Multimodal flight control system for manned vehicles
By constructing a unified energy potential well across the entire envelope and a dynamic control allocation module that senses energy temperature, a single scalar energy potential well function is generated and gradient processing is performed. This solves the problem of discontinuous control commands during the transition phase of a multi-mode aircraft and enables continuous differentiable control and actuator load balancing across the entire envelope.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- NANJING KUAILUN INTELLIGENT TECH CO LTD
- Filing Date
- 2026-04-17
- Publication Date
- 2026-07-14
AI Technical Summary
Existing flight control systems for multi-mode aircraft suffer from torque discontinuities or high-frequency jitter in control commands during the transition phase, making it impossible to achieve continuous differentiability across the entire flight envelope.
By employing a unified energy potential well construction module for the entire envelope, a unified outer loop control law module based on the energy potential well gradient, and a dynamic control allocation module based on energy and temperature sensing, a single scalar energy potential well function is generated. A unified linear acceleration command is generated through the energy gradient and time derivative. The actuator control command is allocated by combining the temperature and power penalty matrix, thereby achieving continuous differentiable control within the entire envelope.
It achieves continuous differentiability of control commands throughout the vertical takeoff and landing, transition phase, and level flight cruise envelope, avoiding mode switching logic and gain scheduling tables, automatically balancing the thermal and power loads of actuators, and possessing fault tolerance capability for partial actuator failure.
Smart Images
Figure CN122387115A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to aerospace science and technology, and more particularly to a multimodal flight control system for manned aircraft. Background Technology
[0002] Multimodal flight control systems for manned aircraft are widely used in electric vertical takeoff and landing (EVTOL) aircraft, tiltrotor aircraft, and lift-cruise hybrid aircraft. These types of aircraft need to frequently switch between multiple flight modes, such as vertical takeoff and landing, transition phases, and level flight cruise, in order to achieve urban air traffic or regional short-distance transportation functions.
[0003] In existing technologies, flight control systems for the aforementioned multimodal aircraft typically employ a sub-modal control law design scheme. This involves designing separate control laws for the vertical takeoff and landing (VTOL) mode based on rotors or ducted fans, and for the level flight and cruise mode based on fixed-wing aerodynamic control surfaces. During transitions, switching between these different control laws is achieved through predefined gain scheduling tables, throttle curves, or tilt angle curves. Furthermore, existing schemes generally assign fixed priorities to each actuator in the control efficiency matrix or use a weighted pseudo-inverse method for control allocation.
[0004] However, in the process of implementing the above technical solution, the inventors of this application discovered that the prior art has at least the following technical problems: due to the essential differences in the mathematical form and gain level actuator allocation logic of the control laws corresponding to different flight modes, the existing solutions inevitably have torque discontinuities or high-frequency jitter in the transition section. Even if a dense gain scheduling table or a complex hybrid control law is used, it can only achieve segmented continuity and cannot achieve continuous differentiability within the entire envelope. Summary of the Invention
[0005] To overcome the above shortcomings, the present invention provides a multimodal flight control system for manned aircraft, which aims to improve the problem that existing solutions inevitably experience torque discontinuities or high-frequency jitter in the transition phase of control commands, resulting in the inability to achieve continuous differentiability across the entire flight envelope.
[0006] In a first aspect, the present invention provides the following technical solution: a multi-modal flight control system for a manned aircraft, comprising a unified energy potential well construction module with a full envelope, a unified outer-loop control law module based on the energy potential well gradient, and a dynamic control allocation module with energy temperature sensing, connected in sequence, wherein:
[0007] The unified energy potential well construction module of the full envelope generates a single scalar energy potential well function in real time based on the current pressure altitude, airspeed, real-time total power of all lift / thrust actuators, and weighting coefficients that vary smoothly with airspeed.
[0008] The unified outer-loop control law module based on the energy potential well gradient generates a unified linear acceleration command based on the gradient and time derivative of the single scalar energy potential well function.
