality and inequality constraints. The comprehensive objective function constructed in step A501 is converted to a quadratic form, and the dynamic model modified in step A502 and the thrust feasible region determined in step A503 are converted into linear equality or inequality constraints. By solving this quadratic programming problem, a set of thrust instructions for a normal power unit can be quickly found. This set of instructions minimizes attitude and position tracking errors while maintaining the proper physical constraints of the flying saucer and the unit, while also ensuring the smoothest thrust changes.

[0120] Specifically, this solution provides a specific, optimization-based method for determining thrust commands, designed to achieve high-precision attitude control of a manned flying saucer using the remaining functioning power units in the presence of a faulty power unit. In step A501, this method constructs a comprehensive objective function. By incorporating attitude tracking error and position tracking error terms, this function directly quantifies the control objective of bringing the actual state of the flying saucer closer to the desired state, ensuring high-precision tracking of attitude and position. Furthermore, by incorporating a thrust adjustment term, thrust smoothness is considered, avoiding drastic changes in thrust commands and helping to improve flight comfort and structural stability. In step A502, based on the thrust information and control mode of the faulty unit, the resultant force and torque generated by the faulty unit on the flying saucer are calculated. These torques are then applied as external disturbance forces to the flying saucer's dynamic model, resulting in a flying saucer dynamic model that accounts for the thrust disturbance caused by the faulty unit. This allows subsequent thrust allocation calculations to directly account for and compensate for the effects of the faulty unit, enhancing the robustness of the control system. In step A503, based on the dynamic constraint model and operational constraint information of the normal power units, the thrust feasible region of each normal power unit is determined, limiting the available thrust range and available thrust change rate range of each normal power unit. This ensures that the calculated thrust command is physically feasible, avoids commands that exceed the unit's capabilities, and ensures the reliability of the control system. Finally, in step A504, using the comprehensive objective function as the optimization target, the UFO dynamics model that considers the thrust interference of the faulty unit and the thrust feasible region as constraints, a quadratic programming method is used to solve the thrust commands for each normal power unit that meet the constraints and minimize the comprehensive objective function. By selecting the computationally efficient optimization method of quadratic programming, the optimal normal unit thrust commands that meet all physical constraints and interference compensation requirements can be quickly solved. This set of commands maximizes attitude and position tracking accuracy while ensuring smooth thrust changes.

[0121] Compared with the existing technology, which has difficulty in achieving high precision, fast response and smooth control under complex constraints and interference, this solution effectively solves these problems by constructing a comprehensive objective function, considering the interference of faulty units, determining the constraints of normal units, and using quadratic programming methods to solve them.

[0122] Through the above technical solution, the present application can effectively handle complex factors such as the unpredictable thrust output of the faulty unit, the complex nonlinear characteristics of the remaining normal power units, response delays, and physical coupling effects when a manned flying saucer power unit experiences a response hysteresis or stuck fault, accurately calculating and rapidly implementing a new thrust distribution scheme. This allows for rapid and high-precision restoration of control in the event of a fault, while maintaining smooth flight.

[0123] Preferably, step A503 may include:

[0124] B1. For each normal power unit, based on its dynamic constraint model, establish a nonlinear function with thrust command as input and attitude angular acceleration and position acceleration as output;

[0125] B2. Using the operational constraints of each normal power unit as a constraint, combined with the nonlinear function, by solving the constrained optimization problem, calculate the feasible range of the attitude angular acceleration and position acceleration of each normal power unit under the constraints of its own manned flying saucer;

[0126] B3. Perform vector superposition of the feasible ranges of attitude angular acceleration and position acceleration corresponding to all normal power units to obtain the feasible region of the overall attitude angular acceleration and overall position acceleration of the manned flying saucer;

[0127] B4. Based on the feasible domain of the overall attitude angular acceleration and the overall position acceleration, inversely solve the thrust range of each normal power unit. Based on the inverse solution, determine the available thrust range and the available thrust change rate range to obtain the thrust feasible domain.

