A torsor-based underactuated unmanned aerial vehicle cascade position and posture control method

Abstract

The application relates to a kind of underactuated unmanned aerial vehicle cascade pose control methods based on torsor, belong to unmanned aerial vehicle control technical field.The method, first, establishes singular-free, constraint-free torsor pose integrated model, avoids dual quaternion normalization calculation burden;Then construct the outer loop position-inner loop attitude cascade structure adapted to the underactuated characteristics of four-rotor, the outer loop uses the Latin hypercube sampling improved gull optimization PD controller to generate the expected attitude instruction of dynamics feasibility, the inner loop adopts the preset performance hyperbolic sliding mode controller to output the unified control force screw containing thrust and torque.The application solves the adaptation problem that existing torsor control relies on full drive assumption, considers modeling uniformity and engineering landing, and can be popularized to various underactuated rigid body systems.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) control technology, specifically relating to a cascade pose control method for an underactuated UAV based on torque. Background Technology

[0002] As a typical underactuated system, quadrotor UAVs exhibit strong coupling between their translational and rotational dynamics. Horizontal movement of the quadrotor must be achieved by tilting the fuselage, a fundamental constraint in its dynamic structure, with position and attitude tightly bound in the equations. Traditional control approaches forcibly separate the coupled dynamics into independent loops. This decoupling design ignores the attitude coupling relationship and is limited in performance when faced with strongly coupled disturbances. To address this, researchers have proposed an integrated attitude control theory, aiming to describe the six-degree-of-freedom motion of a rigid body using a unified mathematical framework, such as dual quaternions. While methods like dual quaternions can achieve integrated attitude modeling, parameter redundancy and normalization constraints increase algorithm complexity and reduce engineering applicability. Furthermore, most existing integrated control methods assume a fully actuated system when designing control laws, which is inconsistent with the actual underactuated characteristics of UAVs and the widely used cascade control architecture in engineering. Simultaneously, the performance improvement strategies for the inner and outer loops in traditional cascade control are based on separate Newton-Euler models, failing to coordinate translation and rotation, thus limiting further performance improvements. Torque, as a simpler mathematical tool based on dual quaternions, provides an effective way to construct singular-free unified dynamic equations. In existing technologies, torque or dual quaternion methods are mostly used for integrated pose modeling and control of UAVs, but their control law design is usually based on the assumption of a fully actuated system, meaning the system is assumed to have independent control capabilities in all six degrees of freedom. However, for quadrotor UAVs, the control inputs are only the total thrust and three torques, making them typical underactuated systems. The position and attitude of a quadrotor are deeply coupled, making translational control independent of attitude impossible. Therefore, integrated pose control based on torque cannot be directly achieved. Consequently, existing torque methods, when applied to quadrotors, often degenerate into a translational and rotational separation control structure, thus losing the coupling information advantage brought by integrated modeling. How to combine torque theory with engineering cascade control architectures to adapt to the underactuated characteristics of UAVs remains a pressing technical challenge. Summary of the Invention

[0003] The purpose of this invention is to provide a torque-based cascade pose control method for underactuated unmanned aerial vehicles (UAVs). This method combines the unified representation capability of torque theory with the engineering practicality of cascade control architecture. It can achieve singularity-free and constraint-free integrated modeling using torque, and adapt to the underactuated characteristics of UAVs through the cascade structure, ultimately achieving high-precision and robust pose tracking control. This invention does not directly use torque for integrated control, but rather, under underactuated constraints, achieves consistent design between modeling and control through torque-based integrated modeling combined with cascade-realizable control methods. This overcomes the technical bottleneck of the difficulty in directly applying torque methods to underactuated quadrotor systems.

[0004] To achieve the above objectives, the technical solution of the present invention is: a cascade pose control method for an underactuated unmanned aerial vehicle based on torque, comprising:

[0005] A torque-based integrated UAV pose model is established, which describes the translation and rotation of the UAV using torque, resulting in a singular and unconstrained six-degree-of-freedom kinematic and dynamic model.

[0006] Based on the integrated pose model of the UAV, a cascade control structure of outer loop position and inner loop attitude is constructed by combining the underactuated characteristics of the quadrotor; the disturbance is uniformly represented as a disturbance force helix, and robust control design is carried out in the torque space.

