An adaptive control allocation method for tilt-rotor UAV

Abstract

The present invention discloses an adaptive control allocation method for a tilt-rotor unmanned aerial vehicle (UAV), belonging to the field of UAV flight control. The method comprises: modeling the tilt-rotor UAV to obtain a control allocation equation between the control variable and the actuator, and determining the uncertainty parameters present during flight; separating the uncertainty parameters into additive and multiplicative forms to design the output equation for the control allocation; designing a reference model based on the principle of model-referenced adaptive control; and designing an adaptive law using the Lyapunov stability analysis method. The present invention utilizes an adaptive law to achieve steady-state control of the aircraft under different pitch angles and variable pitch angle conditions, and can address the uncertainty of platform parameters in the UAV during the control allocation process.

Description

Technical Field

[0001] The present invention relates to a control method for a tilt-rotor dual-rotor unmanned aerial vehicle (UAV), in particular to a control distribution link and adaptive control, and belongs to the technical field of flight control. Background Art

[0002] Currently, most micro-UAVs feature multi-rotor configurations, with quadcopters in particular being widely used in civilian applications. However, the large number of motors reduces efficiency, making endurance a major constraint on the development of quadcopters. In recent years, dual-rotor configurations have garnered attention due to their higher efficiency. Compared to other multi-rotor UAVs, dual-rotor configurations offer the highest aerodynamic efficiency, thus improving endurance at the same weight. However, dual-rotor UAVs also present a problem: they are sensitive to center of gravity shifts. The primary reason for this issue is that dual-rotor UAVs utilize the same roll control as quadcopters, controlled by differential rotation of the left and right motors. However, pitch control requires servos to change the direction of motor thrust. In the pitch channel, the direction of motor thrust must pass through the center of gravity to achieve hovering. Therefore, if the center of gravity shifts during hovering, the direction of the motor thrust must also change accordingly to maintain torque balance. Consequently, the steady-state attitude angle of the aircraft will no longer be 0 degrees, and this angle will change with the shift in the center of gravity. The change of steady-state attitude angle will reduce the accuracy of control allocation and affect the mathematical expression from position control loop to attitude control loop, which will greatly affect the control accuracy.

[0003] For existing multi-rotor UAV control algorithms, considering the uncertainty of platform parameters is an issue that deserves special attention. Existing adaptive control algorithms can handle the effects of uncertainty in parameters such as mass, moment of inertia, and center of gravity of quadrotor UAVs. For example, CN 116736705 A discloses an adaptive attitude control method for quadrotor UAVs. However, for tilt-rotor UAVs, there are still some deficiencies, which are mainly reflected in the following aspects: (1) The adaptive control method based on quadrotor configuration design cannot handle the special effects brought about by the change of the steady-state attitude angle of the dual rotor. (2) The traditional adaptive control derivation process based on the backstepping method is complicated and deeply integrated with the design of the upper-level controller. It cannot be applied to the adaptive control of tilt-rotor UAVs and needs to be redesigned after replacing the upper-level controller algorithm.

[0004] The applicant's prior patent CN118953677A discloses a hybrid-wing aircraft. Its powertrain includes a master tilt-steering servo and two rotors, or left and right servos, located on either side of the master tilt-steering servo. The master tilt-steering servo allows the aircraft to actively change its fuselage angle. The principle is that in the hovering state, the line connecting the tension point and the center of gravity is always vertically downward. By rotating the master tilt-steering servo, the vector from the motor tension point to the center of gravity is changed, thereby changing the steady-state attitude angle of the aircraft, enabling the tilt-steering servo to hover at different attitude angles. By changing the steady-state attitude angle of the fuselage during flight, the aircraft can adapt to complex and confined spaces and enhance its ability to explore unknown and confined environments. In terms of algorithm implementation, the effects of master tilt-steering servo rotation are similar to those of center of gravity shift, both manifesting in changes in the moment arm parameters in the control allocation matrix. Therefore, overcoming the platform parameter uncertainty, particularly the uncertainty in the center of gravity position, in dual-rotor UAVs is a research direction in this field. Therefore, the influence of the total tilt servo rotation and the center of gravity offset are both reflected in the change of the lever arm parameters in the control distribution matrix. Therefore, overcoming the platform parameter uncertainty in the twin-rotor UAV, especially the uncertainty of the center of gravity position, is one of the research directions in this field.

[0005] Furthermore, since the steady-state pitch angle during hovering can be adjusted by changing the angle of the total tilt servo, however, strange problems may occur in the case of large pitch angles. Summary of the Invention

[0006] The first objective of the present invention is to address the impact of platform parameter uncertainty in existing dual-rotor UAVs by providing an adaptive control allocation method to eliminate this impact. To this end, the present invention provides an adaptive control allocation method for a tilt-rotor UAV. This method utilizes the principle of model reference adaptation to construct an adaptive law in the control allocation process, thereby obtaining an estimate of the uncertain parameters.