[0128] Among them, step B1 refers to constructing a mathematical expression for each power unit in normal working state based on the mathematical model that describes its thrust output characteristics, response delay and physical coupling effect. This expression can describe how the thrust generated by the power unit acts on the manned flying saucer when it receives a specific thrust instruction, and causes the flying saucer to generate corresponding attitude angular acceleration and position acceleration. Specifically, it can be implemented using a set of dynamic equations containing nonlinear terms. For example, the force and torque generated by a single power unit can be expressed as a nonlinear function of the thrust instruction, and then converted into the influence on the overall attitude angular acceleration and position acceleration of the flying saucer through the inertia matrix and position relationship of the flying saucer. The purpose of establishing this nonlinear function is to quantify the contribution of a single normal power unit to the flying saucer's motion control capability.

[0129] Among them, step B2 refers to using the nonlinear function established in step B1 and taking into account the operational limitations of each normal power unit itself. These operational limitations include the range of thrust that the power unit can generate and the speed limit of thrust change. By solving a mathematical optimization problem, for example, the goal can be set to maximize or minimize the acceleration in a certain direction, while limiting the thrust instruction to its operating range, so as to calculate the boundary set of attitude angular acceleration and position acceleration that a single normal power unit can independently achieve for the manned flying saucer under the premise of satisfying its own thrust and thrust change rate constraints. This set constitutes the boundary of a single normal unit's control over the flying saucer's motion within its own capabilities. Specifically, methods such as nonlinear programming or convex optimization can be used to solve this constrained optimization problem.

[0130] Step B3 involves combining the feasible ranges of attitude angular acceleration and position acceleration calculated by all normal power units in step B2. Since the overall motion of a manned flying saucer is the result of the net force and torque generated by the combined action of all power units, the total acceleration range that can be generated by all normal power units working together can be obtained by vector summing the feasible ranges of acceleration generated independently by each unit. This vector superposition process can be understood as calculating the Minkowski sum of all individual feasible ranges, thereby obtaining the set of control capabilities of all remaining normal power units combined to control the overall motion of the flying saucer. This set reflects the overall feasible range of attitude angular acceleration and position acceleration that the remaining normal units can achieve after a fault occurs.

[0131] Among them, step B4 refers to the feasible domain of the manned flying saucer's overall attitude angular acceleration and overall position acceleration obtained in step B3, and reversely deriving the thrust combination range of each normal power unit that can produce these accelerations. This usually involves solving the inverse problem of the flying saucer dynamics equation, that is, given the desired acceleration range, solving the set of thrust vectors required to produce these accelerations. Based on the thrust range obtained by this inverse solution, and further considering the thrust change rate limits of each power unit, the actual available thrust range and thrust change rate range of each normal power unit in the current state are finally determined. This ultimately determined available thrust range and available thrust change rate range together constitute the thrust feasible domain. This thrust feasible domain is a direct constraint condition for the subsequent thrust optimization solution, ensuring that the calculated thrust command is both within the actual capabilities of the remaining units and can achieve the desired flying saucer motion control.

[0132] For example, in step B4, first, after knowing the feasible domain of the manned flying saucer's overall attitude angular acceleration and position acceleration, the dynamic model of the flying saucer is used to reversely calculate the thrust combination of each normal power unit that can produce these accelerations. This inverse solution process involves solving a system of linear or nonlinear equations, resulting in a set of thrust vectors, denoted as Q. Each thrust vector in this set is capable of causing the manned flying saucer to produce an acceleration within the feasible domain.

[0133] Next, the thrust rate limits of each normal power unit need to be considered. The speed at which each power unit's thrust changes is limited by its physical structure and cannot be changed infinitely quickly. These limits are typically expressed as upper and lower limits on the thrust rate (i.e., the upper and lower limits of the reference thrust rate range), and the thrust rate of each unit must fall within this range. Specifically, for each thrust vector in Q, its thrust change compared to the actual thrust vector at the previous moment is checked to see if it exceeds the limits of the reference thrust rate range. If so, the thrust vector is removed from Q.

