Flying car control method and system giving consideration to control accuracy and flight stability

Abstract

The invention discloses a hovercar control method and system giving consideration to control accuracy and flight stability, and relates to the technical field of aircraft control. The method comprises the following steps: receiving an aircraft system state and environment parameters; a six-degree-of-freedom dynamical model of the vertical take-off and landing aircraft is constructed according to the aircraft system state and the environmental parameters, and external force and torque borne by the dynamical model are decomposed into aerodynamic force from wings and thrust from propellers. According to the method, a universal aircraft model is constructed, thrust and torque of an electric propeller and aerodynamic force and torque of wings are considered when the aircraft works, the two parts of power are comprehensively considered, and different distribution control is conducted on a power source according to different working conditions such as takeoff, cruising and landing of the aircraft; a nonlinear controller is established, a force distribution module is designed to reasonably distribute two power sources, distributed control is performed based on a control distribution module, and an optimal control distribution scheme is sought through a recursive control algorithm.

Description

Technical Field

[0001] The present invention relates to the field of aircraft control technology, and in particular to a flying car control method and system that balances control accuracy and flight stability. Background Art

[0002] As an emerging aerial vehicle, fixed-wing vertical take-off and landing (eVTOL) aircraft demonstrate tremendous potential in urban air mobility (UAM). These vehicles cleverly combine the vertical take-off and landing capabilities of traditional helicopters with the long-range, high-efficiency flight characteristics of fixed-wing aircraft, enabling flexible operation in congested urban environments and efficient intercity transportation. However, the unique hybrid architecture of eVTOL aircraft presents unprecedented challenges for flight control.

[0003] Traditional control methods typically divide the flight process into vertical takeoff and landing mode and fixed-wing cruise mode, each employing independent control schemes. For example, the vertical takeoff and landing mode primarily relies on multi-rotor propellers to generate lift, while the fixed-wing cruise mode uses the aerodynamic forces of the wings to provide lift and the propellers to generate thrust. While this separate control method can meet basic flight requirements, the complex interaction between aerodynamic forces and propeller thrust during mode switching often leads to a decrease in dynamic stability and even causes the aircraft to lose control. Furthermore, the separate control logic for different modes makes it difficult to achieve smooth and precise control during the transition process. Therefore, the present invention provides a flying car control method and system that balances control accuracy and flight stability. Summary of the Invention

[0004] The purpose of the present invention is to provide a flying car control method and system that takes into account both control accuracy and flight stability, solve the control problem of winged vertical take-off and landing aircraft in different flight modes, achieve seamless transition from take-off, cruising to landing through a unified control architecture, and provide higher flight stability and control accuracy.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a flying car control method that takes into account both control accuracy and flight stability, comprising the following steps:

[0006] Receive aircraft system status and environmental parameters;

[0007] A six-degree-of-freedom dynamic model of the vertical take-off and landing aircraft is constructed based on the aircraft system status and environmental parameters. The external forces and moments acting on the dynamic model are decomposed into the aerodynamic force from the wings and the thrust from the propellers.

[0008] A propeller model is constructed based on the decomposed propeller thrust, and the thrust and torque of the propeller model are calculated;

[0009] The aerodynamic force and moment model of the wing is constructed based on the aerodynamic force from the wing, and the force and moment on the wing are calculated in combination with the relative wind speed;

[0010] Construct a nonlinear trajectory controller and attitude controller, input the calculated thrust and torque of the propeller model, and the force and moment on the wing into the nonlinear trajectory controller to calculate the total desired force and total desired moment required by the aircraft;

[0011] Construct a force distribution model, distribute the desired thrust into thrust in different directions according to the aircraft's dynamic model and force distribution space, and calculate the desired thrust command, desired torque command, and desired attitude. The desired thrust distribution is adjusted according to the flight speed.

[0012] A control distribution model is constructed, and a recursive control algorithm is used to convert the desired thrust instructions and desired torque instructions into aircraft control signals to achieve smooth control of the aircraft during the take-off, cruise, and landing transition stages.

[0013] Furthermore, the system state includes aircraft position, speed, attitude, and angular velocity; and the environmental parameters include wind speed and airflow.

[0014] Furthermore, a six-degree-of-freedom dynamic model of the vertical take-off and landing aircraft is constructed based on the aircraft system status and environmental parameters. The external forces and moments acting on the dynamic model are decomposed into the aerodynamic force from the wing and the thrust from the propeller, as follows:

[0015] (31) The six-degree-of-freedom dynamic model of a vertical take-off and landing aircraft is studied. The system state is determined by the inertial position of the center of mass of the aircraft. speed and the angular velocity in the fuselage coordinates The definition is as follows:

[0016]

[0017] Where, is the inertia matrix in the aircraft body coordinate system; g is the gravity constant vector in the inertial system; It is a skew-symmetric mapping, a×b=S(a)b;

[0018] (32) The forces and moments of the external environment on the aircraft are grouped uniformly as f b and τ b , specifically broken down into:

[0019] f b =f T +f A ,τ b =τ T +τ A

[0020] Where f T and τ T are the thrust and torque from the propeller, respectively, and f A and τ A are the force and moment from the wing, respectively.

[0021] Furthermore, a propeller model is constructed based on the decomposed propeller thrust, and the thrust and torque of the propeller model are calculated, specifically the thrust and torque of the electric propeller:

[0022] (41) The axial thrust t and torque τ are the force and moment generated when the propeller accelerates:

[0023] T=C T ρd 4 n 2

[0024] τ=C Q ρd 5 n 2

[0025] Where C T 、C Q is a dimensionless coefficient, ρ is the air density, d is the propeller diameter, and n is the propeller speed;

[0026] The propulsive force generated by the jth propeller in the fuselage coordinate system is defined as is the unit vector in the propulsion direction, and the axial torque is

[0027] r j Defined as the vector pointing from the center of mass of the fuselage to the rotor axis, the torque of the rotor j on the aircraft at the center of mass is expressed as:

[0028]

[0029] Where, τ j,cm : represents the torque generated by the jth rotor at the center of mass of the aircraft; r j : The vector from the center of mass of the fuselage to the j-th rotor axis; T j : the magnitude of the axial thrust generated by the j-th propeller; The unit vector in the propulsion direction of the jth propeller; τ j : In the first line of the formula Here, τ j Refers to the axial torque vector generated by the j-th propeller; S(r j ): is the vector r j The skew-symmetric matrix representation of C Q : dimensionless torque coefficient; d: propeller diameter; C Tis the dimensionless thrust coefficient; μ j Represents the j-th thruster unit thrust T j The resulting torque contribution vector;

[0030] Under the premise of knowing the control of each axial thruster, the thrust f of the thruster model is obtained T and torque τ T :

[0031]

[0032] Where B T represents the matrix; f T : represents the total thrust vector generated by all thrusters on the aircraft; τ T : represents the total torque vector generated by all thrusters on the aircraft around the center of mass of the aircraft; represents the unit vector in the thrust direction of the j-th propeller; μ j : represents the j-th propeller unit thrust T j The torque contribution vector generated; N T : indicates the total number of thrusters on the aircraft; T j : represents the magnitude of the axial thrust generated by the j-th propeller; B T : thruster control effectiveness matrix; u T : Control input vector of the thruster.

[0033] Furthermore, the aerodynamic force and moment model on the wing is constructed based on the aerodynamic force from the wing, and the force and moment on the wing are calculated in combination with the relative wind speed, as follows:

[0034]

[0035] v w is the ambient wind speed, v is the inertial speed of the aircraft, Transpose the rotation matrix of the aircraft;

[0036] Define V ∞ , V xz and V xy are the free stream wind speed and the expected speed in the xz plane and xy plane of the aircraft, respectively, which are expressed as:

[0037]

[0038] Where V ∞ is the incident wind speed, α is the angle of attack, and β is the sideslip angle;

[0039] Define the lift L and drag D in the xz plane of the aircraft, define the lateral force Y on the y axis of the aircraft, and similarly define the aerodynamic torque in each axis:

[0040]

[0041] Where S ref is the wing reference area, is the chord width, C L ,C D ,G Y is the non-dimensional coefficient of lift, drag and lateral force;

[0042] The moments of the aircraft around the x, y, and z axes are calculated using the following formulas:

[0043]

[0044] C mx ,C my ,C mz is the non-dimensional coefficient of moment, ω x ,ω y ,ω z is the angular velocity of the aircraft, ΔC mx ,ΔC my ,ΔC mz is the additional effect of the control surface deflection on the torque; let the control surface angle δ Aj The linear effect on the torque, the additional torque coefficient can be expressed as:

[0045]

[0046] In summary, the aerodynamic torque is expressed as:

[0047]

[0048] Where τ′ A Determined by aerodynamic characteristics, B A u A Caused by control surface deflection; τ A : represents the total aerodynamic torque vector of the aircraft; ρ: represents the air density; S ref : represents the reference area of ​​the wing; c: represents the average aerodynamic chord length or reference chord length of the wing; V ∞ : Indicates the incident wind speed of the aircraft relative to the airflow; C' mx ,C' my ,C' mz : The dimensionless aerodynamic moment coefficients of the aircraft around the x, y, and z axes of the aircraft coordinate system; τ' A : represents the aerodynamic torque part determined by the aerodynamic characteristics of the aircraft itself, excluding the influence of the control surface deflection; N A : represents the total number of pneumatic control surfaces; δ A,j : represents the deflection angle of the jth aerodynamic control surface; B A: represents the aerodynamic control effectiveness matrix; u A : represents the vector composed of the deflection angles of the aerodynamic control surfaces, that is,

[0049] Furthermore, the non-dimensional coefficients of lift, drag, and lateral force are calculated using a flat plate model, a linear model, or a hybrid model, as follows:

[0050] The flat plate model, used in the high angle of attack region, gives a large-scale trend:

[0051]

[0052] Where, represents the lift coefficient calculated under the flat plate model; k L : represents a proportional coefficient used to calculate the lift coefficient in the flat plate model; α: represents the angle of attack or angle of attack of the aircraft, that is, the angle between the direction of the aircraft's speed and the chord line of the wing; represents the resistance coefficient calculated under the flat plate model; k D1 : represents a proportional coefficient used to calculate the resistance coefficient in the flat plate model; k D0 : represents another constant coefficient used to calculate the resistance coefficient in the flat plate model; represents the lateral force coefficient calculated under the flat plate model; k Y : represents a proportional coefficient used to calculate the side force coefficient in the flat plate model; β: represents the sideslip angle of the aircraft, that is, the angle between the projection of the aircraft's velocity direction on the horizontal plane and the longitudinal axis of the fuselage;

[0053] Linear model for the low angle of attack range:

[0054]

[0055] Where, represents the lift coefficient calculated under the linear model; C L0 : represents the lift coefficient at zero angle of attack, that is, the lift coefficient when the angle of attack α is zero; C L1 : represents the slope of the lift line, that is, the rate at which the lift coefficient changes with the angle of attack; α: represents the angle of attack of the aircraft; represents the drag coefficient calculated under the linear model; C D0 : Indicates the drag coefficient at zero lift (usually corresponding to near zero angle of attack), also known as zero lift drag coefficient or waste drag coefficient; C D1 : represents the coefficient of linear variation of drag coefficient with angle of attack α; C D2 : represents the coefficient of drag coefficient changing with the square of angle of attack α, usually related to induced drag;

[0056] represents the lateral force coefficient calculated under the linear model; C Lβ :According to the formula β, this item represents the proportional coefficient of the side force coefficient with the sideslip angle β; β: represents the sideslip angle of the aircraft;

[0057] Mixed model, mixing the flat model and the linear model through the tanh function:

[0058]

[0059] Where, f A : represents the total aerodynamic force vector of the aircraft; ρ: represents the air density; S ref : represents the reference area of ​​the wing; V ∞ : Indicates the incident wind speed of the aircraft relative to the airflow; C L (α): represents the lift coefficient calculated by the hybrid model and changes with the angle of attack α; C D (α): represents the drag coefficient calculated by the hybrid model and changes with the angle of attack α; α: represents the angle of attack of the aircraft; C Y (β): represents the side force coefficient calculated by the hybrid model and changes with the sideslip angle β; β: represents the sideslip angle of the aircraft.

[0060] Furthermore, a nonlinear trajectory controller and attitude controller are constructed. The calculated thrust and torque of the propeller model, and the force and moment on the wing are input into the nonlinear trajectory controller to calculate the total desired force and total desired moment required by the aircraft, as follows:

[0061] (71) Combining the forces and moments from the propeller model and the wing, the resultant forces and moments acting on the aircraft are unified as follows:

[0062]

[0063] Where, f b : represents the total external force vector acting on the aircraft body coordinate system; τ b : represents the total external torque vector acting on the aircraft body coordinate system; f A (v i ): represents the aerodynamic force generated by the wing, which is the relative wind speed v i function of τ' A (v i ω): represents the part of the aerodynamic torque determined by the aerodynamic characteristics of the aircraft itself, excluding the torque generated by the deflection of the control surface; B A : represents the aerodynamic control effectiveness matrix. It maps the deflection angle of the aerodynamic control surface to the resulting additional aerodynamic torque; u A: represents the control vector composed of the deflection angles of each aerodynamic control surface; B T : represents the thruster control effectiveness matrix. It maps the control input of each thruster to the total thrust and total torque they produce; u T : represents the control input vector of the thruster;

[0064] According to B A u A and B T u T Direct inversion is used to generate the resultant forces and moments, f A The existence of Rf indicates that there are difficulties in determining the flight attitude. b and τ b In the dynamics, position and attitude are used as control inputs, and the and Solve as the desired force and torque, output the desired posture and ensure Rf b Approaching

[0065] (72) Trajectory tracking controller:

[0066] In tracing a given path p d (t), the resulting position error is defined as Design a manifold that makes the position error converge to 0

[0067]

[0068] Λ p is the positive definite gain matrix for the position error, and a trajectory tracking controller using the desired net force is constructed, defined as:

[0069]

[0070] Among them, K v and K p is a positive definite gain matrix, assuming The difference between its realized value is If the difference Bounded by a matrix Then v→v r and p→p d In a matrix ∈ and the gain matrix Λ p ,K p and K v The controlled sphere changes exponentially;

[0071] (73) Design of posture tracking controller:

[0072] Define the attitude error matrix Rd is the target pose matrix, and R is the current actual pose matrix. The error function used in the rotation group SO(3) is equivalent to the error function from The error quaternion

[0073]

[0074] Where, Represents the scalar part of the error quaternion; error quaternion Used to express the difference between the desired posture and the actual posture; represents the attitude error matrix, defined as where R d is the target posture matrix, R is the current actual posture matrix; Represents the vector part of the error quaternion; Represents the attitude error matrix The transpose of

[0075] The attitude change of the aircraft is described by the following formula:

[0076]

[0077] In the formula The time derivative of the attitude error matrix, which represents the rate of change of the aircraft attitude error over time; The attitude error matrix is ​​usually defined as the relationship between the desired attitude and the actual attitude; R: the actual attitude matrix, which refers to the current attitude of the aircraft and is a rotation matrix; The derivative of the actual attitude matrix with respect to time; The time derivative of the desired attitude matrix; S(·): skew-symmetric matrix operator, which converts a three-dimensional vector into a 3×3 skew-symmetric matrix, used to represent the cross product operation of the vector, that is, S(a)b=a×b; ω: the actual angular velocity of the aircraft; ω d : expected angular velocity or target angular velocity;

[0078] The control rules are defined as:

[0079]

[0080] Where τ is the control torque, which is the torque that needs to be applied to the aircraft to achieve the desired attitude and angular velocity; J is the moment of inertia matrix of the aircraft, which is a symmetric positive definite matrix; The time derivative of the reference angular velocity; ω r : reference angular velocity; K ω : angular velocity error gain matrix; Angular velocity error, defined as the difference between the actual angular velocity and the reference angular velocity; kq : attitude quaternion error gain coefficient; The vector part of the attitude error quaternion. Quaternions are often used to represent three-dimensional rotations, and their vector part is related to the rotation axis and rotation angle;

[0081] K ω is a positive definite matrix, k q >0.

[0082] Furthermore, a force distribution model is constructed. According to the dynamic model of the aircraft and the force distribution space, the desired thrust is distributed into thrusts in different directions. The desired thrust command, desired torque command, and desired attitude are calculated. The desired thrust distribution is adjusted according to the flight speed:

[0083] (81) Obtained by changing actuator commands or aircraft attitude and This process is the distribution of force, which is achieved by solving the desired posture and the desired thrust command. First, the desired posture R is obtained by the following formula: d :

[0084]

[0085] (82) Solve for the desired thrust command by minimizing the force residual

[0086] where f T Directly from dynamic fast input u T Decision, and f A Depends on attitude and speed; for fixed-wing forward-flying vertical take-off and landing aircraft, first use the wing lift to make f A Prioritize meeting expectations The rest is made up of T supply;

[0087] (83) Determination of expected posture:

[0088] At low speed:

[0089] In the object coordinate system, The unit vector representing the favorable thrust direction, the current thruster output is:

[0090]

[0091] R d Obtained by any rotation that satisfies the following conditions:

[0092]

[0093] At high speed:

[0094] make represents the required aircraft thrust, given the measured incident wind speed, and the incident wind coordinate system axes are defined as:

[0095]

[0096] The incident wind coordinate system is expressed in the fuselage coordinate system as:

[0097]

[0098] According to the aerodynamic force and moment model of the wing, by setting an ideal angle of attack α d , giving priority to raising the wing over generating lift perpendicular to the wind speed

[0099]

[0100] in Incident wind coordinate system Axis rotation α d , represented by the rotation matrix R a ;

[0101] The desired pose is calculated by successive rotations:

[0102] R d =RR i R α

[0103] (84)Thrust distribution:

[0104] At the current attitude, the thrust of the thruster is calculated by the following formula:

[0105]

[0106] Where M represents the matrix, which shields the body axial force components that cannot be realized by the propeller input; f A Estimated by the current incident wind speed.