[0007] In the outer loop position control loop, a PD controller based on the Latin hypercube sampling-seagull optimization algorithm is designed to generate the desired attitude command based on the position command and feedback. In the inner loop attitude control loop, a hyperbolic sliding mode controller with preset performance is designed to track the desired attitude command and generate a control force spiral including thrust and torque to act on the UAV, thereby realizing closed-loop cascade control of attitude.

[0008] Furthermore, the UAV dynamics and kinematics model satisfies the following formula:

[0009]

[0010] Where m is the mass of the drone. Acceleration of the drone, gravity term , It is the acceleration due to gravity. For the control of drones, For external disturbance force, For the rotational inertia of the drone, For the angular velocity of the drone, For the angular acceleration of the drone, For the control torque of the drone, This refers to the external disturbance torque.

[0011] Furthermore, torque Its dual vector satisfies the following definition:

[0012]

[0013] in, The first derivative of the torsion, These are the real and dual parts of the torque, respectively. As a dual unit, for antisymmetric form, E represents the drone's position, and E represents the unit array. For dual quaternions, satisfying:

[0014]

[0015] in, and They are respectively The real part and the dual part, for The derivative of For the dual quaternion form of velocity;

[0016] remember for Conjugate form, dual quaternion Normalization constraints must be met:

[0017]

[0018] Will express Substituting into the above equation, we get:

[0019] .

[0020] Furthermore, defining the body coordinate system B, the inertial coordinate system I, and the desired coordinate system D, the torque-based UAV pose integration model satisfies the following formula:

[0021]

[0022] in, This represents the variable representing the body coordinate system B used in the current description relative to the desired coordinate system D. This represents the variable in the body coordinate system B used in the current description relative to the inertial coordinate system I. For torque, The first derivative of the torsion, for vector form, For the mass operator of the drone, To control the force spiral, For the disturbance force spiral, It is a gravity-driven spiral.

[0023] Furthermore, the cascade control structure is specifically implemented within the torque framework as follows:

[0024] The outer-loop position controller processes position commands and feedback information in torque space. First, it generates the desired control force helix. Then, it introduces thrust direction constraints and projects the desired control force helix into a subspace to satisfy the physical constraint that the UAV thrust is applied only along the fixed axis of the fuselage. Based on the projected force direction, it constructs the desired attitude command and total thrust.

[0025] The inner-loop attitude controller tracks the desired attitude command in the torque space and synthesizes a unified control force spiral that includes both thrust and torque, which directly acts on the UAV attitude integrated model.

[0026] Furthermore, the design process of the PD controller based on the Latin hypercube sampling-seagull optimization algorithm, namely the DSPD controller, includes:

[0027] Latin hypercube sampling was used to initialize the initial population of the seagull optimization algorithm, with each individual corresponding to a set of PD control parameters;

[0028] The fitness function H is the absolute error of the integration time.

[0029]

[0030] in, For time, For pose error, This is the maximum score.

[0031] The population is iteratively updated by simulating the migration and spiral attack behaviors of seagulls to find the PD parameter combination that minimizes the absolute error of the integration time.

[0032] Based on the obtained PD parameters, a position control law is generated. Substituting this law into the integrated UAV pose model, the acceleration of the UAV's center of mass relative to the desired coordinate system is obtained as follows:

[0033]

[0034] in, This indicates the position of the UAV's center of mass relative to the desired coordinate system. This represents the linear velocity of the UAV relative to the desired coordinate system from the body coordinate system.

[0035] Since the UAV's thrust can only be generated along the z-axis of the aircraft, this mapping process must satisfy the underactuated constraint. From Newton's equations, the required control force in the inertial frame is:

[0036]

[0037] in, Let be the unit vector of the direction of gravity in the inertial frame. It is the acceleration due to gravity;

[0038] The desired z-axis direction of the machine body is determined by the desired control force vector direction:

[0039]

[0040] The desired z-axis direction of the machine body;

[0041] Combined with the given desired yaw angle Calculate the desired x-axis direction of the organism:

[0042]

[0043] in, In the inertial coordinate system, the desired yaw angle is... Defined horizontal reference direction vector;

[0044] This leads to the desired y-axis direction of the organism:

[0045]

[0046] Therefore, the desired rotation matrix is:

[0047]

[0048] in, It is a three-dimensional special orthogonal group. Expectation matrix. This indicates that the three attitude vectors in the matrix are pairwise perpendicular, each with a length of 1, and strictly follow the "right-hand rule";

[0049] Therefore, the desired attitude command can be obtained. and total thrust :

[0050]

[0051] in, This represents the corrected Rodrigues parameter of the body coordinate system relative to the desired coordinate system.