[0134] After filtering by the thrust rate constraints, the remaining set of thrust vectors represents the thrust combinations that, under the current state, allow each normal power unit to both generate feasible acceleration and meet its own thrust rate constraints. Based on this set, the available thrust range and available thrust rate range for each normal power unit can be determined. For example, for the i-th normal power unit, its available thrust range can be determined as the minimum to maximum thrust values ​​of that unit in the remaining set. Similarly, its available thrust rate range can be determined as the minimum to maximum thrust rate values ​​of that unit in the remaining set.

[0135] Ultimately, the available thrust range and available thrust change rate range of each normal power unit together constitute the thrust feasible region. This thrust feasible region is a direct constraint for the subsequent thrust optimization solution, ensuring that the calculated thrust command is both within the actual capabilities of the remaining units and can achieve the desired UFO motion control.

[0136] Through the above technical solution, the present application can accurately determine the actual control capability range of the remaining normal power units, taking into account their own complex dynamic constraints and operational constraints. This transforms the complex constraints of a single unit into constraints on the overall motion of the flying saucer, and takes into account the physical coupling effects between multiple units. This provides a reliable constraint basis for high-precision thrust reconstruction after a failure, allowing thrust optimization solutions to be performed within the actual capabilities of the remaining units, thereby improving the accuracy and reliability of the flying saucer's attitude and position control after a failure.

[0137] Preferably, step A504 may include:

[0138] C1. Based on the real-time attitude, real-time position, desired attitude and desired position of the manned flying saucer, according to the comprehensive objective function, determine the desired attitude angular acceleration and desired position acceleration of the manned flying saucer;

[0139] C2. Based on the UFO dynamics model that takes into account the thrust interference of the faulty unit, establish the functional relationship between the desired attitude angular acceleration and the desired position acceleration of the manned UFO and the thrust command of each normal power unit;

[0140] C3. Based on the functional relationship and the thrust feasible region, construct a quadratic programming problem with the thrust command of each normal power unit as the optimization variable; wherein the objective function of the quadratic programming problem is to minimize the error between the actual thrust of each normal power unit and the desired thrust obtained by inversely solving the desired attitude angular acceleration and the desired position acceleration, with the constraint that the thrust command of each normal power unit is within the thrust feasible region;

[0141] C4. Solve the quadratic programming problem to obtain thrust instructions for each normal power unit that meets the constraints and minimizes the comprehensive objective function.

[0142] Step C1 involves calculating the instantaneous attitude angular acceleration and positional acceleration required for the manned flying saucer to move to the desired state based on the current state information (real-time attitude, real-time position) and target state information (desired attitude, desired position), combined with a function that comprehensively considers attitude tracking error, position tracking error, and thrust adjustment. This can be achieved using a control law based on error feedback, such as proportional-integral-derivative (PID) control or linear quadratic regulator (LQR). The control law takes attitude error, position error, and their rate of change as inputs and calculates the desired acceleration command. The comprehensive objective function can be designed as a weighted sum of the attitude error term, position error term, and thrust rate of change term. Minimizing this function determines the optimal desired acceleration.

[0143] Step C2 involves establishing a mathematical relationship between the control output (desired attitude angular acceleration and desired position acceleration) and the control input (thrust commands for each normal power unit) based on the dynamic model of the flying saucer. This dynamic model takes into account the mass and moment of inertia of the flying saucer, the installation position and orientation of each power unit, and the known or estimated interference forces / torques generated by the faulty unit. Specifically, the Newton-Euler equations can be used to describe the translational and rotational dynamics of the flying saucer. The sum of the resultant forces and torques generated by the normal power units and the interference forces / torques generated by the faulty units is equal to the mass of the flying saucer multiplied by the position acceleration and the moment of inertia multiplied by the attitude angular acceleration. This results in a linear or nonlinear system of equations that express the desired acceleration as a function of the thrust commands of the normal power units.