[0009] The dynamic control allocation module with energy and temperature sensing receives the unified linear acceleration command and the total force and torque command generated by the attitude inner loop. It uses a quadratic programming solution with real-time temperature and energy penalty weights to output control commands for each lift / thrust actuator.
[0010] Preferably, the process by which the full-envelope unified energy potential well construction module generates a single scalar energy potential well function includes:
[0011] Real-time data collection includes current pressure altitude, target altitude, airspeed, economic cruise target airspeed, and the real-time total power of all lift / thrust actuators;
[0012] By employing altitude weighting coefficients, velocity weighting coefficients, and power weighting coefficients that vary continuously with airspeed, the squared terms of altitude error, velocity error, and real-time total power are weighted and summed to obtain the single scalar energy potential well function.
[0013] Preferably, the altitude weighting coefficient and the velocity weighting coefficient achieve a smooth and continuous change through a hyperbolic tangent function with airspeed as the variable, and the power weighting coefficient achieves a smooth and continuous change through a Gaussian decay function with airspeed as the variable, thereby making the single scalar energy potential well function at least second-order differentiable throughout the entire flight envelope.
[0014] Preferably, the process by which the unified outer-loop control law module based on the energy potential well gradient generates the unified linear acceleration command includes:
[0015] Calculate the gradient in the altitude direction and the gradient in the airspeed direction for the single scalar energy potential well function, and differentiate them with respect to time.
[0016] By multiplying the altitude gradient, airspeed gradient, and their time derivative by the proportional coefficient and differential coefficient that vary monotonically only with dynamic pressure, respectively, and then taking the negative values, a unified linear acceleration command for the altitude channel and the forward channel is obtained.
[0017] Preferably, the process of the energy temperature sensing dynamic control allocation module performing quadratic programming solution includes:
[0018] Real-time data collection of current temperature, temperature rise rate, and remaining available power of each lift / thrust actuator; construction of a time-varying temperature energy penalty diagonal matrix.
[0019] By adding the product term of the temperature-energy penalty diagonal matrix and the square of the actuator command vector to the quadratic programming objective function, the control commands of each lift / thrust actuator obtained by the solution will automatically tend to use actuators with lower temperature and larger remaining available power.
[0020] Preferably, each diagonal element of the temperature energy penalty diagonal matrix is a weighted sum of the normalized value of the current temperature relative to the ambient temperature, the reciprocal of the normalized value of the remaining available power, and the positive portion of the temperature rise rate.
[0021] Preferably, when the system detects a sudden drop in the control efficiency of any lift / thrust actuator, it automatically adjusts the upper limit of its control command to the currently achievable value.
[0022] Preferably, in the unified outer loop control law module based on the energy potential well gradient, the proportional coefficient and differential coefficient used to generate the unified linear acceleration command adopt a scheduling method that changes linearly and monotonically with dynamic pressure.
[0023] Preferably, the attitude inner loop uses a dual-loop PID structure of angular rate and attitude angle to generate total force and torque commands.
[0024] Preferably, the lift / thrust actuator includes multiple rotors, ducted fans, aerodynamic control surfaces, and vector nozzles, and its control efficiency matrix is updated in real time by lookup table interpolation based on the angle of attack and Mach number.
[0025] The present invention has the following beneficial effects:
[0026] 1. This invention uses a single scalar energy potential well function and its gradient and time derivative to directly generate a unified linear acceleration command. The system does not require mode switching logic and gain scheduling table within the full envelope of vertical takeoff and landing, transition zone and level flight cruise, so that the control command remains continuously differentiable within the full envelope.
[0027] 2. This invention introduces a temperature energy penalty diagonal matrix that changes in real time with temperature and remaining available power in the objective function of the secondary programming control allocation. Under the premise of satisfying the force and torque commands, the system automatically allocates the load to the actuator with a lower temperature and a larger remaining power margin, thereby achieving dynamic balance between the thermal load and power load of the actuator.