[0052] Furthermore, the design method of the preset performance hyperbolic sliding mode controller is as follows:

[0053] Design sliding surface With the law of convergence for:

[0054]

[0055] in, Let be the angular velocity of the body coordinate system B relative to the desired coordinate system D. , These are the sliding mode control parameters. This represents the corrected Rodrigues parameter of the body coordinate system relative to the desired coordinate system;

[0056] The corresponding control torque satisfies:

[0057]

[0058] in, This represents the control torque of the body coordinate system B relative to the desired coordinate system D. This represents the disturbance moment of the body coordinate system B relative to the desired coordinate system D. The moment of inertia of the drone.

[0059] Furthermore, the preset performance hyperbolic sliding mode controller also includes a preset performance function. It satisfies the following formula:

[0060]

[0061] in, It is the overshoot suppression parameter. For sliding mode 3D unfolding, These are the initial constraint parameters. These are the final constraint parameters, which satisfy... , These are the convergence parameters.

[0062] Furthermore, the preset performance hyperbolic sliding mode controller also transforms the attitude error based on a preset performance function:

[0063] definition Then there is The derivatives of the unconstrained transformation variables satisfy the following:

[0064]

[0065] in, For sliding surface The derivative of It is a real number;

[0066] The adjusted control torque satisfies the following formula:

[0067]

[0068] in, E is the identity matrix.

[0069] The present invention also provides an electronic device, including a memory, a processor, and computer program instructions stored in the memory and executable by the processor, wherein when the processor executes the computer program instructions, it can implement the method described above.

[0070] Compared with existing technologies, this invention has the following advantages: The method of this invention, for the first time, applies torque theory to the cascade control architecture of UAVs, solving the problem that existing torque control methods are difficult to apply to practical underactuated systems due to the use of the full-drive assumption. This method retains the advantages of unified representation, singularity-free, and unconstrained torque modeling, while inheriting the strong engineering practicality and ease of deployment of cascade control structures. Under a unified torque framework, high-performance algorithms are designed for both the inner and outer loops. The outer loop's DSPD controller achieves adaptive online parameter tuning through a Seagull optimization algorithm improved by Latin hypercube sampling, improving adaptability to dynamic environments and disturbances. The inner loop's PHSMC strictly constrains the transient and steady-state performance of the error through a preset performance function and effectively suppresses chattering through the hyperbolic tangent reaching law, significantly improving the system's robustness and tracking accuracy. The method of this invention provides a clear and effective implementation path for the engineering application of torque theory in underactuated systems, applicable not only to quadcopter UAVs but also applicable to other types of underactuated rigid body systems. Attached Figure Description

[0071] Figure 1 This is a flowchart of the DSPD algorithm in an embodiment of the present invention.

[0072] Figure 2 This is the UAV position response curve under wind field disturbance in an embodiment of the present invention.

[0073] Figure 3 This is the attitude response curve of the UAV under wind field disturbance in an embodiment of the present invention.

[0074] Figure 4 This is the output response diagram under PHSMC control in an embodiment of the present invention.

[0075] Figure 5 This is a schematic diagram of a torque-based cascade control architecture in an embodiment of the present invention. Detailed Implementation

[0076] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings.

[0077] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0078] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0079] This invention provides a torque-based underactuated unmanned aerial vehicle (UAV) cascade pose control method, comprising:

[0080] A torque-based integrated UAV pose model is established, which describes the translation and rotation of the UAV using torque, resulting in a singular and unconstrained six-degree-of-freedom kinematic and dynamic model.

[0081] Based on the integrated pose model of the UAV, instead of adopting integrated control under the full-drive assumption, a cascade control structure of outer loop position and inner loop attitude is constructed by combining the underactuated characteristics of the quadrotor to realize the physical realizability of the control input; the disturbance is uniformly represented as a disturbance force helix, and robust control design is carried out in the torque space, so as to retain the coupling relationship between the disturbance in translation and rotation during the compensation process and improve the disturbance rejection performance.

