Direct proportional decoupling control method for tilting three-rotor unmanned aerial vehicle

Abstract

The invention discloses a direct proportional decoupling control method for a tilting three-rotor unmanned aerial vehicle, and relates to the technical field of unmanned aerial vehicle flight control. A control distribution model between expected control force and torque of the tilting three-rotor unmanned aerial vehicle and a rotor inclination angle and a rotor rotating speed is established; selecting different working point balancing control distribution models to obtain a balancing data set; constructing a fitting function of a rotor inclination angle with respect to expected control force and torque, establishing a target optimization function based on the trim data set, and optimizing the target function by adopting a Levenberg-Marquardt algorithm to achieve an optimal function fitting effect; under the condition that expected control force and torque are given, through calculation of a rotor inclination angle, linearization of a control distribution model and solving of a rotor rotating speed, calculation complexity is reduced, and decoupling of the rotor inclination angle and the rotor rotating speed in control channel distribution solving is achieved. In practical engineering application, the method can effectively improve the decoupling performance of a control channel, and reduces the resolving complexity of a hybrid controller of a flight control system at the same time.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle flight control, and in particular to a direct proportional decoupling control method for a tilt-rotor tri-rotor unmanned aerial vehicle. Background Art

[0002] Tilt-rotor UAVs have always been a hot topic of research in the engineering community because they have both the short takeoff and landing capabilities of helicopters and the high-efficiency, high-speed cruising capabilities of fixed-wing aircraft. Tilt-rotor tri-rotor UAVs are typical representatives of this type of UAV. Compared with conventional multi-rotor UAVs that only control the movement of the aircraft by adjusting the speed of each rotor, tilt-rotor UAVs need to adjust the rotor speed and rotor pitch angle at the same time in the rotor mode to achieve motion control. In the control distribution process of the former, the expected control force and torque are linearly related to the rotor speed. Therefore, the pseudo-inverse method can be used to directly solve the rotor speed under the conditions of given expected control force and torque; while the latter, due to the introduction of the rotor pitch angle, there are nonlinear cross terms in the control distribution model, and the pseudo-inverse method cannot be used to solve the rotor speed and rotor pitch angle. One current power distribution solution is to introduce proportional control in each channel, but it does not eliminate the coupling between the channels, which causes oscillations during the movement of the aircraft and reduces the flight quality.

[0003] Therefore, it is necessary to provide a direct proportional decoupling control method for a tilt-rotor tri-rotor UAV to solve the above problems. Summary of the invention

[0004] The purpose of the present invention is to provide a direct proportional decoupling control method for a tilt-rotor tri-rotor UAV, which decouples the rotor pitch angle calculation from the rotor speed calculation in the control allocation model solution to achieve linearization of the control allocation model, thereby reducing the complexity of the control allocation solution.

[0005] To achieve the above object, the present invention provides a direct proportional decoupling control method for a tilt-rotor tri-rotor UAV, comprising the following steps:

[0006] S1: Establish a mapping function between the desired control force and torque of the tilt-rotor tri-rotor UAV and the rotor inclination angle and rotor speed to build a control allocation model;

[0007] S2: Select different working points according to the power performance constraints, balance the control distribution model, and obtain the balance data set;

[0008] S3: Construct a fitting function of the rotor pitch angle with respect to the desired control force and torque, establish a target optimization function based on the trim data set, and use the Levenberg-Marquardt algorithm to optimize the target function to achieve the optimal function fitting effect;

[0009] S4: For any control force and torque within the performance constraints of the UAV, the rotor pitch angle is calculated by fitting the function, the control allocation model is linearized, and then the rotor speed is solved to achieve step-by-step decoupling.