[0107] Furthermore, a control allocation model is constructed, and a recursive control algorithm is used to convert the desired thrust command and the desired torque command into the aircraft control signal, so as to achieve smooth control of the aircraft during the takeoff, cruise, and landing transition phases:

[0108] (91) Control distribution:

[0109] After derivation, the required thrust and torque of the propeller are obtained, which can be expressed as follows:

[0110]

[0111] Define the propulsion vector ω TDefine the feasible control space of distributed propulsion for the combination of thrust and torque of the thrusters The goal of control allocation is to find So that:

[0112]

[0113] For the drive system, when N T >N w When , there does not exist a single inverse mapping that recovers a complete set of feasible w without violating the control constraints. T ; But because B T There are multiple right inverses, and there is still more than one B T Satisfy the above formula; N T Indicates the number of independent thrusters, N w Denotes the dimensions of the resultant total thrust and total torque vectors;

[0114] (92) Reachable thruster propulsion space

[0115] Command control space Its boundaries are

[0116] A feasible propulsion vector ω T Located in B T In the column space of The accessible propulsion space is defined as:

[0117] where u T,j is the vector u T The jth element of b j For B T The j-th column vector of

[0118] Assuming that each thruster is identical, u T,min =0,u T,max =1, definition:

[0119]

[0120] Among them, Conv represents the convex hull of a given set, for Minkowski and;

[0121] (93) Static and recursive control allocation:

[0122] (93.1) The recursive allocation method iteratively calculates the pseudo-inverse or generalized inverse solution, prioritizes the requirements of the main forces, and then distributes the remaining moments. The recursive control algorithm flow is as follows:

[0123] Recursive control algorithm input:

[0124] B τ : control allocation matrix, describing the contribution of thrust devices to the total force and total torque;

[0125] w τ : total target force, including target force and torque;

[0126] Recursive control algorithm output:

[0127] u τ : The thrust output of each thruster meets B τ u τ =w τ andu τ,min ≤u τ ≤u τ,max ;

[0128] (93.2) Steps:

[0129] (93.21) Check the target total force w τ feasibility, if the target total force w τ Out of reach or approximate range Then for w τ Scale it back to within the feasible range;

[0130] (93.22) Pseudo-inverse calculation initial solution

[0131] If B τ Is a square matrix, directly calculate the inverse matrix:

[0132]

[0133] If B τ If is a rectangular matrix, the pseudo-inverse is used:

[0134]

[0135] (93.23) Check constraints:

[0136] For each u τ The component u τ,j :

[0137] If u τ,i >u τ,max , will u τ,max Assign to u τ,i ;

[0138] If u τ,i τ,min , will u τ,min Assign to u τ,i ; ​

[0139] (93.24) Update target force and distribution matrix

[0140] According to the fixed thrust component u τ,i , update the target force w τ and the allocation matrix B τ , delete the assigned columns;

[0141] (93.25) Recursive call

[0142] For the updated target w τ and matrix B τ , recursively call the control allocation algorithm until the remaining matrix B τ Become a square;

[0143] (93.26) Return result

[0144] Merge the allocation results u of all recursive levels τ , and returns.

[0145] The present invention has at least the following beneficial effects:

[0146] The present invention constructs a universal aircraft model, considers the thrust and torque of the electric propeller and the aerodynamic force and torque of the wing when the aircraft is working, comprehensively considers the two parts of power, and performs different distributed control on the power source according to different working conditions of the aircraft, such as takeoff, cruising, landing, etc.; by establishing a nonlinear controller, designing a force distribution module to reasonably distribute the two power sources, performing distributed control based on the control distribution module, and seeking the optimal control distribution scheme through a recursive control algorithm to achieve the purpose of precise control and better flight conditions. This method optimizes the problem that traditional control methods control wings and propellers separately and cannot make accurate responses to the specific working conditions of the aircraft. A unified control architecture is proposed to uniformly optimize the control of the two power sources of the aircraft to achieve better control effects and flight conditions.

[0147] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0148] Figure 1 Schematic diagram of the control method of the present invention. DETAILED DESCRIPTION

[0149] The following will be combined with the accompanying drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the embodiments described are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present disclosure.

[0150] See also Figure 1 The present invention provides a technical solution: a flying car control method that takes into account both control accuracy and flight stability, comprising the following steps:

[0151] Receive aircraft system status and environmental parameters;

[0152] A six-degree-of-freedom dynamic model of the vertical take-off and landing aircraft is constructed based on the aircraft system status and environmental parameters. The external forces and moments acting on the dynamic model are decomposed into the aerodynamic force from the wings and the thrust from the propellers.

[0153] A propeller model is constructed based on the decomposed propeller thrust, and the thrust and torque of the propeller model are calculated;

[0154] The aerodynamic force and moment model of the wing is constructed based on the aerodynamic force from the wing, and the force and moment on the wing are calculated in combination with the relative wind speed;

[0155] Construct a nonlinear trajectory controller and attitude controller, input the calculated thrust and torque of the propeller model, and the force and moment on the wing into the nonlinear trajectory controller to calculate the total desired force and total desired moment required by the aircraft;

[0156] Construct a force distribution model, distribute the desired thrust into thrust in different directions according to the aircraft's dynamic model and force distribution space, and calculate the desired thrust command, desired torque command, and desired attitude. The desired thrust distribution is adjusted according to the flight speed.

[0157] A control distribution model is constructed, and a recursive control algorithm is used to convert the desired thrust instructions and desired torque instructions into aircraft control signals to achieve smooth control of the aircraft during the take-off, cruise, and landing transition stages.

[0158] Next, the technical solution of the present invention is further described with reference to specific embodiments:

[0159] S1. Establish a general vertical take-off and landing aircraft model, taking into account the aircraft's global position, inertial velocity, attitude rotation, and acceleration. Apply this model to obtain the effects of external forces and torques on the aircraft, which are used to control the aircraft's vertical take-off and landing process.

[0160] Study the six-degree-of-freedom dynamic model of vertical take-off and landing aircraft. The system state is determined by the inertial position of the center of mass of the aircraft. speed and the angular velocity in the fuselage coordinates The definition is as follows:

[0161]

[0162] is the inertia matrix in the fuselage coordinate system; g is the gravity constant vector in the inertial system; S(·): It is a skew-symmetric mapping, a×b=S(a)b; the external forces and moments on the aircraft are uniformly grouped as f b and τ b , can be decomposed into:

[0163] f b =f T +f A ,τ b =τ T +τ A

[0164] f T and τ T is the effect from the propeller, f A and τ A The effect comes from the wing.

[0165] S2. Build a propeller model, specifically the thrust and torque of the electric propeller

[0166] Axial thrust T and torque τ are the main forces and moments generated when the propeller accelerates:

[0167] T=C T ρd 4 n 2

[0168] ρ=C Q ρd 5 n 2

[0169] C T 、C Q is a dimensionless coefficient, ρ is the air density, d is the propeller diameter, and n is the propeller speed.

[0170] The propulsive force generated by the jth propeller in the fuselage coordinate system is defined as is the unit vector in the propulsion direction. Similarly, the axial torque is

[0171] r j Defined as the vector pointing from the center of mass of the fuselage to the rotor axis, the torque of the rotor j on the aircraft at the center of mass can be expressed as:

[0172]

[0173] Where, τ j,cm : represents the torque generated by the jth rotor at the center of mass of the aircraft; r j : The vector from the center of mass of the fuselage to the j-th rotor axis; Tj : the magnitude of the axial thrust generated by the j-th propeller; The unit vector in the propulsion direction of the jth propeller; τ j : In the first line of the formula Here, τ j Refers to the axial torque vector generated by the j-th propeller; S(r j ): is the vector r j The skew-symmetric matrix representation of C Q : dimensionless torque coefficient; d: propeller diameter; C T is the dimensionless thrust coefficient; μ j Represents the j-th thruster unit thrust T j The resulting torque contribution vector;

[0174] Under the premise of knowing the control of each axial thruster, we can get:

[0175]

[0176] Where, f T : represents the total thrust vector generated by all propellers on the aircraft; τ T : represents the total torque vector generated by all thrusters (propellers) on the aircraft around the center of mass of the aircraft; (For example ): represents the unit vector in the thrust direction of the j-th propeller; μ j (e.g. μ1,…, ): represents the j-th propeller unit thrust T j The torque contribution vector generated; N T : Indicates the total number of propellers on the aircraft; T j (e.g. T1,…, ): represents the magnitude of the axial thrust generated by the j-th propeller; B T : thruster control effectiveness matrix; u T : control input vector of the thruster;

[0177] By configuring the aircraft, you can get accurate B T matrix;

[0178] S3. Model the aerodynamic forces and moments on the wing

[0179] The forces and moments on the airfoil depend primarily on the relative wind speed.

[0180]

[0181] v w is the ambient wind speed, v is the inertial speed of the aircraft, Transpose the vehicle's rotation matrix.

[0182] Define V ∞ , V xz and V xy are the free stream wind speed and the expected speed in the aircraft's xz and xy planes, respectively.

[0183]

[0184] V ∞ is the incident wind speed, α is the angle of attack, and β is the sideslip angle.

[0185] Define the lift L and drag D in the xz plane of the aircraft, define the lateral force Y on the y axis of the aircraft, and similarly define the aerodynamic torque in each axis:

[0186]

[0187] S ref is the wing reference area, is the chord width, C L ,C D ,C Y are the non-dimensional coefficients of lift, drag and lateral force.

[0188] The non-dimensional coefficient can be obtained through wind tunnel test or model calculation. Here we use two models for calculation:

[0189] The flat plate model, used in the high angle of attack region, gives a large-scale trend:

[0190]

[0191] Where, represents the lift coefficient calculated under the flat plate model; k L : represents a proportional coefficient used to calculate the lift coefficient in the flat plate model; α: represents the angle of attack (or angle of attack) of the aircraft, that is, the angle between the direction of the aircraft's velocity and the chord line of the wing; represents the resistance coefficient calculated under the flat plate model; k D1 : represents a proportional coefficient used to calculate the resistance coefficient in the flat plate model; k D0 : represents another constant coefficient used to calculate the resistance coefficient in the flat plate model; represents the lateral force coefficient calculated under the flat plate model; k Y : represents a proportional coefficient used to calculate the side force coefficient in the flat plate model; β: represents the sideslip angle of the aircraft, that is, the angle between the projection of the aircraft's velocity direction on the horizontal plane and the longitudinal axis of the fuselage;

[0192] Linear model, more accurate in the low angle of attack range:

[0193]

[0194] Where, represents the lift coefficient calculated under the linear model; C L0 : represents the lift coefficient at zero angle of attack, that is, the lift coefficient when the angle of attack α is zero; C L1 : represents the slope of the lift line, that is, the rate at which the lift coefficient changes with the angle of attack; α: represents the angle of attack of the aircraft; represents the drag coefficient calculated under the linear model; C D0 : Indicates the drag coefficient at zero lift (usually corresponding to near zero angle of attack), also known as zero lift drag coefficient or waste drag coefficient; C D1 : represents the coefficient of linear variation of drag coefficient with angle of attack α; C D2 : represents the coefficient of drag coefficient changing with the square of angle of attack α, usually related to induced drag;

[0195] represents the lateral force coefficient calculated under the linear model; C Lβ :According to the formula β, this item represents the proportional coefficient of the side force coefficient with the sideslip angle β; β: represents the sideslip angle of the aircraft;

[0196] Hybrid model, which mixes two models through the tanh function to ensure accuracy and take into account the global range:

[0197]

[0198] The moments of the aircraft around the x, y, and z axes are calculated using the following formulas:

[0199]

[0200] C mx ,C my ,C mz is the non-dimensional coefficient of moment, ω x ,ω y ,ω z is the angular velocity of the aircraft, ΔC mx ,ΔC my ,ΔC mz is the additional effect of control surface deflection on the torque.