[0028] 3. When the actuator control efficiency is detected to suddenly drop, the present invention automatically adjusts its command upper limit to the current actual achievable value, and the control efficiency matrix is updated in real time with the angle of attack, Mach number and dynamic pressure. The system can still maintain the effectiveness of control when the actuator fails or the aerodynamic parameters change significantly. Attached Figure Description
[0029] Figure 1 This is a diagram of the multimodal flight control system for the manned aircraft proposed in this invention. Detailed Implementation
[0030] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0031] Reference Figure 1 In the first embodiment of the present invention, a multimodal flight control system for a manned aircraft is provided, comprising a unified energy potential well construction module with a full envelope, a unified outer loop control law module based on the energy potential well gradient, and a dynamic control allocation module with energy temperature sensing, connected in sequence, wherein:
[0032] The unified energy potential well construction module of the full envelope generates a single scalar energy potential well function in real time based on the current pressure altitude, airspeed, real-time total power of all lift / thrust actuators, and weighting coefficients that vary smoothly with airspeed.
[0033] The unified outer-loop control law module based on the energy potential well gradient generates a unified linear acceleration command based on the gradient and time derivative of a single scalar energy potential well function.
[0034] The dynamic control allocation module with energy and temperature sensing receives unified linear acceleration commands and total force and torque commands generated by the attitude inner loop. It uses quadratic programming with real-time temperature-energy penalty weights to solve the problem and outputs control commands for each lift / thrust actuator, thereby achieving full-envelope continuous control without the need for mode switching logic.
[0035] Specifically, this invention proposes a multimodal flight control system for manned aircraft. The system includes three functional modules connected in sequence: a unified energy potential well construction module with full envelope, a unified outer loop control law module based on energy potential well gradient, and a dynamic control allocation module with energy temperature sensing. It is supplemented by an attitude inner loop, a sensor data acquisition unit, and an actuator execution unit.
[0036] The unified energy potential well construction module receives power feedback signals in real time from the barometric altimeter, pitot tube, inertial navigation system, and various lift / thrust actuators. By using altitude weighting coefficients, velocity weighting coefficients, and power weighting coefficients that vary smoothly with airspeed, it constructs a single scalar energy potential well function covering the entire envelope of vertical takeoff and landing, transition phase, and level flight cruise. The unified outer loop control law module based on the energy potential well gradient calculates the gradient and time derivative of this single scalar energy potential well function to generate unified linear acceleration commands for the altitude and forward channels, which together with the traditional attitude inner loop constitute a complete guidance command. The energy and temperature sensing dynamic control allocation module converts the unified linear acceleration command into total force and torque commands, and then uses a quadratic programming method with a real-time temperature-energy penalty diagonal matrix to rationally allocate the control commands to multiple lift / thrust actuators such as rotors, ducted fans, aerodynamic control surfaces, or vector nozzles.
[0037] The three modules work together to enable the aircraft to achieve complete continuity of control commands, automatic balancing of actuator load and thermal load, and passive fault tolerance for sudden drops in actuator efficiency throughout the entire flight envelope without any mode switching logic, gain scheduling table, or predefined transition lift distribution curve. This solves long-standing technical problems in traditional multimodal flight control systems, such as discontinuous torque in the transition section, high parameter sensitivity, and easy overheating during continuous takeoff and landing.
[0038] The unified energy potential well construction module of the full envelope receives pressure altitude h and air velocity V. a Target height Economic cruise target airspeed and the real-time total power of all lift / thrust actuators The single scalar energy potential well function E is obtained through real-time calculation.
[0039] The unified outer-loop control law module based on the energy potential well gradient calculates the gradient and time derivative of the single scalar energy potential well function E, generating a height channel unified linear acceleration command. Unified linear acceleration command with forward channel .