[0007] The second object of the present invention is to solve the problem that when a dual-rotor UAV with a total servo adjusts the steady-state pitch angle during hovering by changing the angle of the total tilt servo, a large pitch angle may occur, thereby causing a strange problem.

[0008] To achieve the first objective of the present invention, the present invention provides an adaptive control distribution method for a tilt-rotor dual-rotor unmanned aerial vehicle (UAV), comprising a fuselage and a power unit disposed on the fuselage, the power unit comprising a master tilt servo capable of tilting forward from a vertical position, and independently tiltable left and right servos symmetrically disposed on either side of the master tilt servo, the left and right servos being fixedly connected to the master tilt servo and capable of rotating as the master tilt servo rotates; the power unit providing flight power and fuselage tilt control; and the fuselage tilt control method comprising the following steps:

[0009] S1: Model the tiltrotor UAV and obtain the mapping equation between the control variable and the actuator, namely the control allocation equation; and determine the uncertain parameters in the flight process;

[0010] S2: Design an adaptive law for uncertain parameters:

[0011] S21. Separate the uncertainty parameters into additive and multiplicative ones, and design the output equation for control allocation.

[0012] S22. Design a reference model based on the principle of model reference adaptive control.

[0013] S23. Design of adaptive laws using Lyapunov stability analysis method.

[0014] Furthermore, the adaptive control allocation method for the tilt-rotor UAV of the present invention, in step S1, determines that the actual control input of the tilt-rotor UAV is , the control distribution equation of the tilt-rotor UAV is established as:

[0015] (5);

[0016] Among them: E 1 、 E 2 It is to simplify the intermediate quantity of the control allocation matrix:

[0017] (6);

[0018] in, The virtual control quantity calculated by the upper-level attitude controller includes and , Represents the projection of the combined acceleration generated by the left and right motors in the z-axis direction of the body, which is the command value of the nominal pulling force The projection on the z-axis of the body, It is a three-axis (body coordinate system ) Angular acceleration command; B Assign a matrix to the control;T L , T R Indicates the pulling force command of the left and right motors; δ L , δ R Represents the rotation angle command of the left and right servos; Represents the rotation angle command of the total tilt servo; m Represents the mass of the body; is the three-axis moment of inertia of the body; L 、 D represents the moment arm; L p ，L r ，D P are the distance vectors from the center of gravity of the whole machine to the axis of the total tilt steering gear The projection on the three axes of the machine system, u m Represents the components of the left and right motor pulling forces on the x-axis and z-axis respectively.

[0019] Furthermore, in the adaptive control allocation method for the tilt-rotor UAV of the present invention, in step S1, the uncertainty parameters include the body mass m， Three-axis moment of inertia of the body , the center of gravity position of the whole machine: the distance vector from the center of gravity of the whole machine to the axis of the total tilt steering gear Projection of the three axes of the machine system L P ,L r ,D P， Lever L 、 D .

[0020] Furthermore, in the adaptive control allocation method for the tilt-rotor UAV of the present invention, in step S21, the uncertainty parameters are first separated into additive and multiplicative properties, and the control allocation matrix B is split into:

[0021] (9);

[0022] in B 0 、 B 1 represents the matrix after the control allocation matrix is ​​split; is the platform parameter matrix, which contains all the uncertainty parameters in the control allocation matrix B;

[0023] Secondly, the output equation form of the control distribution is designed. The control distribution equation (5) is inverted and substituted into (9) to obtain:

[0024] (12);

[0025] Use the estimated value of the corresponding platform parameter matrix replace , and perform compensation to obtain the output equation of control distribution as follows:

[0026] (13);

[0027] in, is the estimated value of the platform parameter matrix that needs to be adapted; Δ represents the compensation coefficient; B 0 、 B 1 represents the matrix after the control allocation matrix is ​​split, The control quantity calculated by the upper-level attitude controller.

[0028] Furthermore, in the adaptive control allocation method for the tilt-rotor UAV of the present invention, in step S21, the Sherman-Morrison formula is used to decompose the inverse of the estimated value in the control allocation matrix B, and the Split into additive form, specifically:

[0029] The matrix B 0 、 B 1 It is obtained by separating the additive and multiplicative uncertainty parameters, that is, the uncertainty parameters in the control allocation matrix B in Eq. (9) are converted using two platform parameter matrices The control allocation matrix B can be decomposed into , where the platform parameter matrix for:

[0030] (10);

[0031] Where: the matrix B 0 、 B 1 for:

[0032] (11);

[0033] in: B 0. B 1 represents the matrix after the control allocation matrix is ​​split; is the platform parameter matrix, which contains all the uncertainty parameters in the control allocation matrix B.