[0010] Preferably, in step S1, the mapping function between the desired control force and torque of the UAV established according to the mechanical characteristics analysis and the rotor pitch angle and the rotor speed is as follows:

[0011]

[0012] The control distribution model is established based on the body coordinate system. b x b y b z b Take the center of gravity of the body as the origin of the coordinate system, o b x b Along the longitudinal axis of the aircraft, pointing to the nose, o b y b Perpendicular to the longitudinal symmetry plane, pointing to the right side of the aircraft, o b z b According to the right-hand rule, (x i ,y i ,z i ) represents the position of the rotation center of rotor i in the body coordinate system, i∈{1,2,3} in the tilt-rotor tri-rotor UAV;

[0013] In the formula, C L is the rotor thrust coefficient, C M is the rotor torque coefficient, n i is the rotation speed of rotor i; r i is the direction of rotor rotation, the forward propeller value is 1, and the reverse propeller value is -1;

[0014] The centerline of the nacelle of the rotor i is at o b x b z b Projection in the plane and axis o b x b The angle between the rotor and the i , when the rotor is pointing vertically upward, the tilt angle is 0, and when the rotor is tilted vertically backward, the tilt angle is positive. In the rotor mode, the left and right rotors are differentially controlled, and the tail rotor is kept vertical. The rotor tilt angle satisfies the relationship -α 1 =α 2 =β,α 3 =0, β is the rotor differential angle;

[0015] F z is the desired thrust input to the controller, M x is the desired rolling moment input to the controller, M y is the desired pitch moment input to the controller, M z is the yaw moment input to the controller.

[0016] The control allocation model is abbreviated as:

[0017] c=H(β)n

[0018] Where c = (F z ;M x ;M y ;M z ) is the desired control input; is the control quantity of the actuator.

[0019] Preferably, in step S2, the trimming working point t i Includes the following elements:

[0020] t i ={F zi ,M xi ,M yi ,M zi ,β i ,n 1i ,n 2i ,n 3i}

[0021] Different trim working points t i Compose the balanced data set T:

[0022] T={t 1 ,t 2 ,…,t i ,…t n}.

[0023] Preferably, in step S3, the rotor pitch angle fitting function is:

[0024]

[0025] In the formula, K is the fitting coefficient to be optimized, K = (k 1 ;k 2 ;…k 11 );

[0026] Based on the balanced data set T, the target optimization function is constructed according to the fitting error:

[0027]

[0028] When the objective function Γ is minimized, the optimal fitting parameters are expressed as:

[0029]

[0030] Where f(K) is the residual vector:

[0031]

[0032] Use the LM optimization algorithm to calculate the Jacobian matrix J at the jth iteration j for:

[0033]

[0034] In the formula, Represents partial differential.

[0035] Preferably, in step S4, the optimal fitting parameter K * , under the action of the desired control input c, the rotor inclination angle β is directly obtained by fitting the function:

[0036]

[0037] The rotor inclination angle β is substituted into the control allocation model for linearization to solve the rotor speed n:

[0038] n=H(β) -1 c

[0039] In the formula, H(β) -1 is the generalized inverse matrix of H(β).

[0040] Therefore, the present invention adopts the above-mentioned direct proportional decoupling control method of the tilt-rotor three-rotor UAV to decouple the rotor pitch angle calculation and the rotor speed calculation in the control allocation model solution to realize the linearization of the control allocation model, thereby reducing the complexity of the control allocation solution. In actual engineering applications, this method can reduce the complexity of the flight control mixer solution.

[0041] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flow chart of a direct proportional decoupling control method of a tilt-rotor tri-rotor UAV of the present invention;

[0043] Figure 2 It is a schematic diagram of the appearance structure and power analysis of the tilt-rotor tri-rotor UAV in the present invention;

[0044] Figure 3 It is the descent curve of the objective function of the LM algorithm optimization process in the present invention;

[0045] Figure 4 It is the control tracking effect curve of the direct proportional decoupling control method in the present invention; DETAILED DESCRIPTION