[0144] Among them, step C3 refers to converting the problem of solving the thrust command of the normal power unit into a quadratic programming problem. Specifically, the thrust command of each normal power unit is set as the optimization variable to be solved. The objective function of the quadratic programming problem is constructed to minimize a quadratic function, which represents the sum of the squares of the errors between the thrust actually generated by the normal power unit and the ideal thrust required to achieve the desired acceleration calculated in step C1. The ideal thrust can be obtained by inversely solving the dynamic model established in step C2. The constraints ensure that the thrust command obtained by the solution meets the physical limitations, that is, the thrust command of each normal power unit must be within its thrust feasible domain. The thrust feasible domain is determined by the thrust range (minimum and maximum values) and thrust change rate range of the normal power unit.

[0145] Step C4 involves using a standard quadratic programming algorithm to calculate the solution to the optimization problem constructed in step C3. Specifically, established numerical optimization algorithms such as the interior point method and the active set method can be employed. The solution seeks the normal power unit thrust command vector that minimizes the objective function while satisfying all thrust feasible region constraints. The resulting thrust command is the optimal thrust allocation solution at the current moment.

[0146] Specifically, this solution addresses the issue of how to effectively construct and solve a quadratic programming problem to achieve high-precision control and thrust smoothness. First, in step C1, the desired attitude angular acceleration and position acceleration of the flying saucer are calculated based on the saucer's current state, desired state, and a function that comprehensively considers control objectives and thrust smoothness. This desired acceleration serves as the control command for achieving high-precision tracking and thrust smoothness. Next, in step C2, a mathematical relationship is established between the desired acceleration and the thrust command of the normal power unit, using a flying saucer dynamics model that accounts for the thrust disturbance of the faulty unit. This relationship, which forms the basis for thrust allocation, reflects the required net force and torque generated by the normal power unit to achieve the desired acceleration, while also accounting for the need to compensate for the disturbance of the faulty unit. In step C3, a quadratic programming problem is constructed based on the mathematical relationship established in step C2 and the thrust feasible region of the normal power unit. The optimization variable in this problem is the thrust command of each normal power unit. The objective function is set to minimize the error between the actual thrust of the normal power unit and the ideal thrust required to achieve the desired acceleration. The constraints ensure that the thrust instructions for each normal power unit are within its physically permitted range. Finally, in step C4, this quadratic programming problem is solved to obtain a set of normal power unit thrust instructions that satisfy the constraints and minimize the objective function. This set of instructions represents how the normal power units can optimally work together in the current state to achieve the desired acceleration as much as possible, thereby reducing tracking error, while satisfying thrust limitations and indirectly considering thrust smoothness. In this way, the scheme provides a specific and computable thrust allocation method that can effectively handle fault interference and thrust constraints, achieving high-precision control and thrust smoothness.

[0147] refer to Figure 2 The present application provides a manned flying saucer attitude control system, which is applied to a manned flying saucer with multiple power units. The system includes:

[0148] Fault detection module 1 is used to monitor the thrust command and actual thrust output of each power unit and determine whether there is a faulty unit with response hysteresis or stuck fault in each power unit (for the specific process, refer to step A1 above);

[0149] Thrust characteristic identification module 2 is used to identify the thrust output characteristics of a faulty unit based on the relevant data stream of the faulty unit when a faulty unit exists, and obtain thrust information of the faulty unit; the relevant data stream includes the data stream of the thrust command and actual thrust output of the faulty unit, as well as the data stream of the motion state data of the manned flying saucer (for the specific process, refer to step A2 above);

[0150] Control mode determination module 3, used to determine the control mode for the faulty unit based on the thrust information of the faulty unit (for the specific process, refer to step A3 above);

[0151] Data extraction module 4 is configured to extract the dynamic constraint model and operational constraint information of the remaining normal power units from the local database; the dynamic constraint model includes a mathematical model of the thrust output characteristics, response delay, and physical coupling effects of the normal power units; the operational constraint information includes a reference thrust range and a reference thrust change rate range of the normal power units (for details, refer to step A4 above);