[0082] In the outer loop position control loop, a PD controller based on the Latin hypercube sampling-seagull optimization algorithm is designed to generate the desired attitude command based on the position command and feedback, so that the generated desired attitude is consistent with the UAV dynamics, thereby avoiding the unreachable attitude command problem that may occur in traditional separate control. In the inner loop attitude control loop, a hyperbolic sliding mode controller with preset performance is designed to track the desired attitude command and generate a control force spiral including thrust and torque to act on the UAV, realizing closed-loop cascade control of attitude.

[0083] The following is a further explanation of the relevant content involved in the method of the present invention.

[0084] 1. Torque-based integrated pose modeling of unmanned aerial vehicles

[0085] The Newton-Euler equations of dynamics for drones satisfy:

[0086]

[0087] Where m is the mass of the drone. Accelerate the drone It is the acceleration due to gravity. For the control of drones, For external disturbance force, For the rotational inertia of the drone, For the angular velocity of the drone, For the angular acceleration of the drone, For the control torque of the drone, This refers to the external disturbance torque.

[0088] Define the body coordinate system (B), the inertial coordinate system (U), and the desired coordinate system (D). Introduce dual quaternions. To describe the relative pose between coordinate systems, which satisfies:

[0089]

[0090] in, It is a quaternion. and They are respectively The real part and the dual part, for The derivative of Dual velocity vector The dual quaternion form.

[0091] To overcome the normalization constraint problem of dual quaternions, a torsion is introduced. and its dual vector :

[0092]

[0093] in, These are the real and dual parts of the torque, respectively. As a dual unit, for antisymmetric form, E represents the location of the drone and E represents the unit array. It is a quaternion. It is a dual quaternion that satisfies the normalization constraint:

[0094]

[0095] remember for Conjugate form, express Substituting into the formula, we get:

[0096] .

[0097] It is evident that its torque form can automatically satisfy this constraint, thereby avoiding additional normalization calculations and simplifying the model complexity.

[0098] To achieve trajectory tracking, a relative dynamic model based on torque is derived by combining the Newton-Euler dynamics equations of the UAV:

[0099]

[0100] Among them, parameters This represents the variable in the current body coordinate system B, relative to the desired coordinate system D. (Parameter) This represents the variable in the body coordinate system B used in the current description, relative to the inertial coordinate system I. For the mass operator of the drone, To control the force spiral. For the disturbance force spiral, The gravity-driven spiral is specifically represented as:

[0101]

[0102] 2. Construction of Cascade Control Structure under Torque Framework

[0103] Quadrotor drones are a typical underactuated system, with four control inputs but controlling six degrees of freedom. Due to their physical limitations, quadrotors can only generate translational forces in different directions by adjusting their attitude. Therefore, the traditional control approach involves calculating the pose and then performing cascade control. However, the position controller only considers the position error and not the real-time attitude of the drone, which may result in the position controller generating an unattainable desired attitude command.

[0104] Torque can project the desired force helically onto the UAV's realizable subspace, thus ensuring that the generated desired attitude naturally satisfies reachability constraints. However, torque has previously been mostly used in fully actuated systems. To apply torque theory to quadcopter UAVs, their underactuated characteristics must be considered, and cascaded control structures cannot be avoided.

[0105] In this invention, a dynamic model of a quadrotor UAV is constructed within a torque system. Considering the underactuated characteristics of the quadrotor, a cascade control approach is used to achieve attitude coupling control of the quadrotor UAV within the torque system. The desired attitude is not generated based on the traditional "acceleration-attitude inverse kinematics" method, but rather constructed based on torque dynamic consistency constraints. The advantage of this approach is that the desired force helix is ​​projected onto the UAV's realizable subspace, ensuring that the generated attitude naturally satisfies the underactuated reachability constraints. That is, position control considers attitude information, preventing unattainable desired attitude commands. Furthermore, external disturbances are no longer decomposed into independent force and torque disturbances, but are uniformly represented as a disturbance force helix. The system disturbances are considered in a coupled manner, which is something that traditional discrete disturbance rejection cannot do.