[0040] The energy-temperature sensing dynamic control distribution module receives the total force and torque commands output from the attitude inner loop. It also provides a unified linear acceleration command, and through a quadratic programming solver with a real-time temperature-energy penalty diagonal matrix, outputs normalized control commands u* for each lift / thrust actuator.
[0041] The system eliminates mode switching logic, gain scheduling tables, and predefined transition curves throughout the entire flight envelope.
[0042] The formula for calculating the single scalar energy potential well function E is: (1)
[0043] In the formula,
[0044] It is the acceleration due to gravity;
[0045] For variable altitude weighting coefficients related to airspeed;
[0046] These are variable speed weighting coefficients related to airspeed;
[0047] For airspeed-related variable power weighting coefficients;
[0048] h represents the current air pressure altitude;
[0049] The target altitude set for the current mission phase;
[0050] Airspeed;
[0051] The preset target airspeed for economical cruising;
[0052] This is the sum of the power feedback from all lift / thrust actuators, normalized to the rated power, and ranges from 0 to 1.
[0053] Height weighting coefficient The calculation formula is:
[0054] ;
[0055] Speed weighting coefficient The real-time calculation formula is as follows:
[0056] ;
[0057] Power weighting coefficient The real-time calculation formula is as follows:
[0058] ;
[0059] The symbols in the formula are defined as follows:
[0060] , The pre-defined maximum weight value;
[0061] Predefined base power weights;
[0062] The power weight increment is a pre-calibrated value.
[0063] The kurtosis coefficient of the hyperbolic tangent function is pre-calibrated;
[0064] The characteristic airspeed for the transition section is pre-calibrated;
[0065] This is the pre-calibrated Gaussian attenuation width coefficient.
[0066] The above method of calculating weighting coefficients ensures that the single scalar energy well function E is at least second-order continuously differentiable within the entire envelope.
[0067] The system converts uniform linear acceleration commands into total force and torque commands through the attitude inner loop. Then, the dynamic control and distribution module that senses energy and temperature completes the final actuator command distribution, thus forming a complete closed-loop control link.
[0068] The system hardware platform includes a flight control computer, an inertial measurement unit, a barometric altimeter, a pitot tube, temperature sensors built into each lift / thrust actuator, and servo drivers. All sensor data and actuator status feedback are connected to the flight control computer via ARINC429 or CAN bus. The normalized control commands u* output by the flight control computer are sent to each servo driver via PWM signals or CAN bus.
[0069] Furthermore, the process of generating a single scalar energy well function using the unified energy well construction module for the entire envelope includes:
[0070] Real-time data collection includes current pressure altitude, target altitude, airspeed, economic cruise target airspeed, and the real-time total power of all lift / thrust actuators;
[0071] By using the altitude weighting coefficient, velocity weighting coefficient, and power weighting coefficient that vary continuously with airspeed, the squared terms of altitude error, velocity error, and real-time total power are weighted and summed to obtain a single scalar energy potential well function.
[0072] The altitude and speed weighting coefficients achieve smooth and continuous variation through a hyperbolic tangent function with airspeed as the variable, and the power weighting coefficient achieves smooth and continuous variation through a Gaussian decay function with airspeed as the variable, thus making the single scalar energy potential well function at least second-order differentiable throughout the entire flight envelope.
[0073] Specifically, it collects the current air pressure altitude h and the target altitude set for the current mission phase in real time. airspeed V a Economic cruise target airspeed The real-time power feedback signals of all lift / thrust actuators are added together after being normalized to their respective rated power to obtain the normalized total power. .
[0074] Using an altitude weighting factor that varies continuously with airspeed Speed weighting coefficient and power weighting coefficient For the squared term of height error , speed error squared term and the normalized total power square term By performing a weighted summation, we obtain the single scalar energy well function E. The formula for calculating the single scalar energy well function E is formula (1).
[0075] The above weighting coefficient calculation process makes the height weighting coefficient... Approaching at low speeds The speed weighting coefficient approaches 0 at high speeds. Approaching 0 at low speeds and approaching 0 at high speeds. Power weighting coefficient It reaches its maximum value in the low-speed range and decreases monotonically as airspeed increases.