[0034] Furthermore, in the adaptive control allocation method for the tilt-rotor UAV of the present invention, in step S23, the Lyapunov function is used to calculate the Lyapunov stability. Design the adaptive law separately, first determine The possible range of variation and the projection operator is used to The adaptive law is used to limit the boundaries to prevent the error adaptation in certain cases from causing parameter divergence, and finally the estimated value of the adaptive platform parameter matrix is ​​obtained. , and its derivative expression is:

[0035] (26);

[0036] (27);

[0037] in, is an adjustable parameter and is a constant; is the error of the virtual system; B 11 、 B 12 yes B 1 is split into two matrices of rank 1, namely ,in Depend on B 1 The second and third rows of the matrix are composed of, By matrix B The fourth row of 1 is composed; is the projection operator, used to restrict The range of change, for The derivative of is the virtual control quantity calculated by the upper attitude controller; According to equations (26) and (27), The integration can be used to obtain the estimated value of the platform parameter matrix .

[0038] Furthermore, the adaptive control allocation method of the tilt-rotor UAV described in the present invention is to obtain the error of the virtual system in equations (26) and (27): In step S22, the specific method for designing the reference model according to the principle of model reference adaptive control is:

[0039] Define two state variables ξ and ξ m , defines the error of the virtual system ,in ξ and ξ m The derivative form of is as follows:

[0040] (14);

[0041] (15);

[0042] in is the system state quantity; It is the acceleration disturbance caused by the outside world, included in the measurement value of the acceleration sensor middle; ξ Based on acceleration measurement information and angular acceleration command Update the difference; and are all diagonal Hurwitz matrices;

[0043] According to the error of the virtual system , it can be seen that The derivative of for:

[0044] (16);

[0045] in: .

[0046] In order to achieve the second object of the present invention, the present invention prevents strange problems that may occur when the angle of the total tilt servo is changed to adjust the steady-state pitch angle of hovering.

[0047] Furthermore, the tilt-rotor UAV adaptive control allocation method of the present invention further includes the following steps:

[0048] S3: Establish multiple coordinate systems, establish control allocation equations and control allocation output equations in the multiple coordinate systems, and switch them according to the current total tilt servo and the estimated value of the steady-state pitch angle, so that the control allocation output equation always maintains the uniqueness of the control allocation solution.

[0049] Step S3 is to design a switching control architecture to address the singularity problem that may occur when the tilt-rotor UAV is in a large pitch angle state. By establishing control allocation equations in multiple coordinate systems and switching according to conditions, the uniqueness of the control allocation solution is always maintained, and singularity problems will not occur.

[0050] Specifically, in step S3, three body coordinate systems are constructed , And derive the control allocation equation and the output equation of control allocation respectively, according to the estimated value of the adaptive platform parameter matrix Determine the new steady-state pitch angle estimate of the tiltrotor after the center of gravity shifts , which is expressed as follows:

[0051] (28);

[0052] in Representative Matrix The (3,3)th element of Switch the coordinate system when the steady-state pitch angle estimate Time Control distribution is performed in the coordinate system; when = Time In other cases, the control is distributed in the coordinate system. Control allocation is performed in the coordinate system.

[0053] Using the steady-state pitch angle estimate Switch; define the intermediate variable r and rotation matrix as follows:

[0054] (29);

[0055] The three-axis angular acceleration command calculated by the attitude controller By left multiplication Transform to or In the coordinate system, the total acceleration generated by the left and right motors at the same time (nominal tension command) Component on the z-axis of the body Calculated by the following formula:

[0056] (30);

[0057] in, R Represents the transformation matrix from the machine system to the ground system; , is the identity matrix; is the 3D identity matrix; It is the linear acceleration command calculated by the upper position controller.

[0058] In step S3, the steady-state pitch angle estimate is used Compensate the attitude angle command. A single-axis rotation matrix is ​​constructed and used to perform coordinate transformation on the attitude angle command calculated by the position control loop in traditional dual-rotor control. The new attitude angle command after considering the center of gravity offset compensation is obtained as the reference trajectory of the attitude angle control loop.