[0106] In the cascade control of this invention, the outer-loop position controller processes position commands and feedback in the torque space, converting them into desired attitude commands. The inner-loop attitude controller also tracks the desired attitude commands in the torque space, ultimately synthesizing a unified control force spiral that includes both thrust and torque, directly acting on the quadrotor dynamics model.

[0107] 3. Design of outer loop DSPD position controller

[0108] The outer loop employs a PD controller based on the Seagull Optimization Algorithm using Latin hypercube sampling, namely DSPD. The algorithm flow is as follows: Figure 1 As shown.

[0109] To avoid collisions between individual "seagulls", an additional variable is introduced. The formula for adjusting the direction of motion is shown below:

[0110]

[0111] in, A random number within the interval [0,1]; This variable, which decays linearly with the number of iterations, is used to balance the algorithm's exploration and development capabilities. Its calculation formula is as follows:

[0112]

[0113] in, This represents the maximum number of iterations.

[0114] Towards the current optimal position The formula for calculating the direction of motion is as follows:

[0115]

[0116] in, This represents the current position of the individual. This is the distance vector between the individual and the optimal position; The random coefficient for adjusting exploration behavior is calculated using the following formula:

[0117]

[0118] in, It is a random number between [0,1].

[0119] Combining collision avoidance behavior with movement towards the optimal position, the position update of individual seagulls during migration is given:

[0120]

[0121] When an individual seagull approaches its prey, it employs a spiral descent attack strategy to perform a fine-grained local search. The position update formula is shown below:

[0122]

[0123] Where b is a constant that defines the shape of the spiral. It is a random number in the range [0,1].

[0124] The algorithm selects either migration or attack behavior with equal probability to balance exploration and exploitation in each iteration. The selection mechanism is as follows:

[0125]

[0126] in, It is a random number in the interval [0,1], used to control behavior selection.

[0127] when At this point, the individual no longer rotates around its current optimal position but instead performs random exploration, thereby enhancing the algorithm's ability to escape local optima. In this case, an individual is randomly selected from the reference population for position updates. The position is calculated using the following formula:

[0128]

[0129] To evaluate the performance of the PD controller parameters, ITAE is used as the fitness function, and its expression is:

[0130]

[0131] in, It is a pose error.

[0132] Based on the relative dynamics model of the UAV using torque, the traditional PD control law is given as follows:

[0133]

[0134] in, This is a gravity coupling term. and For the dual operators of the controller, they satisfy:

[0135]

[0136] in, For the quality of drones, and It is a positive real number.

[0137] Therefore, the quadcopter unmanned aerial vehicle system satisfies:

[0138]

[0139] The closed-loop position motion equations of the system are shown below:

[0140]

[0141] Since the UAV's thrust can only be generated along the z-axis of the fuselage, this mapping process must satisfy underactuated constraints. From Newton's equations, the required control force (in the inertial frame) is:

[0142]

[0143] in, It is the unit vector of the direction of gravity in the inertial frame.

[0144] The desired z-axis direction of the machine body is determined by the desired control force vector direction:

[0145]

[0146] Combined with the given desired yaw angle The desired x-axis direction of the organism can be calculated:

[0147]

[0148] This leads to the desired y-axis of the organism: Therefore, the expected rotation matrix is...

[0149]

[0150] Receive the desired attitude command :

[0151]

[0152] Simultaneously, by converting virtual control force into drone lift, the total thrust is obtained:

[0153]

[0154] To enhance the algorithm's global exploration capability in the parameter space, Latin hypercube sampling is used to initialize the initial population positions in the Seagull Optimization Algorithm. Subsequently, the optimized SOA is integrated into a torque-based PD controller to obtain the proposed DSPD controller.

[0155] 4. Inner Loop PHSMC Attitude Controller Design

[0156] The inner loop employs a pre-defined performance hyperbolic sliding mode controller (PHSMC) to ensure transient and steady-state performance in attitude tracking. The sliding surface is designed. With the law of convergence for

[0157]

[0158] in, , These are the sliding mode control parameters. The control torque at this time is:

[0159]

[0160] Build a pre-defined performance function Used to constrain the transient and steady-state performance of attitude tracking errors:

[0161]

[0162] in, , . These are the initial constraint parameters. These are the final constraint parameters, which satisfy... , These are the convergence parameters.