[0076] The above weighting coefficient calculation method ensures that the single scalar energy potential well function E and its first and second derivatives are continuous throughout the entire flight envelope. The unified energy potential well construction module of the entire envelope completes all calculations and outputs the real-time single scalar energy potential well function E to the unified outer loop control law module based on the energy potential well gradient.
[0077] Furthermore, the process of generating unified linear acceleration commands by the unified outer-loop control law module based on the energy potential well gradient includes:
[0078] Calculate the gradient along the altitude and the gradient along the airspeed for a single scalar energy potential well function, and then differentiate them with respect to time.
[0079] By multiplying the altitude gradient, airspeed gradient, and their time derivative by the proportional coefficient and differential coefficient that vary monotonically only with dynamic pressure, respectively, and then taking the negative values, a unified linear acceleration command for the altitude channel and the forward channel is obtained.
[0080] Specifically, the altitude gradient and airspeed gradient, as well as their time derivatives, are calculated for the single scalar energy potential well function E. The altitude gradient, airspeed gradient, and their time derivatives are then multiplied by the proportional coefficient and differential coefficient, which vary monotonically only with dynamic pressure, respectively, and the negative values are taken to obtain the uniform linear acceleration command for the altitude channel. Unified linear acceleration command with forward channel .
[0081] The gradient in the height direction is The gradient in the airspeed direction is .
[0082] The formula for calculating the uniform linear acceleration command for the height channel is:
[0083] ;
[0084] The formula for calculating the forward channel uniform linear acceleration command is:
[0085] ;
[0086] The symbols in the formula are defined as follows:
[0087] The dynamic pressure is calculated in real time. Atmospheric density;
[0088] , , , These are the proportional and differential coefficients that vary monotonically with dynamic pressure;
[0089] The rate of change of air pressure at altitude is provided directly by the inertial navigation system.
[0090] The airspeed rate of change is obtained by differentiating the airspeed tube or integrating the acceleration of the inertial navigation system.
[0091] The dynamic pressure regulation relationship between the proportional coefficient and the differential coefficient is expressed as a linear function:
[0092] ;
[0093] ;
[0094] ;
[0095] ;
[0096] In the formula, , , , For reference dynamic pressure The reference gain below, 、k_{dz}、 , For the pre-calibrated rate of change of gain, Use standard sea-level dynamic pressure.
[0097] To prevent excessively low gain in the low dynamic voltage region, lower limits are set for the four coefficients in actual implementation:
[0098] ;
[0099] The same applies to the other three coefficients.
[0100] The calculations of the gradient and time derivative described above are completed within each control cycle. and The formula E is obtained by direct analytical differentiation, avoiding the noise and delay introduced by numerical differentiation, and unifying the output of the outer loop control law module. and To the inner attitude loop, the inner attitude loop according to and And the current attitude angle command generates total force and torque commands. In this way, the linear acceleration commands of the altitude channel and the forward channel remain continuously differentiable across the entire envelope from hovering to cruise, without the need to switch control laws or consult the gain scheduling table.
[0101] Furthermore, the process of the energy-temperature sensing dynamic control allocation module performing quadratic programming solutions includes:
[0102] Real-time data collection of current temperature, temperature rise rate, and remaining available power of each lift / thrust actuator; construction of a time-varying temperature energy penalty diagonal matrix.
[0103] By adding a product term of the temperature-energy penalty diagonal matrix and the square of the actuator command vector to the quadratic programming objective function, the control commands of each lift / thrust actuator obtained by the solution will automatically tend to use actuators with lower temperature and larger remaining available power.
[0104] Each diagonal element of the temperature energy penalty diagonal matrix consists of a weighted sum of the normalized value of the current temperature relative to the ambient temperature, the reciprocal of the normalized value of the remaining available power, and the positive part of the rate of temperature rise.