[0046] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0047] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0048] The words "include" or "comprises" and the like used in the present invention mean that the elements before the word include the elements listed after the word, and do not exclude the possibility of also including other elements. The orientation or position relationship indicated by the terms "inside", "outside", "upper", "lower", etc. is based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation of the present invention. When the absolute position of the described object changes, the relative position relationship may also change accordingly. In the present invention, unless otherwise clearly specified and limited, the terms "attachment" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral body; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0049] Example

[0050] like Figure 1 As shown, the present invention provides a direct proportional decoupling control method for a tilt-rotor tri-rotor UAV, comprising the following steps:

[0051] S1: Establish a mapping function between the desired control force and torque of the tilt-rotor tri-rotor UAV and the rotor pitch angle and rotor speed to build a control allocation model; In step S1, the mapping function between the desired control force and torque of the UAV with respect to the rotor pitch angle and rotor speed established according to the mechanical characteristics analysis is as follows:

[0052]

[0053] The control distribution model is established based on the body coordinate system. The UAV structure is as follows: Figure 2 As shown, the body coordinate system o b x b y b z b Take the center of gravity of the body as the origin of the coordinate system, o b x b Along the longitudinal axis of the aircraft, pointing to the nose, o b y b Perpendicular to the longitudinal symmetry plane, pointing to the right side of the aircraft, o b z b According to the right-hand rule, (x i ,y i ,z i) represents the position of the rotation center of rotor i in the body coordinate system, i∈{1,2,3} in the tilt-rotor tri-rotor UAV;

[0054] The thrust f generated by rotor i i and the magnitude of the reaction torque m i Calculated by empirical formula, the force and torque generated by rotors 1, 2, and 3 are shown in the figure below: Figure 2 As shown;

[0055]

[0056] In the formula, C L is the rotor thrust coefficient, C M is the rotor torque coefficient, n i is the rotation speed of rotor i; r i is the direction of rotor rotation, the forward propeller value is 1, and the reverse propeller value is -1;

[0057] The thrust f generated by rotor i i and the reaction torque m i Decompose it in the body coordinate system as follows:

[0058]

[0059] In the formula is the tension f i The components in the body coordinate system, is the reaction torque m i Components in the body coordinate system;

[0060] For the tilt-rotor tri-rotor aircraft, the forces and moments generated by the three rotors are simplified to the coordinate origin, and the resultant force on the drone is obtained as:

[0061]

[0062] The resultant moment is:

[0063]

[0064] The centerline of the nacelle of the rotor i is at o b x b z b Projection in the plane and axis o b x b The angle between the rotor and the i , when the rotor is pointing vertically upward, the tilt angle is 0, and when the rotor is tilted vertically backward, the tilt angle is positive. In the rotor mode, the left and right rotors are differentially controlled, and the tail rotor is kept vertical. The rotor tilt angle satisfies the relationship -α 1 =α 2 =β,α 3 =0, β is the rotor differential angle;

[0065] After simplification and merging, the control allocation model of the tilt-rotor tri-rotor UAV is expressed as:

[0066]

[0067] The control allocation model is abbreviated as:

[0068] c=H(β)n

[0069] Where c = (F z ;M x ;M y ;M z ) is the desired control input; is the control quantity of the actuator.

[0070] In this embodiment, the rotor pull coefficient C L 7×10 -8 , rotor torque coefficient C M For 10 -9 , the rotor position parameters are shown in Table 1:

[0071] Table 1

[0072]

[0073] Since there is a nonlinear term β in the control allocation model, the rotor speed cannot be directly obtained by finding the pseudo-inverse method like conventional multi-rotor UAVs. Therefore, the nonlinear term β in the control allocation model is fitted in advance, and a fitting equation for the rotor pitch angle β with respect to the expected control input c is established. This decouples the calculation of the rotor pitch angle from the calculation of the rotor speed, thereby converting the problem of solving a set of nonlinear equations into a two-step problem of solving linear equations, greatly reducing the computational complexity.