[0059] The beneficial effects of the present invention are as follows:

[0060] 1) This invention designs a control law for a tilt-rotor UAV with a main servo. Typically, the accuracy of control methods depends on the physical parameters of the model itself. Large errors in these parameters can affect control effectiveness and even cause system divergence, posing a risk. However, this invention determines the uncertain parameters of the platform during control allocation. Adaptive laws tailored to the characteristics of the platform are designed to estimate the uncertain parameters of the controlled object in real time, eliminating the impact of parameter inaccuracies. This adaptive control allocation method is particularly sensitive to center of gravity shift in tilt-rotor UAVs. The adaptive control allocation method estimates the change in the moment arm after center of gravity shift, compensating for this shift and improving control accuracy under eccentric loads. Furthermore, the adaptive law is decoupled from the upper-layer attitude and position controllers, offering modularity. Designing an adaptive control allocation based on the UAV's mechanical model allows for integration with different upper-layer control algorithms.

[0061] 2) This invention leverages the unique steady-state pitch angle variations caused by center of gravity shift in dual-rotor aircraft. By estimating and compensating for these variations in the steady-state pitch angle in adaptive control allocation, it improves the control accuracy of dual-rotor UAVs. Experimental results demonstrate superior tracking accuracy compared to a PID controller that does not account for center of gravity uncertainty.

[0062] 3) For tilt-rotor twin-rotors with total tilt servos, the ability to actively change the steady-state pitch angle is realized. By establishing a multi-coordinate system, the problem of large pitch angles and strange phenomena occurring when the angle of the total tilt servo is changed to adjust the steady-state pitch angle during hovering is solved. The aircraft can hover at different fuselage angles, enhancing the passability of the twin-rotor UAV in confined spaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 This is a framework diagram of the adaptive control allocation method for a tilt-rotor UAV of the present invention;

[0064] Figure 2 This is a schematic structural diagram of the tilt-rotor UAV of the present invention;

[0065] Figure 3 Schematic diagram of the three axes of the basic coordinate system of the tilt-rotor UAV of the present invention;

[0066] Figure 4 The center of gravity of the tilt-rotor UAV of the present invention is To the main tilt servo axis The distance vector Schematic diagram of projection on three axes;

[0067] Figure 5 、 67 is the switching control law coordinate system of the tilt-rotor UAV of the present invention 、 、 Schematic diagram. DETAILED DESCRIPTION

[0068] This paper first proposes a control allocation model for a tiltrotor UAV through mechanistic analysis and identifies the parameters requiring adaptive estimation. Next, the control allocation output equation is designed, and the Sherman-Morison formula is used to decompose the control allocation equation, separating the uncertain parameters into additive and multiplicative components. Adaptive laws for the estimated parameters are then designed using the Lyapunov stability analysis method. Finally, a switching control architecture is designed to address singularity issues that may arise in tiltrotor UAVs operating at large steady-state angles. The control allocation equation is established in multiple coordinate systems, and switching is performed based on the current total tilt servo and estimated steady-state pitch angle to ensure the uniqueness of the control allocation solution.

[0069] The present invention will be further described below with reference to the accompanying drawings and examples. It should be understood that the examples described below are intended to facilitate understanding of the present invention and do not have any limiting effect on the present invention.

[0070] See also Figure 1 The control architecture of the tilt-rotor UAV is divided into a posture controller and an adaptive control allocation module, wherein the posture controller 300 is based on the position command defined by the user. P d , speed command , acceleration command and yaw angle command Calculate the virtual control quantity: nominal tension instruction , angular acceleration command and attitude angle command R c The adaptive control allocation module 400 uses the virtual control amount calculated by the upper attitude controller and the difference between the reference model and the virtual model The adaptive law can be obtained. Then, through the formula U can be calculated m , U m Represents the components of the left and right motor pulling forces on the x-axis and z-axis respectively. is the estimated value of the platform parameter matrix that needs to be adapted, also known as the adaptive law; Δ represents the compensation coefficient; B0 and B1 are defined in formula (11) and represent the matrix after the control allocation matrix B in formula (5) is split. Finally, the actual control input of the UAV can be calculated by formula (8): ,in T R , T Lare the pulling forces of the right and left motors respectively; They are the rotation angles of the total tilt servo and the left and right servos respectively. The layout of the actuators is shown in Figure 2 .

[0071] Figure 2 A tilt-rotor UAV is presented, comprising a fuselage 100 and a power unit 200 mounted thereon. The power unit comprises a master tilt servo 1, a left servo 2, and a right servo 3. The master tilt servo 1 is mounted on the fuselage so that it can tilt forward from a vertical position. The left and right servos 2 and 3 are symmetrically arranged on either side of the master tilt servo. The left and right servos 2 and 3 can tilt independently and are fixedly connected to the master tilt servo 1 and rotate with the rotation of the master tilt servo 1. The power unit provides flight power and fuselage tilt control. Its structure can be referenced by the applicant's prior patent 2024110505010. The main inventive point of this invention lies in the design of an adaptive control distribution module, the upper layer of which can be combined with different attitude controllers.