[0201] Effects of control surface deflection:

[0202] Assuming the control surface angle δ Aj The linear effect on the torque, the additional torque coefficient can be expressed as:

[0203]

[0204] In summary, the aerodynamic torque can be written as:

[0205]

[0206] Where τ′ A Determined by aerodynamic characteristics, B A u A Caused by control surface deflection; τ A : represents the total aerodynamic torque vector of the aircraft; ρ: represents the air density; S ref : represents the reference area of ​​the wing; c: represents the average aerodynamic chord length or reference chord length of the wing; V ∞ : Indicates the incident wind speed of the aircraft relative to the airflow; C' mx ,C' my ,C' mz : The dimensionless aerodynamic moment coefficients of the aircraft inherent (i.e., excluding the influence of control surface deflection) around the x, y, and z axes of the aircraft coordinate system; τ' A : represents the aerodynamic torque part determined by the aerodynamic characteristics of the aircraft itself, excluding the influence of the control surface deflection; N A : represents the total number of pneumatic control surfaces; δ A,j : represents the deflection angle of the jth aerodynamic control surface; B A : represents the aerodynamic control effectiveness matrix; u A : represents the vector composed of the deflection angles of the aerodynamic control surfaces, that is,

[0207] S4. Establish a unified control architecture, including: Nonlinear controller: Design a nonlinear controller based on force and torque input to achieve target attitude and position control, specifically:

[0208] Combining the forces and moments from the two models, the resultant forces and moments acting on the aircraft are unified as follows:

[0209]

[0210] Where, f b : represents the total external force vector acting on the aircraft body coordinate system; τ b : represents the total external torque vector acting on the aircraft body coordinate system; f A (v i ): represents the aerodynamic force generated by the wing, which is the relative wind speed v i function of τ' A (v i ,ω): represents the aerodynamic characteristics of the aircraft itself (and the relative wind speed v iThe part of the aerodynamic torque determined by the body angular velocity ω, excluding the torque generated by the deflection of the control surface; B A : represents the aerodynamic control effectiveness matrix. It maps the deflection angle of the aerodynamic control surface to the resulting additional aerodynamic torque; u A : represents the control vector composed of the deflection angles of each aerodynamic control surface; B T : represents the thruster (propeller) control effectiveness matrix. It maps the control input of each thruster to the total thrust and total torque they produce; u T : represents the control input vector of the propeller, such as the thrust command of each propeller;

[0211] By B A u A and B T u T It can be seen that direct inversion can be used to generate forces and moments, f A The existence of Rf indicates that there are difficulties in determining the flight attitude. b and τ b In dynamics, position and attitude are used as control inputs. and It can be solved as the desired force and torque. Then use the new force distribution method to output the desired posture and ensure Rf b Approaching

[0212] Trajectory tracking controller:

[0213] In tracing a given path p d (t), the resulting position error is defined as A manifold is designed to make the position error converge to 0

[0214]

[0215] Λ p A trajectory tracking controller using the desired net force is proposed, which is defined as:

[0216]

[0217] Among them, K v and K p Is a positive definite gain matrix. Assume The difference between it and its implementation is If it is bounded by a matrix Then v→v r and p→p d In a matrix ∈ and the gain matrix Λ p,K p and K v The controlled sphere changes exponentially.

[0218] Posture Tracking Controller:

[0219] A controller is designed to provide global tracking of any pose

[0220] Define the attitude error matrix R d is the target pose matrix, and R is the current actual pose matrix. The error function used in the rotation group SO(3) is equivalent to the error function from The error quaternion

[0221]

[0222] Where, Represents the scalar part of the error quaternion; error quaternion Used to express the difference between the desired posture and the actual posture; represents the attitude error matrix. It is defined as where R d is the target posture matrix, R is the current actual posture matrix; Represents the vector part of the error quaternion; Represents the attitude error matrix The transpose of

[0223] The attitude change of the aircraft is described by the following formula:

[0224]

[0225] In the formula The time derivative of the attitude error matrix. It represents the rate of change of the vehicle's attitude error over time. The attitude error matrix is ​​usually defined as the relationship between the desired attitude and the actual attitude, such as R d : The desired attitude matrix or target attitude matrix, which is the attitude that the aircraft expects to achieve, is a rotation matrix:

[0226] R: actual attitude matrix, which is the current attitude of the aircraft and is a rotation matrix; The derivative of the actual attitude matrix with respect to time; The time derivative of the desired attitude matrix; S(·): skew-symmetric matrix operator. It converts a three-dimensional vector into a 3×3 skew-symmetric matrix, which is used to represent the cross product operation of the vector, that is, S(a)b = a×b; ω: The actual angular velocity of the aircraft. Usually refers to the angular velocity vector in the body coordinate system; ω d : Desired angular velocity or target angular velocity. This is the angular velocity that the aircraft is expected to achieve;

[0227] The control rules are defined as:

[0228]

[0229] Where τ is the control torque. The controller calculates the torque that needs to be applied to the aircraft to achieve the desired attitude and angular velocity. J is the aircraft's moment of inertia matrix. This is a symmetric positive definite matrix that describes the aircraft's rotational inertia about its center of mass. The time derivative of the reference angular velocity; ω r : Reference angular velocity. This is an intermediate variable used to assist in calculating the control torque; K ω : angular velocity error gain matrix. This is a positive definite matrix used to adjust the response strength of the control law to the angular velocity error; Angular velocity error, defined as the difference between the actual angular velocity and the reference angular velocity; k q : attitude quaternion error gain coefficient, which is a scalar value greater than zero and is used to adjust the response strength of the control law to the attitude quaternion error vector; The vector part of the attitude error quaternion. Quaternions are often used to represent three-dimensional rotations, and their vector part is related to the rotation axis and rotation angle;

[0230] K ω is a positive definite matrix, k q >0;

[0231] The total forces and total moments required by a vehicle are the desired net forces and net moments calculated by the control system to accurately track the planned trajectory and maintain the desired attitude. They are the outputs of the nonlinear trajectory controller and attitude controller.

[0232] Expected total force Calculated by the trajectory tracking controller, the goal is to make the actual position p and velocity v of the aircraft track the desired path p d (t).

[0233] Desired total torque τ: calculated by the attitude tracking controller, the goal is to make the actual attitude R and angular velocity ω of the aircraft track the desired attitude R d and the desired angular velocity ω d .

[0234] Expected total force In the inertial coordinate system, it is expressed as:

[0235]

[0236] in:

[0237] g is the acceleration due to gravity vector.

[0238] is the reference speed The derivative of is the second derivative of the desired path (desired acceleration).

[0239] is the speed error.

[0240] is the position error.

[0241] K p ,K v ,Λ p is a positive definite gain matrix.

[0242] this It is the total force that the aircraft needs to be subjected to, calculated by the trajectory controller in order to eliminate position and velocity errors and make the aircraft move according to the desired acceleration.

[0243] The expected total torque τ is expressed in the body coordinate system as:

[0244]

[0245] in:

[0246] J is the inertia matrix of the aircraft.

[0247] ω r =R T R d ω d -2Λ ω q ν , is its derivative, representing the quantity related to the desired angular acceleration.

[0248] S(Jω) is a skew-symmetric matrix containing Jω, -S(Jω)ω r (or -S(Jω)ω in the document) represents the torque term related to the gyroscopic effect. The document formula is S(Jω)ω r , here may r Contains information about the current angular velocity ω.

[0249] is the angular velocity error.

[0250] q ν is the vector part of the attitude error quaternion.

[0251] K ω ,k q ,Λ ω is a positive definite gain or a scalar.

[0252] This τ is the total torque that the aircraft needs to be subjected to, calculated by the attitude controller in order to eliminate attitude and angular velocity errors;

[0253] S5. Force distribution module, force distribution module: according to the dynamic model of the aircraft and the force distribution space, the target is distributed into thrusts in different directions.

[0254] Obtained by changing actuator commands or vehicle attitude and This process is the distribution of force, which is achieved by solving the desired posture and the desired thrust command. First, the desired posture R is obtained by the following formula: d :

[0255]

[0256] Secondly, the thrust command is solved by minimizing the force residual

[0257] Note f T Directly from dynamic fast input u T Decision, and f A Depends on attitude and speed. For a fixed-wing forward-flying vertical take-off and landing aircraft, it is important to utilize the wing lift as much as possible to make f A Prioritize meeting expectations The rest is made up of T supply.

[0258] Desired posture determines:

[0259] At low speed:

[0260] Assume that in the object coordinate system, The unit vector representing the favorable thrust direction, the current thruster output is

[0261]

[0262] R d Obtained by any rotation that satisfies the following conditions:

[0263]

[0264] At high speed:

[0265] make represents the required aircraft thrust, given the measured incident wind speed, and the incident wind coordinate system axes are defined as:

[0266]

[0267] The incident wind coordinate system is expressed in the fuselage coordinate system as:

[0268]

[0269] According to the lift model, by setting an ideal angle of attack α d , giving priority to raising the wing over generating lift perpendicular to the wind speed

[0270]

[0271] in Incident wind coordinate system Axis rotation α d , represented by the rotation matrix R a .