[0152] Thrust optimization module 5 is used to solve the thrust command of each normal power unit based on the real-time attitude, real-time position, desired attitude and desired position of the manned flying saucer, as well as the thrust information of the faulty unit, the control mode of the faulty unit, the operating constraint information of the normal power unit and the dynamic constraint model of the normal power unit, with the goal of minimizing the attitude tracking error, position tracking error and thrust change rate of the manned flying saucer (for the specific process, refer to step A5 above);

[0153] The thrust execution module 6 is used to send the calculated thrust instructions of each normal power unit to the corresponding power unit execution mechanism (for the specific process, please refer to step A6 above).

[0154] The above description is merely an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, various modifications and variations of the present application are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A manned flying saucer attitude control method, applied to a manned flying saucer with multiple power units, characterized in that: The steps of the method include: A1. Monitor the thrust command and actual thrust output of each power unit to determine whether any power unit has a response delay or a stuck fault. A2. If a faulty unit exists, the thrust output characteristics of the faulty unit are identified based on the relevant data stream of the faulty unit, and the thrust information of the faulty unit is obtained. The relevant data stream includes the data stream of the thrust command and actual thrust output of the faulty unit, as well as the data stream of the manned flying saucer's motion status data. A3. Determine the control mode for the faulty unit based on the thrust information of the faulty unit; A4. Extract the dynamic constraint models and operational constraint information of the remaining normal power units from the local database. The dynamic constraint models include mathematical models of the normal power unit's thrust output characteristics, response delay, and physical coupling effects. The operational constraint information includes the reference thrust range and reference thrust change rate range of the normal power unit. A5. With the goal of minimizing the attitude tracking error, position tracking error, and thrust change rate of the manned vehicle, solve the thrust command for each normal power unit based on the vehicle's real-time attitude, real-time position, desired attitude, and desired position, as well as the thrust information and control mode of the faulty unit, the operational constraint information of the normal power unit, and the dynamic constraint model of the normal power unit. A6. Send the calculated thrust instructions of each normal power unit to the corresponding power unit actuator.

2. A manned flying saucer attitude control method according to claim 1, characterized in that: Step A1 includes: A101. Periodically obtain the thrust command and actual thrust output of each power unit at a preset frequency; A102. For each power unit, calculate the absolute value of the deviation between the actual thrust output and the thrust command in the current cycle, recording this as the first deviation. A103. If the first deviation exceeds the preset deviation threshold, the cumulative duration of the first deviation exceeding the preset deviation threshold; A104. If the duration exceeds a preset delay time threshold, the corresponding power unit is determined to be a faulty unit with a response delay fault; A105. For each power unit, monitor whether the actual thrust output of the power unit changes within a preset observation time after receiving a new thrust command. If no change occurs, determine that the corresponding power unit is a faulty unit with a stuck fault.

3. A manned flying saucer attitude control method according to claim 1, characterized in that: Step A2 includes: A201. Extract the thrust command sequence of the faulty unit, the actual thrust output sequence, and the motion state data sequence of the manned flying saucer; the motion state data sequence includes an attitude angular velocity sequence and a moving speed sequence; A202. Call a thrust output characteristic identification model pre-established based on the least squares method; the thrust output characteristic identification model takes thrust command, attitude angular velocity, and position velocity as input, and actual thrust output as output, and includes three parameters to be identified: thrust proportional coefficient, hysteresis time, and hysteresis threshold; A203. Using the recursive least squares method, based on the extracted thrust command sequence, actual thrust output sequence, attitude angular velocity sequence, and movement velocity sequence, the parameters to be identified in the thrust output characteristic identification model are updated in real time to obtain the real-time thrust proportional coefficient, hysteresis time, and stuck threshold of the faulty unit. A204. Calculate the nominal thrust value of the faulty unit based on the real-time thrust proportional coefficient of the faulty unit. A205. Determine whether the actual thrust output of the faulty unit is affected by hysteresis or sticking based on the hysteresis time and stuckness threshold of the faulty unit. If so, modify the nominal thrust value to obtain the effective thrust value of the faulty unit. Otherwise, use the nominal thrust value as the effective thrust value of the faulty unit. A206. The real-time thrust proportional coefficient, hysteresis time, stuck threshold and effective thrust value are used as the thrust information of the faulty unit.