[0105] Specifically, the current temperature, rate of temperature rise, and remaining available power of each lift / thrust actuator are collected in real time. Construct a time-varying temperature-energy penalty diagonal matrix Q(t);
[0106] By adding a product term of the temperature-energy penalty diagonal matrix Q(t) and the square of the actuator command vector to the quadratic programming objective function, the control commands of each lift / thrust actuator obtained by the solution will automatically tend to use actuators with lower temperature and larger remaining available power.
[0107] Each diagonal element of the temperature-energy penalty diagonal matrix consists of a weighted sum of the normalized value of the current temperature relative to the ambient temperature, the reciprocal of the normalized value of the remaining available power, and the positive part of the rate of temperature rise.
[0108] The quadratic programming problem is expressed as:
[0109] ;
[0110] Subject to:
[0111] ;
[0112] in, The control efficiency matrix is given by the real-time dynamic pressure q, angle of attack α, and Mach number. The value is obtained by looking up a pre-stored 3D table and performing linear interpolation.
[0113] This is the total force and torque command vector output by the inner loop of the attitude control;
[0114] W and R are fixed diagonal weighted matrices;
[0115] λ is the scalar coefficient for the control quantity penalty;
[0116] The actuator command vector is an m-dimensional normalized vector, with rotor speed, control surface deflection angle, and vector nozzle angle all normalized to [-1,1] or [0,1].
[0117] , These are the normalized values for the physical limits of each actuator;
[0118] , This represents the maximum permissible rate of change for each actuator;
[0119] This is the duration of the current control cycle;
[0120] This is the actuator instruction from the previous cycle.
[0121] The formula for calculating the i-th diagonal element Q_{ii}(t) of the temperature-energy penalty diagonal matrix Q(t) is as follows:
[0122] ;
[0123] in:
[0124] The real-time temperature of the i-th actuator;
[0125] The current ambient temperature;
[0126] This is the maximum allowable temperature for the actuator;
[0127] This represents the current remaining available power of the i-th actuator;
[0128] The rated power of the i-th actuator;
[0129] The temperature change rate of the i-th actuator is obtained from the difference between two adjacent temperature samples;
[0130] , , These are pre-calibrated weighting coefficients, ranging from 0.1 to 1.0.
[0131] When the temperature of a certain actuator is close to Or when the remaining available power is close to 0, The corresponding increase reduces the actuator's command amplitude while satisfying the force and torque commands. The quadratic programming is solved in real time in the flight control computer using the active set method or interior point method. The solution obtained is... It outputs directly to the underlying servo driver of each lift / thrust actuator.
[0132] By constructing and adding the temperature-energy penalty diagonal matrix Q(t) in real time, the control allocation achieves dynamic balance between actuator thermal load and power load while ensuring the accuracy of force and torque tracking.
[0133] The quadratic programming solver adopts a sparse QP solution algorithm based on the active set method, which is applied to actuators with a size of m≤16 on the flight control computer.
[0134] Furthermore, when the system detects a sudden drop in the control efficiency of any lift / thrust actuator, it automatically adjusts the upper limit of its control command to the current actual achievable value; in the unified outer loop control law module based on the energy potential well gradient, the proportional coefficient and differential coefficient used to generate the unified linear acceleration command adopt a scheduling method that changes linearly and monotonically with dynamic pressure.
[0135] Specifically, the process for detecting sudden drops in control efficiency and automatically adjusting the command limit is as follows:
[0136] The energy-temperature sensing dynamic control allocation module adjusts the control efficiency matrix in each control cycle. Norm monitoring is performed on each column vector. The real-time efficiency scalar of the i-th actuator is defined as:
[0137] ;
[0138] in,
[0139] This refers to the i-th column of the current cycle control efficiency matrix B;
[0140] This is a pre-stored nominal efficiency column vector, representing ground calibration or first flight calibration values.