[0272] The desired pose can be calculated by performing the successive rotations:

[0273] R d =RR i R α

[0274] Thrust distribution:

[0275] At the current attitude, the thruster force is calculated using the following formula:

[0276]

[0277] Where M is a matrix that shields the body axis force components that cannot be realized by the thruster input;

[0278] The aerodynamic force f on the wing A is the current incident wind speed v i The specific steps for estimation are as follows:

[0279] Calculate the incident wind speed (v i ), it is necessary to determine the speed of the aircraft relative to the air, that is, the incident wind speed v i ;

[0280] The velocity is determined by the inertial velocity v of the aircraft, the attitude rotation matrix R of the aircraft (specifically, its transpose R T ) and ambient wind speed v w Calculated [cite:1,9]: v i =R T vv w

[0281] Calculate the free stream wind speed (V ∞ ), angle of attack (α) and sideslip angle (β): free stream wind speed V ∞ The incident wind speed vector v i Size (module):

[0282] V ∞ =‖v i ‖

[0283] The angle of attack α and the sideslip angle β are determined by the incident wind speed v i The components in the aircraft body coordinate system are calculated. i =[v ix ,v iy ,v iz ] T ,but:

[0284]

[0285] Determine the aerodynamic coefficient (C L ,C D ,C Y ): The lift L, drag D, and lateral force Y generated by the wing can be expressed by the following formula:

[0286]

[0287] Where ρ is the air density, S ref is the wing reference area, C L ,C D ,C Y are the lift coefficient, drag coefficient, and side force coefficient respectively;

[0288] These dimensionless aerodynamic coefficients C L ,C D ,C Y is a function of the angle of attack α and the sideslip angle β;

[0289] Calculate the total aerodynamic force (f A ): After obtaining the aerodynamic coefficient, combined with the free stream wind speed V ∞ , air density ρ and reference area S ref , the total aerodynamic force f can be calculated A ;

[0290]

[0291] Therefore, f AThe estimation of is a multi-step process that relies on accurate measurement or estimation of the current incident wind speed, combined with the aerodynamic model of the aircraft to calculate. This estimate is essential for the force distribution module to correctly calculate the required propeller thrust f T It is crucial;

[0292] S6. Control distribution module: Calculates the control signals of each motor through optimization methods to ensure the robustness and accuracy of thrust distribution.

[0293] Control Allocation:

[0294] After derivation, the required thrust and torque are obtained:

[0295]

[0296] Define the propulsion vector ω T Define the feasible control space of distributed propulsion for the combination of thrust and torque of the thrusters The goal of control allocation is to find So that:

[0297]

[0298] For overdrive systems, when N T >N w When , there is no single inverse mapping that can recover the complete set of feasible w without violating some control constraints T ; N T Indicates the number of independent thrusters, N w represents the dimensions of the total thrust and total torque vector; however, due to B T There are multiple right inverses, and there is still more than one B T Satisfy the above formula;

[0299] Reachable thruster propulsion space

[0300] Command control space Its boundaries are A feasible propulsion vector ω T Located in B T In the column space of The accessible propulsion space is defined as

[0301] where u T,j is the vector u T The jth element of b j For B T Assuming that each thruster is exactly the same, u T,min =0,u T,max =1

[0302] definition:

[0303]

[0304] Among them, Conv represents the convex hull of a given set, For Minkowski and.

[0305] Static and recursive control allocation:

[0306] Method 1: Pseudo-inverse method, calculate u by least squares solution T

[0307] B T The right pseudo-inverse of Therefore This method is computationally simple and suitable for real-time control, but it cannot handle physical constraints.

[0308] Method 2: Optimization method, defining the objective function to simultaneously meet the torque requirements and physical constraints

[0309]

[0310] This method can handle the constraints and obtain the optimal allocation result, but the computational complexity is high and may affect the real-time performance.

[0311] Method 3: Recursive allocation method, which iteratively calculates the pseudo-inverse or generalized inverse solution, prioritizes the requirements of the main forces, and then distributes the remaining moments.

[0312] Recursive control algorithm flow:

[0313] enter:

[0314] B τ : Control distribution matrix, describing the contribution of thrust devices to the total force and total torque.

[0315] w τ : Total target force, including target force and torque.

[0316] Output:

[0317] u τ :The output of each thrust device meets B τ u τ =w τ andu τ,min ≤u τ ≤u τ,max .

[0318] step:

[0319] 1. Check target w τ feasibility, if the target w τ Out of reach or its approximate range Then for w τ Scale it back to within a feasible range.

[0320] 2. Pseudo-inverse calculation of initial solution

[0321] If B τ Is a square matrix, directly calculate the inverse matrix:

[0322]

[0323] If B τ If is a rectangular matrix, the pseudo-inverse is used:

[0324]

[0325] 3. Check constraints:

[0326] For each u τ The component u τ,j :

[0327] If μ τ,i >μ τ,max , μ τ,max Assign to μ τ,i

[0328] If μ τ,i <μ τ,min , will u τ,min Assign to μ τ,i

[0329] 4. Update target force and distribution matrix

[0330] According to the fixed inference component μ τ,i , update the target force w τ and the allocation matrix B τ , delete the assigned column.

[0331] 5. Recursive calls

[0332] For the updated target w τ and matrix B τ , recursively call the control allocation algorithm until the remaining matrix B τ Become a square formation.

[0333] 6. Return results

[0334] Merge the allocation results u of all recursive levels τ , and returns.

[0335] Recursive allocation has the following characteristics:

[0336] (1) Gradual convergence: Each iteration gradually reduces the target force w by correcting a thrust component that exceeds the constraint. τ In the worst case, the algorithm will perform nm iterations, (n is the number of thrusters and m is the number of independent constraints)

[0337] (2) Each iteration requires a pseudo-inverse or inverse matrix operation, and the total time complexity is O(nm).

[0338] (3) Robustness: It can ensure that the final allocation result meets all physical constraints.

[0339] Example 2:

[0340] To improve the performance of flying car control systems in the face of model parameter uncertainty, external environmental disturbances (such as strong winds and atmospheric turbulence), and flight state changes (such as load adjustments), this paper proposes an enhanced control method that incorporates robust and adaptive mechanisms into the design of nonlinear controllers.

[0341] Robust Adaptive Enhancement of Aircraft Attitude Control

[0342] The expected control torque τ of the aircraft attitude is calculated as follows:

[0343]

[0344] in:

[0345] J: Inertia matrix in the aircraft body coordinate system.

[0346] ω: The actual angular velocity of the aircraft in the fuselage coordinate system.

[0347] ω r : Reference angular velocity, which is calculated as

[0348] The attitude error matrix represents the deviation between the desired attitude and the actual attitude.

[0349] ω d : Desired vehicle angular velocity.

[0350] Λ ω : A positive definite gain matrix used to adjust the contribution of the attitude error quaternion to the reference angular velocity.

[0351] Reference angular velocity ω r The derivative with respect to time.

[0352] S(·): skew-symmetric mapping operator, for any three-dimensional vector a, b, a×b=S(a)b.

[0353] Angular velocity error, defined as

[0354] q v : The vector portion of the attitude error quaternion, which describes the rotation error from the actual attitude to the desired attitude.

[0355] K ω : A positive definite angular velocity error gain matrix.

[0356] k q : A scalar attitude error gain greater than zero.

[0357] τ adapt : Adaptive compensation torque is used to online estimate and compensate for slow time-varying uncertainties and parameter changes in the system model.

[0358] τ robust : Robust compensation torque to suppress rapidly changing external disturbances and inaccurately modeled dynamic characteristics in the system.

[0359] Adaptive compensation torque τ adapt Ways to achieve this:

[0360] The adaptive compensation torque is intended to cope with changes in inertial parameters caused by, for example, changes in aircraft loads, or slow drifts in aerodynamic characteristics with flight conditions.

[0361] Adaptive method based on parameter identification:

[0362] If the uncertainty in the system can be linearly represented by a set of unknown parameters Θ, for example, the effect of uncertainty on the torque can be expressed as a known function (regression matrix) The product of the unknown parameter vector Θ. Then the adaptive compensation torque is designed as in It is an online estimate of the unknown parameter Θ. Parameter estimation algorithm Usually designed based on Lyapunov stability theory, e.g. Where Γ is the adaptive gain matrix, P is a positive definite matrix, is some measure of the posture tracking error.

[0363] Robust compensation torque τ robust Ways to achieve this:

[0364] Robust compensation torque is mainly used to resist sudden and unknown external disturbances, such as gust impact.

[0365] Robust method based on sliding mode control:

[0366] First, a sliding surface (or switching function) s is defined, which is the attitude tracking error (such as angular velocity error and the attitude error quaternion vector part ) linear or nonlinear combination, such as where λ s is a positive constant. The robust compensation torque is designed to include two parts: one is the linear feedback term -K s s(K s is a positive definite gain matrix), which is used to make the system state approach the sliding surface; the other part is the sign function term -ρ s sgn(s) or its smooth approximation (such as the saturation function sat(s / ∈ s ) or the hyperbolic tangent function tanh(s / ∈ s )), used to overcome bounded external disturbances, where ρ s The size of needs to be set according to the upper bound of the disturbance.

[0367] Similarly, in order to achieve accurate trajectory tracking of the aircraft under uncertainty and disturbance, the total force f expected to be applied to the aircraft is calculated as follows:

[0368]

[0369] in:

[0370] g: gravitational acceleration vector.

[0371] p: The actual inertial position of the aircraft.

[0372] p d : Desired inertial position of the vehicle.