[0072] The adaptive control allocation method of the tilt-rotor UAV of the present invention comprises the following steps:

[0073] S1: Model the tiltrotor UAV, obtain the mapping equation between the control variable and the actuator, that is, the control distribution equation, and determine the uncertainty parameters in the flight process; specifically:

[0074] First, the tilt-rotor UAV body coordinate system Definition Figure 3 , Coordinate system The axis points to the nose of the machine. The axis points to the right side of the fuselage, The axis points downwards towards the fuselage.

[0075] According to the principles of flight mechanics, the dynamic and kinematic models of the tilt-rotor UAV can be shown as follows:

[0076] (1);

[0077] (2);

[0078] (3);

[0079] (4);

[0080] in, p 、 v Represent the position and velocity of the body respectively; R represents the transformation matrix from the aircraft system to the ground system; Ω represents the angular velocity of the aircraft; τRepresents the nominal tension, which is the ratio of the resultant force generated by the actuator to the mass of the body; Represents the angular acceleration generated by the actuator; Represents the antisymmetric matrix of Ω, that is, for any vector satisfy ; is the identity matrix; is the moment of inertia, in the three axis directions respectively ; and are the linear acceleration disturbance and the angular acceleration disturbance respectively.

[0081] Determine the actual control input of the tilt-rotor UAV as ,in T R 、 T L are the pulling forces of the right and left motors respectively; are the rotation angles of the total tilt servo and the left and right servos respectively. Therefore, the control distribution equation is designed as follows:

[0082] (5);

[0083] in: E 1. E 2 is to simplify the intermediate quantity of the control allocation matrix B:

[0084] (6);

[0085] in, The virtual control quantity calculated by the upper attitude controller; including and , Represents the projection of the combined acceleration generated by the left and right motors in the z-axis direction of the body, which is the nominal tension command The component on the z-axis of the body; is the three-axis angular acceleration command; B is the control allocation matrix; T L 、 T R Indicates the pulling force command of the left and right motors; Represents the rotation angle command of the right and left servos; Represents the rotation angle command of the total tilt servo; m Represents the mass of the body; is the three-axis moment of inertia of the body; L , D represents the moment arm; L P ，L r，D P Represents the distance vector from the center of gravity to the axis of the total tilt servo The projection of the three axes of the machine system (see Figure 4 ), u m Represents the components of the left and right motor pulling forces on the x-axis and z-axis respectively. Using the control distribution equations (5) and (6), we can calculate u m as follows:

[0086] (7).

[0087] Determine the uncertainty parameters that may exist during the flight process, including: body mass m, body three-axis moment of inertia , the center of gravity position of the whole machine: the distance vector from the center of gravity to the axis of the total tilt steering gear Projection of the three axes of the machine system L P ， L r ，D P ， Lever length L , D .

[0088] Since the total tilt servo is used, the platform is overdriven and can be controlled individually according to the flight environment and mission requirements. Therefore, using Equation (5) u m The expression of the total tilt servo rotation angle given , the actual control input of the platform can be calculated as:

[0089] (8);

[0090] in: Representative vector u m No. i A portion.

[0091] S2, the design process of adaptive control allocation, includes steps S21-S23.

[0092] S21. First, the uncertainty parameters are separated into additive and multiplicative properties. To facilitate the subsequent design, the platform physical parameters (i.e., uncertainty parameters) in the control allocation matrix B in Equation (5) are converted into two platform parameter matrices: Expressed as follows, the control allocation matrix B can be split into:

[0093] (9);

[0094] in:

[0095] (10);

[0096] in, B 0. B 1 represents the matrix after the control allocation matrix B in formula (5) is split:

[0097] (11);

[0098] Inverse equation (5) and substitute it into equation (9), u m It can be expressed as follows:

[0099] (12);

[0100] From the above formula, we can see that all unknown platform parameters are included in the platform parameter matrix These parameters can be obtained through measurement, but there is still a certain degree of uncertainty. The main reasons can be divided into the following three aspects: 1. There will be measurement errors during the measurement process; 2. In flight missions that require different payloads, the loading and unloading of the payload will change the mass. m , moment of inertia J and the center of gravity L P ，L r ，D P ; 3. During the flight process of changing the fuselage attitude angle, the angle of the total tilt servo If a change occurs, the platform parameter matrix will also be and in E 1 and E 2 changes have occurred.