[0074] S2: Select different working points according to the power performance constraints, balance the control distribution model, and obtain a balance data set; in step S2, the balance working point t i Includes the following elements:

[0075] t i ={F zi ,M xi ,M yi ,M zi ,β i ,n 1i ,n 2i ,n 3i}

[0076] Different trim working points t i Compose the balanced data set T:

[0077] T={t 1 ,t 2,…,t i ,…t n}.

[0078] According to the UAV performance constraints, F z (Unit N) takes 7 values ​​evenly in the range [-25, -10], M x ,M y ,M z (unit N·m) takes 11 values ​​evenly in the range [-1,1]. z ,M x ,M y ,M z}, forming a total of 9317 desired control inputs c i =[F zi ;M xi ;M yi ;M zi ], the value range of i is 1 to 9317;

[0079] The i-th desired control input c i Substitute into the equation c = H (β) n, use the fslove function in MATLAB to solve the remaining unknown parameters, and obtain the actuator control amount {β i ,n 1i ,n 2i ,n 3i}, and the balancing point t is also obtained i , showing the first 5 items of the balanced data set of this embodiment, as shown in Table 2;

[0080] Table 2

[0081]

[0082] S3: Construct a fitting function of the rotor pitch angle with respect to the desired control force and torque, establish a target optimization function based on the trim data set, and use the Levenberg-Marquardt algorithm to optimize the target function to achieve the optimal function fitting effect; in step S3, the rotor pitch angle fitting function is:

[0083]

[0084] In the formula, K is the fitting coefficient to be optimized, K = (k 1 ;k 2 ;…k 11 );

[0085] Based on the balanced data set T, the target optimization function is constructed according to the fitting error:

[0086]

[0087] When the objective function Γ is minimized, the optimal fitting parameters are expressed as:

[0088]

[0089] Where f(K) is the residual vector:

[0090]

[0091] The Levenberg-Marquardt LM optimization algorithm is used to solve the optimal fitting parameters. The LM algorithm has the characteristics of the gradient descent method. Compared with the Gauss-Newton method, this algorithm uses an iterative method to search for the best variable K that minimizes the objective function. * , and the step size can be adjusted in real time to achieve faster convergence. Suppose the fitting coefficient vector at the beginning of the jth iteration is K j , the damping coefficient is μ j ,υ j The specific calculation steps of the iteration are as follows:

[0092] S31: Calculate the j-th iteration residual vector f(K j );

[0093] S32: Calculate the Jacobian matrix J at the jth iteration j for:

[0094]

[0095] In the formula, Represents partial differential.

[0096] S33: Calculate the optimal increment ΔK for the jth iteration j :

[0097]

[0098] S34: Update coefficient vector K j+1 :

[0099] K j+1 =K j +ΔK j

[0100] S35: Calculate the drop rate ρ j :

[0101]

[0102] S36: Update the damping coefficient μ j ,υ j :

[0103]

[0104] S37: Determine the iteration termination condition: If the objective function value of this iteration Γ(K j ) and the previous iteration objective function value Γ(K j-1 ) is less than 10 -6 Then exit the iteration and take K j is the optimal fitting parameter K * .

[0105] Construct the residual vector f(K) according to the trimming working point sequence obtained in step S2, and take the iterative initial fitting coefficient vector K 0 =(0; 0; ... 0), initial fitting coefficient vector increment ΔK 0 =(0; 0; ... 0), initial damping coefficient μ 0 =0,υ 0 = 0 to start iterative optimization, the optimization process objective function Γ decline curve is as follows Figure 3 As shown, it can be seen that the termination condition is met at the third iteration and the optimization is exited.

[0106] K * =(-2.5047×10 -4 ; 0.0473;

[0107] -0.0197; 0.7166; 0.0221; 0.0015; -3.1756×10 -4 ; 0.0208; -0.0078; -6.0686×10 -4 ; -0.0747)

[0108] At this time, the root mean square error is Γ(K * ) / 9317=0.0024, indicating a good fitting effect;

[0109] S4: For any control force and torque within the performance constraints of the UAV, the rotor pitch angle is calculated by fitting the function, the control allocation model is linearized, and then the rotor speed is solved to achieve step-by-step decoupling.