[0373] Position tracking error, defined as

[0374] v r : Reference speed, defined as in is the velocity of the desired trajectory, Λ p is a positive definite position error gain matrix.

[0375] Reference speed v r The derivative with respect to time.

[0376] v: The actual inertial speed of the aircraft.

[0377] Velocity tracking error, defined as

[0378] K p : A positive definite position error gain matrix.

[0379] K v : A positive definite velocity error gain matrix.

[0380] f adapt : Adaptive compensation force used to compensate for factors such as unknown changes in aircraft mass and unmodeled air resistance.

[0381] f robust : Robust compensation force, used to resist interference with the flight trajectory caused by sudden gusts of wind, for example.

[0382] f adapt and f robust The specific implementation method is different from the τ in the attitude controller adapt and τ robust The design ideas and approaches are similar.

[0383] The desired total force f and total torque τ obtained through the above-mentioned robust adaptive enhancement method will serve as input instructions for the aircraft's force distribution and control distribution system, and then calculate the specific control instructions for each propulsion unit and aerodynamic control surfaces, thereby achieving precise, stable and reliable flight of the flying car in complex environments.

[0384] Example 3:

[0385] When a flying car transitions from vertical takeoff and landing (VTOL) mode to forward cruise mode and vice versa, its primary lift source and control methods change significantly. This module details how to achieve smooth, seamless, and stable control during this critical transition through a unified control architecture, optimized flight envelope planning, and a refined aerodynamic and propeller thrust fusion strategy.

[0386] Planning and tracking of the flight envelope during the transition phase

[0387] To ensure the safety and efficiency of the transition phase, an optimized transition flight envelope needs to be planned in advance.

[0388] Envelope definition:

[0389] The transition flight envelope is the ideal trajectory of the aircraft in the state space of speed, altitude, and attitude. It connects the terminal stable state of the vertical take-off and landing mode with the initial stable state of the fixed-wing cruise mode.

[0390] Envelope parameterization example:

[0391] Assume s trans ∈[0,1] is the normalized process parameter of the transition phase, for example, it can be based on the current airspeed V air Function:

[0392]

[0393] in:

[0394] V air : The current airspeed of the aircraft.

[0395] V trans,start : Characteristic airspeed at the start of the transition.

[0396] V trans,end : Characteristic airspeed at the end of the transition.

[0397] sat(·,min,max): saturation function, which limits the input to the interval [min,max].

[0398] Based on this process parameter, the desired flight state can be planned, for example:

[0399] Expected flight speed v d (s trans )=v VTOL,end +s trans (v Cruise,start -v VTOL,end )

[0400] Expected flight altitude h d (s trans )=h VTOL,end +s trans (h Cruise,start -h VTOL,end )

[0401] Desired pitch angle θ d (s trans )=θ VTOL,end +s trans (θ Cruise,start -θ VTOL,end )

[0402] The subscript “VTOL,end” indicates the state value at the end of the vertical take-off and landing mode, and “Cruise,start” indicates the state value at the start of the cruise mode.

[0403] Envelope tracking:

[0404] The trajectory tracking controller (responsible for position and velocity control) and attitude controller (responsible for attitude and angular velocity control) of the aircraft will work together to make the actual state of the aircraft accurately track the expected trajectory p planned above. d (t)(by h d (s trans ) and other integrals to obtain), the expected speed v d (s trans ) and the desired posture R d (t)(by θ d (s trans) etc.).

[0405] Smooth handover and fusion strategy between aerodynamic force and propeller thrust

[0406] The core of the transition is the smooth transition of the lift source from propeller-dominated to wing-dominated.

[0407] Lift source weight factor (λ lift ):

[0408] Introduce a transition process parameter s trans (or directly with the airspeed V air )The lift source weight factor λ changes lift For example, when transitioning from VTOL to cruise:

[0409]

[0410] Or use a smoother sigmoid function:

[0411]

[0412] The value range of this factor is [0,1]. trans =0 (transition start, low speed), λ lift ≈1, indicating that the lift is mainly provided by the thrusters. trans =1 (transition end, high speed), λ lift ≈0, indicating that the lift is mainly provided by the wings.

[0413] Desired lift distribution:

[0414] Assume that the total vertical lift expected by the aircraft calculated by the trajectory controller is L total_d (This is the magnitude of the vertical component of the desired total force f in the inertial frame.) This lift will be distributed between the propulsion system and the wing as follows:

[0415] The vertical lift L expected to be provided by the thruster propeller_d =λ lift (s trans )·L total_d

[0416] The aerodynamic lift L expected to be generated by the wing wing_d =(1-λ lift (s trans ))·L total_d

[0417] Adaptation of the force distribution module:

[0418] The goal of the force distribution module is to calculate the desired posture R required to produce the desired total force f and total torque τ dand the total thrust command of the thrusters (in vector form).

[0419] To generate the desired aerodynamic lift L wing_d , the wing needs to achieve a specific lift coefficient C L,target :

[0420]

[0421] in:

[0422] ρ: air density.

[0423] S ref : Wing reference area.

[0424] The aircraft needs to be adjusted to a specific angle of attack α d To achieve this C L,target If a linear lift region is assumed, then α d ≈C L,target / C Lα , where C Lα is the slope of the lift line.

[0425] Expected posture R d The calculation logic will take into account the angle of attack α d , meet the lateral force requirements (through the sideslip angle β d or rudder) and tracking the desired roll angle φ d During the transition, the R d The calculations will smoothly transition from the attitude control logic of a multi-rotor-like aircraft (mainly generating horizontal force by tilting the thrust vector) to the attitude control logic of a fixed-wing aircraft (mainly controlling the flight path by adjusting aerodynamic control surfaces and angle of attack).

[0426] Smooth transition of thruster operating modes:

[0427] The control distribution module converts the total desired thrust f T (the force generated by the thruster section calculated by the force distribution module) and the desired torque τ T (The total desired control torque, provided primarily by the propellers at low speeds, or by the aerodynamic control surfaces and propeller differential at high speeds) is distributed to the various propulsion actuators (e.g., rotors, thrust fans, control surfaces).

[0428] For tilt-rotor / tilt-ducted aircraft, the tilt angle δ tilt will serve as an additional control variable, and its change should also be related to s trans Synchronous and smooth:

[0429] δ tilt (s trans )=δ VTOL +strans (δ Cruise -δ VTOL )

[0430] For example, from the vertical (δ VTOL ≈90°) to horizontal (δ Cruise ≈0°) smooth tilt.

[0431] Thrust command T for each thruster j Will be based on f T ,τ T and the current δ tilt It is calculated by a control allocation algorithm (such as a pseudo-inverse method, an optimization method, or a recursive allocation method) to ensure that the instruction changes are continuous and smooth.

[0432] Quantitative Analysis and Guarantee Measures of Aircraft Stability during Transition

[0433] Gain Scheduling:

[0434] The feedback gains of the nonlinear trajectory controller and attitude controller (e.g., the position error gain matrix K p , velocity error gain matrix K v , attitude error gain k q , angular velocity error gain matrix K ω ) can be calculated based on the transition process parameter s trans or airspeed V air To schedule:

[0435] K(s trans )=K VTOL ·λ gain (s trans )+K Cruise ·(1-λ gain (s trans ))

[0436] where K VTOL is the gain value applicable to vertical take-off and landing mode, K Cruise is the gain value applicable to cruise mode. gain (s trans ) is a lift Similar but possibly different change rates or shapes of the dispatch factors ensure that the closed-loop system has good dynamic performance and stability during the entire transition process. For example, λ gain (s trans ) can be designed as:

[0437]

[0438] Alternatively, a more complex nonlinear scheduling function can be designed based on the analysis results of a specific mode (such as the damping ratio of the short-period mode).

[0439] Stability Margin Assessment and Maintenance:

[0440] By measuring the aircraft at different transition state points (different s trans The linearized model of the modal damping ratio ζ is analyzed to evaluate its characteristic m and the natural frequency ω nm Ensure that the damping ratio and frequency of key flight modes (such as pitch short period, Dutch roll) meet the design requirements (for example, ζ m >ζ min If the analysis shows that the stability margin of some sections is insufficient, the gain scheduling function K(s) needs to be adjusted inversely. trans ) or the transition envelope itself.

[0441] Active disturbance suppression:

[0442] The disturbance observer or specific disturbance suppression loop included in the controller (such as the robust adaptive part mentioned above) estimates and compensates for unknown force and torque disturbances caused by airflow changes, propeller-wing complex interference, etc. during the transition phase to maintain smooth flight.

[0443] Through the synergistic effects of the flight envelope planning and tracking described above, including specific formula examples, the smooth integration strategy of aerodynamic forces and propeller thrust, and stability assurance measures, the control system proposed in the present invention can effectively achieve safe, smooth, and efficient transitions between different flight modes for a flying car.

[0444] Example 4:

[0445] This embodiment provides a flying car control system that balances control accuracy and flight stability, and is used to implement the above-mentioned flying car control method that balances control accuracy and flight stability, including:

[0446] A receiving module, used to receive aircraft system status and environmental parameters;

[0447] A decomposition module is used to construct a six-degree-of-freedom dynamic model of the vertical take-off and landing aircraft based on the aircraft system status and environmental parameters, and decompose the external forces and moments acting on the dynamic model into the aerodynamic force from the wing and the thrust from the propeller;

[0448] A propeller model building module is used to build a propeller model based on the decomposed propeller thrust and calculate the thrust and torque of the propeller model;

[0449] A wing model building module is used to build an aerodynamic force and moment model of the wing according to the aerodynamic force from the wing, and calculate the force and moment on the wing in combination with the relative wind speed;

[0450] The target total force calculation module is used to build a nonlinear trajectory controller and attitude controller. The calculated thrust and torque of the propeller model, and the force and moment on the wing are input into the nonlinear trajectory controller to calculate the total desired force and total desired moment required by the aircraft.