4. A manned flying saucer attitude control method according to claim 3, characterized in that: Step A205 includes: Calculate the time difference between the moment the thrust command changes and the moment the actual thrust begins to respond. If the time difference is greater than the hysteresis time of the faulty unit, then determine that the faulty unit is affected by hysteresis. Monitor the change in actual thrust output over a number of consecutive preset observation periods. If the change is less than a stuck threshold of the faulty unit, the faulty unit is determined to be affected by stuckness. If the faulty unit is only affected by hysteresis, then based on the time difference, a pre-calibrated hysteresis compensation model is used to perform time compensation on the nominal thrust value to obtain a corrected thrust value; If the faulty unit is affected by sticking, the average value of the actual thrust output during several consecutive preset observation periods is used as the corrected thrust value; If the faulty unit is not affected by hysteresis or sticking, the original nominal thrust value is used as the corrected thrust value; The corrected thrust value is taken as the effective thrust value of the faulty unit.

5. A manned flying saucer attitude control method according to claim 4, characterized in that: Step A3 includes: A301. According to the thrust information of the faulty unit, determine the state type of the faulty unit; the state type includes a complete failure state, a response hysteresis state, and a stuck state; A302. If the status type of the faulty unit is a complete failure state, the control mode of the faulty unit is set to the disabled mode and the thrust output of the faulty unit is turned off; A303. If the faulty unit is in a delayed response state, assess whether the faulty unit has any remaining thrust capability based on the real-time thrust proportional coefficient in the thrust information. A304. If the faulty unit has residual thrust capacity, set the control mode of the faulty unit to hysteresis compensation mode. Based on the hysteresis time in the thrust information, use a pre-calibrated hysteresis compensation model to compensate the thrust command of the faulty unit and send the compensated thrust command to the actuator of the faulty unit. A305. If the faulty unit has no remaining thrust capacity, set the control mode of the faulty unit to the disabled mode and shut down the thrust output of the faulty unit; A306. If the state type of the faulty unit is a stuck state, set the control mode of the faulty unit to a fixed thrust mode, and fix the thrust output of the faulty unit to the effective thrust value in the thrust information.

6. A manned flying saucer attitude control method according to claim 5, characterized in that: Step A301 includes: If the real-time thrust proportional coefficient is lower than the preset failure threshold and the duration exceeds the preset failure time threshold, the state type of the faulty unit is determined to be a complete failure state; When the state type of the faulty unit is not a complete failure state, if the stuck threshold is lower than the preset stuck amplitude threshold, the state type of the faulty unit is determined to be a stuck state; otherwise, the state type of the faulty unit is determined to be a response hysteresis state.

7. A manned flying saucer attitude control method according to claim 5, characterized in that: Step A5 includes: A501. Construct a comprehensive objective function that includes an attitude tracking error term, a position tracking error term, and a thrust adjustment term. The attitude tracking error term and the position tracking error term are calculated based on the real-time attitude, real-time position, desired attitude, and desired position of the manned vehicle. The thrust adjustment term is calculated based on the thrust command change rate of each normal power unit. A502. Based on the thrust information and control mode of the faulty unit, calculate the net force and net torque generated by the faulty unit on the manned flying vehicle. Apply this as an external disturbance force to the dynamic model of the manned flying vehicle, thereby obtaining a dynamic model of the flying vehicle that takes into account the thrust disturbance caused by the faulty unit. A503. Determine the thrust feasible region of each normal power unit based on the dynamic constraint model and operational constraint information of the normal power unit; the thrust feasible region defines the available thrust range and available thrust change rate range of each normal power unit; A504. Taking the comprehensive objective function as the optimization target, the UFO dynamics model taking into account the thrust interference of the faulty unit and the thrust feasible domain as constraints, the quadratic programming method is used to solve and obtain the thrust instructions of each normal power unit that meets the constraints and minimizes the comprehensive objective function.