[0110] In step S4, the best fitting parameter is known to be K * , Under the action of a certain desired control input c, the rotor inclination angle β is directly obtained by substituting it into the fitting function:

[0111]

[0112] The rotor inclination angle β is substituted into the control allocation model c=H(β)n for linearization, and the rotor speed n is solved:

[0113] n=H(β) -1 c

[0114] In the formula, H(β)-1 is the generalized inverse matrix of H(β).

[0115] For a tilt-rotor tri-rotor UAV, given the desired control input c = (F z ;M x ;M y ;M z ), first calculate the desired rotor inclination angle β, and then calculate the desired rotation speed of each rotor The step-by-step decoupling of the desired inclination angle calculation and the desired speed calculation is achieved, avoiding the solution of nonlinear equations and thus reducing the amount of calculation. In practical engineering applications, this method can reduce the complexity of the flight control mixer solution.

[0116] For any control input c that satisfies the dynamic constraints, the control allocation includes the following steps:

[0117] Take first:

[0118] β=-2.5047×10 -4 F z +0.0473M x +…-0.0747M y M z

[0119] Then find: n = H (β) -1 c;

[0120] For the control input c = (-13.82; -0.13; 0.65; -0.11), we can calculate β = -0.0290, n = (7.2696 × 10 7 ;7.9625×10 7 ;4.5170×10 7 ), substituting c into the formula c = H (β) n, and using the fsolve function in MATLAB to perform a nonlinear solution, we get β = -0.0259, n = (7.2688 × 10 7 ;7.9617×10 7 ;4.5175×10 7 ); The results obtained by this method are very close to those obtained by nonlinear solution, and the computational complexity of this method is much lower than that of nonlinear solution, so the effectiveness of this method has been preliminarily verified;

[0121] To further verify the effectiveness of this method, the UAV power system model used in this example will be simulated using the cascade decoupling control allocation method of the present invention to verify the decoupling effect. The specific simulation scheme is as follows:

[0122] Assume that the desired control input c is given d (t)=(F zd(t); M xd (t); M yd (t); M zd (t)), where:

[0123]

[0124] According to the desired control input c d (t) The actuator control quantity β(t),n(t) is calculated using this method and loaded into the tilt-rotor tri-rotor dynamic model to obtain the control torque c generated by the dynamic model a (t), and then analyze c a (t) for c d (t) tracking effect. The simulation results are shown in Figure 4 As shown in the figure, the expected thrust F zd The thrust F generated by the power system za Basically consistent; expected rolling moment M xd The rolling moment M generated by the power system xa Basically consistent; expected pitch moment M xd The pitching moment M generated by the power system xa Basically consistent; the yaw moment M generated by the dynamic model za Relative to the expected yaw moment M zd The deviation is within 7.82%; in summary, c a (t) for c d The tracking effect of (t) is good. The direct proportional decoupling control method of the tilt-rotor can effectively reduce the complexity of the control allocation calculation while ensuring the accuracy of the control allocation.