[0101] In summary, the platform parameter matrix There are uncertainties in the flight process, and its value cannot be accurately obtained, so it is necessary to make real-time estimates during the flight. Therefore, in the actual program implementation, it is necessary to replace the platform parameter matrix with the estimated value. , and the control allocation matrix B Unknown, cannot be directly calculated by (7) u m , a new distributive law needs to be designed.

[0102] Secondly, design the output equation form of the control distribution: define the platform parameter matrix and The estimated value is and , design a new distributive law to replace formula (7), as follows:

[0103] (13) ;

[0104] in: is the estimated value of the platform parameter matrix that needs to be adapted, also known as the adaptive law; Δ represents the compensation coefficient; B 0. B 1 is defined in Equation (11) and represents the control allocation matrix in Equation (5) B The matrix after splitting;

[0105] use u m And using formula (8), the control input of the system can be calculated .

[0106] In the design process of adaptive control allocation, the adaptive law and The derivative of : and By integrating the above two derivatives, we can get the estimated value of the platform parameter matrix and The following is the adaptive law and design process.

[0107] S22. Design a reference model based on the principle of model reference adaptive control: Using the model reference adaptive method, define two state quantities ξ and ξ m , its derivative form is as follows:

[0108] (14);

[0109] (15);

[0110] in is the system state quantity; It is the acceleration disturbance caused by the outside world, included in the measurement value of the acceleration sensor middle; ξ Based on acceleration measurement information (measured values) and angular acceleration command Update the difference; and are both diagonal Hurwitz matrices; Formula (14) is a virtual dynamic model that takes into account the control allocation error, and Formula (15) is a closed-loop reference model.

[0111] Defining the error of the virtual system , and we can know The derivative of for:

[0112] (16);

[0113] in , is the error of the virtual system. ξ The influence of control allocation error is added to the dynamic equation of The difference between the virtual dynamics and the closed-loop reference model is reflected, and the adaptive law is designed based on this error.

[0114] S23. Design of adaptive law using Lyapunov stability analysis method: The design of the adaptive law is derived through Lyapunov stability analysis method. By selecting Lyapunov function and taking its derivative, the adaptive law is designed so that the derivative of Lyapunov function is negative definite.

[0115] The designed alternative Lyapunov function is: ,and is a positive definite matrix and satisfies . , They are and The parameter estimation error.

[0116] right V Taking the derivative we get:

[0117] (17);

[0118] in, Adjustable parameters; process the right side Term, use (13) and split the control allocation matrix B We can get:

[0119] (18);

[0120] In the above formula B 0 is a non-singular matrix, and B 1 can be split into the sum of two rank 1 matrices, namely .in Depend on B 1 The second and third rows of the matrix are composed of, Depend on B 1 is composed of the fourth row of the matrix. Therefore, in (18) It can be decomposed into:

[0121] (19);

[0122] Substitute equation (19) into equation (18) and use and , we can get:

[0123] (20);

[0124] Therefore, in formula (17) It can be expressed as:

[0125] (twenty one);

[0126] Finally, design and The form of is used to offset the second and third lines of Equation (21). And the compensation coefficient Δ is designed to satisfy You can get , thereby ensuring the stability of the adaptive control allocation (as long as The system stability can be deduced from the negative definiteness). The designed adaptive law is as follows:

[0127] (twenty two);

[0128] (twenty three);

[0129] On this basis, the projection operator is defined as:

[0130] (twenty four);

[0131] in: and represent and Y No. i 、 j elements; f It is a continuously differentiable convex function with the following form:

[0132] (25);

[0133] in and They are The upper and lower boundaries of is an adjustable parameter and satisfies Using the projection operator defined above to constrain, the final adaptive law is:

[0134] (26);

[0135] (27);

[0136] in: is an adjustable parameter; is the error of the virtual system, is the projection operator, used to restrict range of change.

[0137] According to the above adaptive laws (26) and (27), the estimated value can be calculated and Using formula (13), we can calculate the variable u m Finally, the actual control input of the system can be calculated using formula (8): .

[0138] The control distribution module designed according to the above method can be combined with different position and posture control methods. Figure 1 Finally, different posture controllers calculate the expected angular acceleration based on the expected position command and the current actual position of the body. and the projection of the desired nominal tension on the z-axis As the input of control allocation, see formula (5).

[0139] For a tilt-rotor with a total tilt servo, the steady-state pitch angle during hovering can be adjusted by changing the angle of the total tilt servo. However, due to the nominal thrust command used in the control distribution model, Component on the z-axis As a pull command, therefore, there will be strange problems in the case of large pitch angles. For example: in the hovering state with a pitch angle of 90°, according to the force analysis, the left and right motors need to provide thrust to offset the gravity of the platform. However, at this time, the nominal pull command Component on the z-axis =0, resulting in the left and right motor pulling forces calculated using equations (5) and (8) being T L 、 T R Both are 0. This erroneous thrust command will cause the tilt-rotor to be unable to fly stably.