[0451] A force distribution model construction module is used to construct a force distribution model. According to the dynamic model of the aircraft and the force distribution space, the desired thrust is distributed into thrusts in different directions, and the desired thrust command, the desired torque command, and the desired attitude are calculated. The desired thrust distribution is adjusted according to the flight speed.

[0452] The control distribution model construction module is used to construct the control distribution model and adopt the recursive control algorithm to convert the desired thrust command and the desired torque command into the aircraft control signal to achieve smooth control of the aircraft during the take-off, cruise and landing transition stages.

[0453] It should be noted that, in this document, relational terms such as first and second, etc. are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

Claims

1. A flying car control method that balances control accuracy and flight stability, applied to the field of vertical take-off and landing aircraft control, characterized by: The following steps are involved: Receive aircraft system status and environmental parameters; A six-degree-of-freedom dynamic model of the vertical take-off and landing aircraft is constructed based on the aircraft system status and environmental parameters. The external forces and moments acting on the dynamic model are decomposed into the aerodynamic force from the wings and the thrust from the propellers. A propeller model is constructed based on the decomposed propeller thrust, and the thrust and torque of the propeller model are calculated; The aerodynamic force and moment model of the wing is constructed based on the aerodynamic force from the wing, and the force and moment on the wing are calculated in combination with the relative wind speed; Construct a nonlinear trajectory controller and attitude controller, input the calculated thrust and torque of the propeller model, and the force and moment on the wing into the nonlinear trajectory controller to calculate the total desired force and total desired moment required by the aircraft; Construct a force distribution model, distribute the desired thrust into thrust in different directions according to the aircraft's dynamic model and force distribution space, and calculate the desired thrust command, desired torque command, and desired attitude. The desired thrust distribution is adjusted according to the flight speed. A control distribution model is constructed, and a recursive control algorithm is used to convert the desired thrust instructions and desired torque instructions into aircraft control signals to achieve smooth control of the aircraft during the take-off, cruise, and landing transition stages.

2. The flying car control method according to claim 1, wherein: The system status includes the aircraft position, speed, attitude, and angular velocity; the environmental parameters include wind speed and airflow.

3. The flying car control method according to claim 2, wherein: A six-degree-of-freedom dynamic model of the vertical take-off and landing aircraft is constructed based on the aircraft system status and environmental parameters. The external forces and moments acting on the dynamic model are decomposed into the aerodynamic force from the wing and the thrust from the propeller, as follows: (31) The six-degree-of-freedom dynamic model of a vertical take-off and landing aircraft is studied. The system state is determined by the inertial position of the center of mass of the aircraft. speed and the angular velocity in the fuselage coordinates The definition is as follows: Where, is the inertia matrix in the aircraft body coordinate system; g is the gravity constant vector in the inertial system; It is a skew-symmetric mapping, a×b=S(a)b; (32) The forces and moments of the external environment on the aircraft are grouped uniformly as f b and τ b , specifically broken down into: f b =f T +f A ,t b =t T +t A Where f T and τ T are the thrust and torque from the propeller, respectively, and f A and τ A are the force and moment from the wing, respectively.

4. The flying car control method according to claim 3, wherein: The propeller model is constructed based on the decomposed propeller thrust, and the thrust and torque of the propeller model are calculated, specifically the thrust and torque of the electric propeller: (41) The axial thrust T and torque τ are the force and moment generated when the propeller accelerates: T=C T ρd 4 n 2 τ=C Q ρd 5 n 2 Where C T 、C Q is a dimensionless coefficient, ρ is the air density, d is the propeller diameter, and n is the propeller speed; The propulsive force generated by the jth propeller in the fuselage coordinate system is defined as is the unit vector in the propulsion direction, and the axial torque is r j Defined as the vector pointing from the center of mass of the fuselage to the rotor axis, the torque of the rotor j on the aircraft at the center of mass is expressed as: Where, τ j,cm : represents the torque generated by the jth rotor at the center of mass of the aircraft; r j : The vector from the center of mass of the fuselage to the j-th rotor axis; T j : the magnitude of the axial thrust generated by the j-th propeller; The unit vector in the propulsion direction of the jth propeller; τ j : In the first line of the formula Here, τ j Refers to the axial torque vector generated by the j-th propeller; S(r j ): is the vector r j The skew-symmetric matrix representation of C Q : dimensionless torque coefficient; d: propeller diameter; C T is the dimensionless thrust coefficient; μ j Represents the j-th thruster unit thrust T j The resulting torque contribution vector; Under the premise of knowing the control of each axial thruster, the thrust f of the thruster model is obtained T and torque τ T : Where B T represents the matrix; f T : represents the total thrust vector generated by all thrusters on the aircraft; τ T : represents the total torque vector generated by all thrusters on the aircraft around the center of mass of the aircraft; represents the unit vector in the thrust direction of the j-th propeller; μ j : represents the j-th propeller unit thrust T j The torque contribution vector generated; N T : Indicates the total number of thrusters on the aircraft; T j : represents the magnitude of the axial thrust generated by the j-th propeller; B T : thruster control effectiveness matrix; u T : Control input vector of the thruster.

5. The flying car control method according to claim 4, wherein: The aerodynamic force and moment model on the wing is constructed based on the aerodynamic force from the wing, and the force and moment on the wing are calculated in combination with the relative wind speed, as follows: v w is the ambient wind speed, v is the inertial speed of the aircraft, Transpose the rotation matrix of the aircraft; Define V ∞ , V xz and V xy are the free stream wind speed and the expected speed in the xz plane and xy plane of the aircraft, respectively, which are expressed as: V ∞ =||v i || Where V ∞ is the incident wind speed, α is the angle of attack, and β is the sideslip angle; Define the lift L and drag D in the xz plane of the aircraft, define the lateral force Y on the y axis of the aircraft, and similarly define the aerodynamic torque in each axis: Where S ref is the wing reference area, is the chord width, C L ,C D ,C Y is the non-dimensional coefficient of lift, drag and lateral force; The moments of the aircraft around the x, y, and z axes are calculated using the following formulas: C mx ,C my ,C mz is the non-dimensional coefficient of moment, ω x ,ω y ,ω z is the angular velocity of the aircraft, ΔC mx ,ΔC my ,ΔC mz is the additional effect of the control surface deflection on the torque; let the control surface angle δ Aj The linear effect on the torque, the additional torque coefficient can be expressed as: In summary, the aerodynamic torque is expressed as: Where τ′ A Determined by aerodynamic characteristics, B A u A Caused by control surface deflection; τ A : represents the total aerodynamic torque vector of the aircraft; ρ: represents the air density; S ref : represents the reference area of ​​the wing; c: represents the average aerodynamic chord length or reference chord length of the wing; V ∞ : Indicates the incident wind speed of the aircraft relative to the airflow; C' mx ,C' my ,C' mz : The dimensionless aerodynamic moment coefficients of the aircraft around the x, y, and z axes of the aircraft coordinate system; τ' A : represents the aerodynamic torque part determined by the aerodynamic characteristics of the aircraft itself, excluding the influence of the control surface deflection; N A : represents the total number of pneumatic control surfaces; δ A,j : represents the deflection angle of the jth aerodynamic control surface; B A : represents the aerodynamic control effectiveness matrix; u A : represents the vector composed of the deflection angles of the aerodynamic control surfaces, that is, 6. The flying car control method according to claim 5, wherein: The non-dimensional coefficients of lift, drag, and lateral force are calculated using a flat plate model, a linear model, or a hybrid model, as follows: The flat plate model, used in the high angle of attack region, gives a large-scale trend: Where, represents the lift coefficient calculated under the flat plate model; k L : represents a proportional coefficient used to calculate the lift coefficient in the flat plate model; α: represents the angle of attack or angle of attack of the aircraft, that is, the angle between the direction of the aircraft's speed and the chord line of the wing; represents the resistance coefficient calculated under the flat plate model; k D1 : represents a proportional coefficient used to calculate the resistance coefficient in the flat plate model; k D0 : represents another constant coefficient used to calculate the resistance coefficient in the flat plate model; represents the lateral force coefficient calculated under the flat plate model; k Y : represents a proportional coefficient used to calculate the side force coefficient in the flat plate model; β: represents the sideslip angle of the aircraft, that is, the angle between the projection of the aircraft's velocity direction on the horizontal plane and the longitudinal axis of the fuselage; Linear model for low angle of attack range: Where, represents the lift coefficient calculated under the linear model; C L0 : represents the lift coefficient at zero angle of attack, that is, the lift coefficient when the angle of attack α is zero; C L1 : represents the slope of the lift line, that is, the rate at which the lift coefficient changes with the angle of attack; α: represents the angle of attack of the aircraft; represents the drag coefficient calculated under the linear model; C D0 : Indicates the drag coefficient at zero lift (usually corresponding to near zero angle of attack), also known as zero lift drag coefficient or waste drag coefficient; C D1 : represents the coefficient of linear variation of drag coefficient with angle of attack α; C D2 : represents the coefficient of drag coefficient changing with the square of angle of attack α, usually related to induced drag; represents the lateral force coefficient calculated under the linear model; C Lβ :According to the formula This item represents the proportional coefficient of the side force coefficient with the sideslip angle β; β: represents the sideslip angle of the aircraft; Mixed model, mixing the flat model and the linear model through the tanh function: Where, f A : represents the total aerodynamic force vector of the aircraft; ρ: represents the air density; S ref : represents the reference area of ​​the wing; V ∞ : Indicates the incident wind speed of the aircraft relative to the airflow; C L (α): represents the lift coefficient calculated by the hybrid model and changes with the angle of attack α; C D (α): represents the drag coefficient calculated by the hybrid model and changes with the angle of attack α; α: represents the angle of attack of the aircraft; C Y (β): represents the side force coefficient calculated by the hybrid model and changes with the sideslip angle β; β: represents the sideslip angle of the aircraft.