[0141] || represents the Euclidean norm.
[0142] When three consecutive control cycles meet the requirements When the i-th actuator experiences a sudden drop in control efficiency, it is determined that the i-th actuator has experienced a sudden drop in control efficiency.
[0143] Once the determination is successful, immediately set the upper limit of the actuator's command range. Updated to:
[0144] ;
[0145] In the formula, This is the nominal physical upper limit.
[0146] Updated This is directly applied as a secondary planning constraint in this cycle and subsequent cycles, until efficiency is restored or ground maintenance is performed. When efficiency recovers to... When ≥0.95, automatically Restore to The above process does not require a dedicated fault detection and isolation module; passive fault tolerance can be achieved solely using the real-time interpolation results of the control efficiency matrix B.
[0147] Furthermore, the attitude inner loop uses a dual-loop PID structure of angular rate and attitude angle to generate total force and torque commands; the lift / thrust actuator includes multiple rotors, ducted fans, aerodynamic control surfaces, and vector nozzles, and its control efficiency matrix is updated in real time by lookup table interpolation based on the angle of attack and Mach number.
[0148] Specifically, the attitude inner loop is composed of an outer loop attitude angle PID and an inner loop angular rate PID cascaded together.
[0149] The attitude angle error vector is defined as:
[0150] ;
[0151] In the formula, , , Each is controlled by a unified linear acceleration command. , It is obtained through small angle approximation or table lookup transformation.
[0152] Outer ring attitude angle PID output desired angular rate command:
[0153] ;
[0154] The inner-loop angular rate PID directly outputs torque commands for the roll, pitch, and yaw channels based on the deviation between the actual angular rates p, q, and r and the desired angular rate command. , , :
[0155] ;
[0156] The attitude inner loop simultaneously converts the uniform linear acceleration command into a vertical force command. and forward force command This ultimately forms a six-dimensional total force and torque command vector:
[0157] ;
[0158] The integral term employs a windmill integrator with anti-saturation properties. The lift / thrust actuator consists of m actuators, including either a rotor or a ducted fan, and either aerodynamic control surfaces or vectoring nozzles. Control efficiency matrix. It is 6×m dimensional, with each row corresponding to one degree of freedom. Each column corresponds to the force and torque contribution of an actuator.
[0159] The matrix elements are obtained through ground wind tunnel tests, CFD calculations, or first flight parameters, forming a three-dimensional data table.
[0160] The real-time update process is as follows:
[0161] Read the current angle of attack α and Mach number Ma;
[0162] exist and Perform bilinear interpolation to obtain the nominal efficiency matrix at the current angle of attack and Mach number. ;
[0163] Multiply by real-time dynamic pressure q and reference dynamic pressure The ratio of these ratios yields the final control efficiency matrix:
[0164] ;
[0165] Interpolation and dynamic compression are performed within each control cycle.
[0166] By using the aforementioned attitude inner-loop dual-loop PID structure and real-time lookup table interpolation update of the control efficiency matrix, the system maintains a precise mapping from force and torque commands to actuator commands across the entire envelope.
[0167] After the cycle begins, the sensor data acquisition, the calculation of the unified energy potential well construction module for the entire envelope, the calculation of the unified outer loop control law module based on the energy potential well gradient, and the angle of attack / Mach number / dynamic pressure interpolation update of the control efficiency matrix B are completed. The attitude inner loop runs independently and in parallel. The remaining time is dedicated to solving the quadratic programming problem. The actuator command u* is output and sent to each servo drive no later than 1ms before the end of the cycle.