8. A manned flying saucer attitude control method according to claim 7, characterized in that: Step A503 includes: B1. For each normal power unit, based on its dynamic constraint model, establish a nonlinear function with thrust command as input and attitude angular acceleration and position acceleration as output; B2. Using the operational constraints of each normal power unit as a constraint, combined with the nonlinear function, by solving the constrained optimization problem, calculate the feasible range of the attitude angular acceleration and position acceleration of each normal power unit under the constraints of its own manned flying saucer; B3. Perform vector superposition of the feasible ranges of attitude angular acceleration and position acceleration corresponding to all normal power units to obtain the feasible region of the overall attitude angular acceleration and overall position acceleration of the manned flying saucer; B4. Based on the feasible domain of the overall attitude angular acceleration and the overall position acceleration, inversely solve the thrust range of each normal power unit. Based on the inverse solution, determine the available thrust range and the available thrust change rate range to obtain the thrust feasible domain.

9. A manned flying saucer attitude control method according to claim 8, characterized in that: Step A504 includes: C1. Based on the real-time attitude, real-time position, desired attitude and desired position of the manned flying saucer, according to the comprehensive objective function, determine the desired attitude angular acceleration and desired position acceleration of the manned flying saucer; C2. Based on the UFO dynamics model that takes into account the thrust interference of the faulty unit, establish the functional relationship between the desired attitude angular acceleration and the desired position acceleration of the manned UFO and the thrust command of each normal power unit; C3. Based on the functional relationship and the thrust feasible region, construct a quadratic programming problem with the thrust command of each normal power unit as the optimization variable; wherein the objective function of the quadratic programming problem is to minimize the error between the actual thrust of each normal power unit and the desired thrust obtained by inversely solving the desired attitude angular acceleration and the desired position acceleration, with the constraint that the thrust command of each normal power unit is within the thrust feasible region; C4. Solve the quadratic programming problem to obtain thrust instructions for each normal power unit that meets the constraints and minimizes the comprehensive objective function.

10. A manned flying saucer attitude control system, applied to a manned flying saucer with multiple power units, characterized in that: The system includes: A fault detection module is used to monitor the thrust command and actual thrust output of each power unit and determine whether there is a faulty unit with response hysteresis or stuck fault in each power unit; A thrust characteristic identification module is used to identify the thrust output characteristics of a faulty unit based on the relevant data stream of the faulty unit when a faulty unit exists, and obtain thrust information of the faulty unit; the relevant data stream includes the data stream of the thrust command and actual thrust output of the faulty unit, and the data stream of the motion state data of the manned flying saucer; A control mode determination module, configured to determine a control mode for the faulty unit based on thrust information of the faulty unit; a data extraction module, configured to extract the dynamic constraint model and operational constraint information of the remaining normal power units from a local database; the dynamic constraint model includes a mathematical model of the thrust output characteristics, response delay, and physical coupling effects of the normal power units; and the operational constraint information includes a reference thrust range and a reference thrust change rate range of the normal power units; A thrust optimization module is used to minimize the attitude tracking error, position tracking error and thrust change rate of the manned flying saucer, and solve the thrust command of each normal power unit based on the real-time attitude, real-time position, expected attitude and expected position of the manned flying saucer, as well as the thrust information of the faulty unit, the control mode of the faulty unit, the operation constraint information of the normal power unit and the dynamic constraint model of the normal power unit; The thrust execution module is used to send the calculated thrust instructions of each normal power unit to the corresponding power unit execution mechanism.

Images