7. The flying car control method according to claim 6, wherein: Construct a nonlinear trajectory controller and attitude controller, input the calculated thrust and torque of the propeller model, and the force and moment on the wing into the nonlinear trajectory controller, and calculate the total desired force and total desired moment required by the aircraft, as follows: (71) Combining the forces and moments from the propeller model and the wing, the resultant forces and moments acting on the aircraft are unified as follows: Where, f b : represents the total external force vector acting on the aircraft body coordinate system; τ b : represents the total external torque vector acting on the aircraft body coordinate system; f A (v i ): represents the aerodynamic force generated by the wing, which is the relative wind speed v i function of τ' A (v i ω): represents the part of the aerodynamic torque determined by the aerodynamic characteristics of the aircraft itself, excluding the torque generated by the deflection of the control surface; B A : represents the aerodynamic control effectiveness matrix. It maps the deflection angle of the aerodynamic control surface to the resulting additional aerodynamic torque; u A : represents the control vector composed of the deflection angles of each aerodynamic control surface; B T : represents the thruster control effectiveness matrix. It maps the control input of each thruster to the total thrust and total torque they produce; u T : represents the control input vector of the thruster; According to B A u A and B T u T Direct inversion is used to generate the resultant forces and moments, f A The existence of Rf indicates that there are difficulties in determining the flight attitude. b and τ b In dynamics, position and attitude are used as control inputs, and and Solve as the desired force and torque, output the desired posture and ensure Rf b Approaching (72) Trajectory tracking controller: In tracing a given path p d (t), the resulting position error is defined as Design a manifold that makes the position error converge to 0 Λ p is the positive definite gain matrix for the position error, and a trajectory tracking controller using the desired net force is constructed, defined as: Among them, K v and K p is a positive definite gain matrix, assuming The difference between its realized value is If the difference Bounded by a matrix Then v→v r and p→p d In a matrix ∈ and the gain matrix Λ p ,K p and K v The controlled sphere changes exponentially; (73) Design of posture tracking controller: Define the attitude error matrix R d is the target pose matrix, and R is the current actual pose matrix. The error function used in the rotation group SO(3) is equivalent to the error function from The error quaternion Where, Represents the scalar part of the error quaternion; error quaternion Used to express the difference between the desired posture and the actual posture; represents the attitude error matrix, defined as where R d is the target posture matrix, R is the current actual posture matrix; Represents the vector part of the error quaternion; Represents the attitude error matrix The transpose of The attitude change of the aircraft is described by the following formula: In the formula The time derivative of the attitude error matrix, which represents the rate of change of the aircraft attitude error over time; The attitude error matrix is ​​usually defined as the relationship between the desired attitude and the actual attitude; R: the actual attitude matrix, which refers to the current attitude of the aircraft and is a rotation matrix; The derivative of the actual attitude matrix with respect to time; The time derivative of the desired attitude matrix; S(·): skew-symmetric matrix operator, which converts a three-dimensional vector into a 3×3 skew-symmetric matrix, used to represent the cross product operation of the vector, that is, S(a)b=a×b; ω: the actual angular velocity of the aircraft; ω d : expected angular velocity or target angular velocity; The control rules are defined as: Where τ is the control torque, which is the torque that needs to be applied to the aircraft to achieve the desired attitude and angular velocity; J is the moment of inertia matrix of the aircraft, which is a symmetric positive definite matrix; The time derivative of the reference angular velocity; ω r : reference angular velocity; K ω : angular velocity error gain matrix; Angular velocity error, defined as the difference between the actual angular velocity and the reference angular velocity; k q : attitude quaternion error gain coefficient; The vector part of the attitude error quaternion. Quaternions are often used to represent three-dimensional rotations, and their vector part is related to the rotation axis and rotation angle; K ω is a positive definite matrix, k q >0.

8. The flying car control method according to claim 7, wherein: Construct a force distribution model. Based on the aircraft's dynamic model and force distribution space, distribute the desired thrust into thrusts in different directions. Calculate the desired thrust command, desired torque command, and desired attitude. The desired thrust distribution is adjusted according to the flight speed: (81) Obtained by changing actuator commands or aircraft attitude and This process is the distribution of force, which is achieved by solving the desired posture and the desired thrust command. First, the desired posture R is obtained by the following formula: d : (82) Solve for the desired thrust command by minimizing the force residual where f T Directly from dynamic fast input u T Decision, and f A Depends on attitude and speed; for fixed-wing forward-flying vertical take-off and landing aircraft, first use the wing lift to make f A Prioritize meeting expectations The rest is made up of T supply; (83) Determination of expected posture: At low speed: In the object coordinate system, The unit vector representing the favorable thrust direction, the current thruster output is: R d Obtained by any rotation that satisfies the following conditions: At high speed: make represents the required aircraft thrust, given the measured incident wind speed, and the incident wind coordinate system axes are defined as: The incident wind coordinate system is expressed in the fuselage coordinate system as: According to the aerodynamic force and moment model of the wing, by setting an ideal angle of attack α d , giving priority to raising the wing over generating lift perpendicular to the wind speed in Incident wind coordinate system Axis rotation α d , represented by the rotation matrix R a ; The desired pose is calculated by successive rotations: R d =RR i R α (84)Thrust distribution: At the current attitude, the thrust of the thruster is calculated by the following formula: Where M represents the matrix, which shields the body axial force components that cannot be realized by the propeller input; f A Estimated by the current incident wind speed.

9. The flying car control method according to claim 8, wherein: A control allocation model is constructed, and a recursive control algorithm is used to convert the desired thrust and torque commands into aircraft control signals to achieve smooth control of the aircraft during takeoff, cruise, and landing transitions: (91) Control distribution: After derivation, the required thrust and torque of the propeller are obtained, which can be expressed as follows: Define the propulsion vector ω T Define the feasible control space of distributed propulsion for the combination of thrust and torque of the thrusters The goal of control allocation is to find So that: For the drive system, when N T >N w When , there does not exist a single inverse mapping that recovers the complete set of feasible w without violating the control constraints. T ; But because B T There are multiple right inverses, and there is still more than one B T Satisfy the above formula; N T Indicates the number of independent thrusters, N w Denotes the dimensions of the resultant total thrust and total torque vectors; (92) Reachable thruster propulsion space Command control space Its boundary is A feasible propulsion vector ω T Located in B T In the column space of The accessible propulsion space is defined as: where u T,j is the vector u T The jth element of b j B T The j-th column vector of Assuming that each thruster is identical, u T,min =0,u T,max =1, definition: Among them, Conv represents the convex hull of a given set, for Minkowski and; (93) Static and recursive control allocation: (93.1) The recursive allocation method iteratively calculates the pseudo-inverse or generalized inverse solution, giving priority to satisfying the requirements of the main forces, and then distributes the remaining moments. The recursive control algorithm flow is as follows: Recursive control algorithm input: B τ : control allocation matrix, describing the contribution of thrust devices to the total force and total torque; w τ : total target force, including target force and torque; Recursive control algorithm output: u τ : The thrust output of each thruster meets B τ u τ =w τ And u τ,min ≤u τ ≤u τ,max ; (93.2) Steps: (93.21) Check the target total force w τ feasibility, if the target total force w τ Out of reach or approximate range Then for w τ Scale it back to within the feasible range; (93.22) Pseudo-inverse calculation initial solution If B τ Is a square matrix, directly calculate the inverse matrix: If B τ If is a rectangular matrix, the pseudo-inverse is used: (93.23) Check constraints: For each u τ The component u τ,j : If u τ,i >u τ,max , will u τ,max Assign to u τ,i ; If u τ,i τ,min , will u τ,min Assign to u τ,i ;​ (93.24) Update target force and distribution matrix According to the fixed thrust component u τ,i , update the target force w τ and the allocation matrix B τ , delete the assigned columns; (93.25) Recursive call For the updated target w τ and matrix B τ , recursively call the control allocation algorithm until the remaining matrix B τ Become a square; (93.26) Return result Merge the allocation results u of all recursive levels τ , and returns.

10. A flying car control system that takes into account both control accuracy and flight stability, used to implement the flying car control method that takes into account both control accuracy and flight stability as described in any one of claims 1 to 9, characterized in that: include: A receiving module, used to receive aircraft system status and environmental parameters; A decomposition module is used to construct a six-degree-of-freedom dynamic model of the vertical take-off and landing aircraft based on the aircraft system status and environmental parameters, and decompose the external forces and moments acting on the dynamic model into the aerodynamic force from the wing and the thrust from the propeller; A propeller model building module is used to build a propeller model based on the decomposed propeller thrust and calculate the thrust and torque of the propeller model; A wing model building module is used to build an aerodynamic force and moment model of the wing according to the aerodynamic force from the wing, and calculate the force and moment on the wing in combination with the relative wind speed; The target total force calculation module is used to build a nonlinear trajectory controller and attitude controller. The calculated thrust and torque of the propeller model, and the force and moment on the wing are input into the nonlinear trajectory controller to calculate the total desired force and total desired moment required by the aircraft. A force distribution model construction module is used to construct a force distribution model. According to the dynamic model of the aircraft and the force distribution space, the desired thrust is distributed into thrusts in different directions, and the desired thrust command, the desired torque command, and the desired attitude are calculated. The desired thrust distribution is adjusted according to the flight speed. The control distribution model construction module is used to construct the control distribution model and adopt the recursive control algorithm to convert the desired thrust command and the desired torque command into the aircraft control signal to achieve smooth control of the aircraft during the take-off, cruise and landing transition stages.

Images