[0125] Therefore, the present invention adopts the above-mentioned direct proportional decoupling control method for a tilt-rotating three-rotor UAV, establishes a target optimization function based on a balancing data set, and optimizes the target function using the Levenberg-Marquardt algorithm to achieve the optimal function fitting effect; under the conditions of a given desired control force and torque, the control allocation model is linearized by calculating the rotor pitch angle, and then the rotor speed is solved, so as to reduce the computational complexity and realize the decoupling of the rotor pitch angle and the rotor speed in the control channel allocation solution. The use of this method in practical engineering applications can effectively improve the decoupling of the control channel and reduce the complexity of the mixer solution of the flight control system.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A direct proportional decoupling control method for a tilt-rotor tri-rotor UAV, characterized in that: The following steps are involved: S1: Establish a mapping function between the desired control force and torque of the tilt-rotor tri-rotor UAV and the rotor inclination angle and rotor speed to build a control allocation model; S2: Select different working points according to the power performance constraints, balance the control distribution model, and obtain the balance data set; S3: Construct a fitting function of the rotor pitch angle with respect to the desired control force and torque, establish a target optimization function based on the trim data set, and use the Levenberg-Marquardt algorithm to optimize the target function to achieve the optimal function fitting effect; S4: For any control force and torque within the performance constraints of the UAV, the rotor pitch angle is calculated by fitting the function, the control allocation model is linearized, and then the rotor speed is solved to achieve step-by-step decoupling.

2. A direct proportional decoupling control method for a tilt-rotor tri-rotor UAV according to claim 1, characterized in that: In step S1, the mapping function between the desired control force and torque of the UAV with respect to the rotor pitch angle and the rotor speed established according to the mechanical characteristics analysis is as follows: The control distribution model is established based on the body coordinate system. b x b y b z b Take the center of gravity of the body as the origin of the coordinate system, o b x b Along the longitudinal axis of the aircraft, pointing to the nose, o b y b Perpendicular to the longitudinal symmetry plane, pointing to the right side of the aircraft, o b z b According to the right-hand rule, (x i ,y i ,z i ) represents the position of the rotation center of rotor i in the body coordinate system, i∈{1,2,3} in the tilt-rotor tri-rotor UAV; In the formula, C L is the rotor thrust coefficient, C M is the rotor torque coefficient, n i is the rotation speed of rotor i; r i is the direction of rotor rotation, the forward propeller value is 1, and the reverse propeller value is -1; The centerline of the nacelle of the rotor i is at o b x b z b Projection in the plane and axis o b x b The angle between the rotor and the i , when the rotor is pointing vertically upward, the tilt angle is 0, and when the rotor is tilted vertically backward, the tilt angle is positive. In the rotor mode, the left and right rotors are differentially controlled, and the tail rotor is kept vertical. The rotor tilt angle satisfies the relationship -α1=α2=β,α3=0, β is the rotor differential angle; F z is the desired thrust input to the controller, M x is the desired rolling moment input to the controller, M y is the desired pitch moment input to the controller, M z is the yaw moment input to the controller. The control allocation model is abbreviated as: c=H(β)n Where c = (F z ;M x ;M y ;M z ) is the desired control input; is the control quantity of the actuator.

3. A direct proportional decoupling control method for a tilt-rotor tri-rotor UAV according to claim 1, characterized in that: In step S2, the trim working point t i Includes the following elements: t i ={F zi ,M xi ,M yi ,M zi ,β i ,n 1i ,n 2i ,n 3i } Different trim working points t i Compose the balanced data set T: T={t1,t2,…,t i ,…t n }。 4. A direct proportional decoupling control method for a tilt-rotor tri-rotor UAV according to claim 1, characterized in that: In step S3, the rotor pitch angle fitting function is: Where K is the fitting coefficient to be optimized, K = (k1; k2; ...k 11 ); Based on the balanced data set T, the target optimization function is constructed according to the fitting error: When the objective function Γ is minimized, the optimal fitting parameters are expressed as: Where f(K) is the residual vector: Use the LM optimization algorithm to calculate the Jacobian matrix J at the jth iteration j for: In the formula, Represents partial differential.

5. The direct proportional decoupling control method for a tilt-rotor tri-rotor UAV according to claim 1, characterized in that: In step S4, the optimal fitting parameter K * , under the action of the desired control input c, the rotor inclination angle β is directly obtained by fitting the function: The rotor inclination angle β is substituted into the control allocation model for linearization to solve the rotor speed n: n=H(β) -1 c In the formula, H(β) -1 is the generalized inverse matrix of H(β).

Images