[0168] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A multi-modal flight control system for a manned aircraft, characterized in that, It includes a fully enveloped unified energy potential well construction module, a unified outer loop control law module based on the energy potential well gradient, and an energy temperature sensing dynamic control allocation module connected in sequence, wherein: The unified energy potential well construction module of the full envelope generates a single scalar energy potential well function in real time based on the current pressure altitude, airspeed, real-time total power of all lift / thrust actuators, and weighting coefficients that vary smoothly with airspeed. The unified outer-loop control law module based on the energy potential well gradient generates a unified linear acceleration command based on the gradient and time derivative of the single scalar energy potential well function. The dynamic control allocation module with energy and temperature sensing receives the unified linear acceleration command and the total force and torque command generated by the attitude inner loop. It uses a quadratic programming solution with real-time temperature and energy penalty weights to output control commands for each lift / thrust actuator.
2. The multi-modal flight control system for a manned aircraft according to claim 1, characterized in that, The process by which the full envelope unified energy potential well construction module generates a single scalar energy potential well function includes: Real-time data collection includes current pressure altitude, target altitude, airspeed, economic cruise target airspeed, and the real-time total power of all lift / thrust actuators; By employing altitude weighting coefficients, velocity weighting coefficients, and power weighting coefficients that vary continuously with airspeed, the squared terms of altitude error, velocity error, and real-time total power are weighted and summed to obtain the single scalar energy potential well function.
3. The multi-modal flight control system for a manned aircraft according to claim 2, characterized in that, The altitude weighting coefficient and velocity weighting coefficient achieve smooth and continuous variation through a hyperbolic tangent function with airspeed as the variable, and the power weighting coefficient achieves smooth and continuous variation through a Gaussian decay function with airspeed as the variable, thereby making the single scalar energy potential well function at least second-order differentiable throughout the entire flight envelope.
4. The multi-modal flight control system for a manned aircraft according to claim 3, characterized in that, The process by which the unified outer-loop control law module based on the energy potential well gradient generates unified linear acceleration commands includes: Calculate the gradient in the altitude direction and the gradient in the airspeed direction for the single scalar energy potential well function, and differentiate them with respect to time. By multiplying the altitude gradient, airspeed gradient, and their time derivative by the proportional coefficient and differential coefficient that vary monotonically only with dynamic pressure, respectively, and then taking the negative values, a unified linear acceleration command for the altitude channel and the forward channel is obtained.
5. The multi-modal flight control system for a manned aircraft according to claim 1, characterized in that, The process of the energy temperature sensing dynamic control allocation module performing quadratic programming solution includes: Real-time data collection of current temperature, temperature rise rate, and remaining available power of each lift / thrust actuator; construction of a time-varying temperature energy penalty diagonal matrix. By adding the product term of the temperature energy penalty diagonal matrix and the square of the actuator command vector to the quadratic programming objective function, the control commands of each lift / thrust actuator obtained by the solution will automatically tend to use actuators with lower temperature and larger remaining available power.
6. The multi-modal flight control system for a manned aircraft according to claim 5, characterized in that, Each diagonal element of the temperature energy penalty diagonal matrix consists of a weighted sum of the normalized value of the current temperature relative to the ambient temperature, the reciprocal of the normalized value of the remaining available power, and the positive portion of the temperature rise rate.
7. The multi-mode flight control system for a manned aircraft according to claim 1, characterized in that, When the system detects a sudden drop in the control efficiency of any lift / thrust actuator, it automatically adjusts the upper limit of its control command to the current actual achievable value.
8. The multi-modal flight control system for a manned aircraft according to claim 1, characterized in that, In the unified outer loop control law module based on the energy potential well gradient, the proportional coefficient and differential coefficient used to generate the unified linear acceleration command adopt a scheduling method that changes linearly and monotonically with dynamic pressure.
9. The multi-modal flight control system for a manned aircraft according to claim 1, characterized in that, The attitude inner loop uses a dual-loop PID structure of angular rate and attitude angle to generate total force and torque commands.
10. The multi-modal flight control system for a manned aircraft according to claim 1, characterized in that, The lift / thrust actuator includes multiple rotors, ducted fans, aerodynamic control surfaces, and vector nozzles, and its control efficiency matrix is updated in real time by lookup table interpolation based on the angle of attack and Mach number.