[0140] Therefore, in order to avoid the impact of singularity problems, a switchable control allocation method is designed. Figure 5 、 6 , 7, respectively build three body coordinate systems and .in X coordinate system B1 Axis pointing to the nose, Y B1 The axis points to the right side of the fuselage, Z B1 The axis points downwards into the fuselage; X coordinate systemB2 The Y axis points upwards. B2 The axis points to the right side of the fuselage, Z B2 The axis points to the nose of the machine; X coordinate system B3 The Y axis points downwards. B3 The axis points to the right side of the fuselage, Z B3 The axis points toward the rear of the fuselage.

[0141] According to the estimated value of the adaptive platform parameter matrix Determine the new steady-state pitch angle estimate of the tiltrotor after the center of gravity shifts , using the steady-state pitch angle estimate Switch the coordinate system, and the expression of the estimated value is:

[0142] (28);

[0143] in Representative Matrix The (3,3)th element of . Estimated value based on the steady-state pitch angle Switch the coordinate system when the steady-state pitch angle estimate Time Control allocation is performed in the coordinate system; Time In other cases, the control is distributed in the coordinate system. Control distribution is performed in the coordinate system. Define the intermediate variable r and the rotation matrix as follows:

[0144] (29);

[0145] The angular acceleration command calculated by the attitude controller is multiplied by the rotation matrix Transform to or In the coordinate system, at the same time, the nominal tension instruction Component on the z-axis of the body Calculated by the following formula:

[0146] (30);

[0147] in: R Represents the transformation matrix from the machine system to the ground system; , is the identity matrix; is the 3D identity matrix; It is the linear acceleration command calculated by the upper position controller.

[0148] By establishing a multi-coordinate system, the problem of a large pitch angle when changing the angle of the total tilt servo to adjust the steady-state pitch angle during hovering is solved, avoiding the occurrence of singular problems.

[0149] The above description is illustrative and non-restrictive of the present invention. The present invention is intended to provide an adaptive control allocation method for a tilt-rotor UAV. The design of an adaptive control allocation module for a tilt-rotor UAV can be combined with various upper-layer position and attitude control methods. Persons skilled in the art will appreciate that various modifications, variations, and equivalents may be made without departing from the spirit and scope of the claims, such as varying the upper-layer attitude control algorithm, and all such modifications and equivalents will fall within the scope of the present invention.

Claims

1. A tilt-rotor UAV adaptive control allocation method, characterized in that: The invention comprises a fuselage and a power unit mounted on the fuselage, wherein the power unit comprises a main tilt servo that can be tilted forward from a vertical position, and left and right servos that can be tilted independently and are symmetrically arranged on either side of the main tilt servo. The left and right servos are fixedly connected to the main tilt servo and can rotate with the rotation of the main tilt servo. The power unit provides flight power and fuselage tilt control. The method comprises the following steps: S1: Model the tilt-rotor UAV, obtain the mapping equation between the control variable and the actuator, that is, the control distribution equation, and determine the uncertainty parameters in the flight process; S2: Design an adaptive law for uncertain parameters: S21. Separate the uncertainty parameters into additive and multiplicative ones and design the output equation for control allocation; S22. Design a reference model based on the principles of model reference adaptive control; S23. Design of adaptive law using Lyapunov stability analysis method; S3: Establish multiple coordinate systems, establish control distribution equations and control distribution output equations in multiple coordinate systems, and calculate the total tilt servo and steady-state pitch angle estimation value based on the current total tilt servo and steady-state pitch angle estimation value. Switching is performed so that the control assignment solution remains unique at all times.

2. The adaptive control allocation method for a tilt-rotor UAV according to claim 1, characterized in that: In step S1, the actual control input of the tilt-rotor UAV is determined to be , the control distribution equation of the tilt-rotor UAV is established as: (5)； Among them: E 1 、E 2 To simplify the control allocation matrix B The intermediate amount: （6）； in, The control quantity calculated by the upper attitude controller includes and , Represents the projection of the combined acceleration generated by the left and right motors on the z-axis direction of the body, Represents the three-axis angular acceleration command; B Assign a matrix to the control; T L ，T R Indicates the pulling force command of the left and right motors; δ L , δ R Represents the rotation angle command of the left and right servos; δ M Represents the rotation angle command of the total tilt servo; m Represents the mass of the body; J X , J Y , J Z is the three-axis moment of inertia of the body; L , D represents the moment arm; L p , L r , D p are the distance vectors from the center of gravity to the axis of the total tilt servo r p The projection on the three axes of the machine system, u m Represents the components of the left and right motor pulling forces on the x-axis and z-axis respectively.

3. The adaptive control allocation method for a tilt-rotor UAV according to claim 2, characterized in that: The uncertainty parameters that need to be adapted include the body mass m， Three-axis moment of inertia of the body J X、 J Y、 J Z ， Center of gravity position of the whole machine: distance vector from the center of gravity to the axis of the total tilt servo r p Projection of the three axes of the machine system L p、 L r、 D p， Lever L, D .

4. The adaptive control allocation method for a tilt-rotor UAV according to claim 3, characterized in that: In step S21, first, the uncertainty parameters are separated into additive and multiplicative properties, and the control allocation matrix B is split into: （9）； in, B 0 , B 1 Represents the matrix after the control allocation matrix B is split; η 1 , η 2 is the platform parameter matrix, including the control allocation matrix B All uncertainty parameters in ; Secondly, design the output equation for control distribution, invert Equation (5) and substitute it into Equation (9) to obtain: （12）； Use the estimated value of the corresponding platform parameter matrix replace η 1 , η 2 , the output equation of the design control allocation is as follows: （13）； in, is the estimated value of the platform parameter matrix that needs to be adapted; Δ represents the compensation coefficient; B 0 ，B 1 represents the matrix after the control allocation matrix B is split, The control quantity calculated by the upper-level attitude controller.

5. The adaptive control allocation method for a tilt-rotor UAV according to claim 4, characterized in that: The matrix B 0 ，B 1 It is obtained by separating the additive and multiplicative uncertainty parameters, that is, the control allocation matrix in Eq. (9) is B The uncertainty parameters in use two platform parameter matrices η 1 , η 2 Expressed as, then the control allocation matrix B Can be split into ,in: （10）； Where: the matrix B 0 ，B 1 for: （11）； in: B 0 ，B 1 Represents the matrix after the control allocation matrix B is split; η 1， η 2 is the platform parameter matrix, including the control allocation matrix B All the uncertainty parameters in diag represents a diagonal matrix.

6. The adaptive control allocation method for a tilt-rotor UAV according to claim 5, characterized in that: The step S23 is specifically as follows: deriving the estimated value of the adaptive platform parameter matrix by Lyapunov function , its derivative expression is: （26）； （27）； in, is an adjustable parameter and is a constant; is the error of the virtual system; B 11 、B 12 is a matrix B 1 is split into two matrices of rank 1, namely ,in By matrix B 1 The second and third rows of By matrix B 1 The fourth line is composed of; 、 is the projection operator, used to restrict the range of change; for The derivative of The virtual control quantity calculated by the upper-level attitude controller; According to formulas (26) and (27), The integration can be used to obtain the estimated value of the platform parameter matrix .

7. The adaptive control allocation method for a tilt-rotor UAV according to claim 6, characterized in that: The specific method of designing the reference model according to the principle of model reference adaptive control in step S22 is: Define two state variables ξ and ξ m , defines the error of the virtual system ,in ξ and ξ m The form is as follows: （14）； （15）； in, is the system state quantity; It is the acceleration disturbance caused by the outside world, included in the measurement value of the acceleration sensor Bu m + d middle; ξ Based on the acceleration measurement Bu m + d and angular acceleration command Update the difference; and are all diagonal Hurwitz matrices; According to the error of the virtual system , it can be seen that The derivative of for: （16）； in: .

8. The adaptive control allocation method for a tilt-rotor UAV according to claim 1, characterized in that: In step S3, three body coordinate systems are constructed , And derive the control allocation equation and the output equation of control allocation respectively, according to the estimated value of the platform parameter matrix Determine the new steady-state pitch angle estimate of the tiltrotor after the center of gravity shifts , which is expressed as follows: （28）； in, Represents the estimated value of the platform parameter matrix The (3,3)th element of Switch the coordinate system when the steady-state pitch angle estimate When Control allocation is performed in the coordinate system; When In other cases, the control is distributed in the coordinate system. Control allocation is performed in the coordinate system.

9. The method for adaptive control allocation of a tilt-rotor UAV according to claim 8, characterized in that: Using the steady-state pitch angle estimate Switch; define intermediate variables r and the rotation matrix as follows: （29）； Angular acceleration command calculated by the attitude controller By left-multiplying the rotation matrix Transform to or In the standard system, the total acceleration generated by the left and right motors at the same time is Component on the z-axis of the body Calculated by the following formula: （30）； in, R Represents the transformation matrix from the machine system to the ground system; , is the identity matrix; is the 3D identity matrix; It is the linear acceleration command calculated by the upper position controller.

Images