[0163] The attitude error is transformed based on the performance function, resulting in the following unconstrained transformation variables:

[0164]

[0165] in, .

[0166] Combining the sliding surface and the reaching law, the control torque under the preset performance function is obtained, specifically:

[0167]

[0168] in, For real numbers, , .

[0169] 5. Simulation Analysis

[0170] To verify the effectiveness and superiority of the torque-based underactuated UAV cascade pose control method (i.e., torque-based DSPD-PHSMC) proposed in this invention, numerical simulation experiments were conducted in this embodiment. In the simulation, the proposed method was compared with two comparative methods: the traditional PD-SMC method and the torque-based cascade PD-SMC method. All methods used the same UAV physical and control parameter benchmarks to ensure a fair comparison.

[0171] The UAV was designed to complete a complex trajectory tracking task comprising six stages within 35 seconds, with the desired yaw angle set to 0 throughout the flight. To test the robustness of the algorithm, continuous wind disturbance was applied throughout the simulation. The simulation results are as follows: Figure 2-4 As shown.

[0172] Figure 2 The position response curves of three controllers in trajectory tracking tasks are shown. As can be seen from the figures, both torque-based controllers can track the desired trajectory well, and their tracking accuracy is superior to the traditional PD-SMC. Among them, the torque-based DSPD-PHSMC method proposed in this invention exhibits the smallest position fluctuation and the highest steady-state accuracy during the hovering phase. This is attributed to the outer-loop DSPD controller using online parameter tuning via the Seagull optimization algorithm, which enhances the system's adaptability to environmental changes.

[0173] Figure 3 Attitude tracking error curves for three controllers are presented. The traditional PD-SMC method exhibits significant chattering in the roll and pitch channels with large error amplitudes. While the torque-based PD-SMC method suppresses chattering to some extent, the error still exceeds the preset limits. In contrast, the torque-based DSPD-PHSMC method proposed in this invention can strictly constrain roll and pitch angle errors within preset performance boundaries, effectively suppressing chattering and achieving smooth, high-precision attitude tracking. Regarding yaw angle control, all three methods achieve high steady-state accuracy, but the traditional method exhibits slight fluctuations, while the two torque-based methods further reduce the error, verifying the advantages of torque modeling.

[0174] Figure 4 For input control comparison, all three methods produce relatively smooth outputs in terms of total thrust. However, in terms of control torque, the traditional PD-SMC exhibits significant high-frequency chattering, which reflects the inherent switching characteristics of sliding mode control. In contrast, the two torque methods, due to their unified geometric description of translation and rotation, can more naturally handle system coupling, resulting in significantly smoother torque output and a substantial reduction in chattering amplitude. The method of this invention further incorporates a preset performance hyperbolic sliding mode, achieving the smoothest torque output.

[0175] The control system flowchart of the torque-based underactuated UAV cascade pose control method of the present invention is as follows: Figure 5 As shown.

[0176] When the control system performs pose and motion planning, it first clarifies the target position, attitude, and motion path of the UAV through target modeling and desired trajectory setting. Based on this, considering external disturbances in the actual flight environment such as wind interference, a relative dynamics model of the UAV based on torque theory is established to accurately describe the UAV's pose information. Subsequently, the PD controller parameters are optimized using a Seagull optimization algorithm based on Latin hypercube sampling. The control strategy is adjusted in real time through position loop performance evaluation to ensure accurate tracking of the desired trajectory by the UAV. Finally, preset performance constraints are introduced at the attitude control layer, a hyperbolic sliding mode surface is designed, and a corresponding sliding mode control law is constructed, forming an integrated cascade control architecture. This achieves high-precision and robust control of the underactuated UAV's pose, ensuring stable operation of the system in complex disturbance environments.

[0177] The present invention also provides an electronic device, including a memory, a processor, and computer program instructions stored in the memory and executable by the processor, wherein when the processor executes the computer program instructions, it can implement the method described above.

[0178] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0179] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0180] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0181] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0182] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A cascade pose control method for an underactuated unmanned aerial vehicle based on torque, characterized in that, include: A torque-based integrated UAV pose model is established, which describes the translation and rotation of the UAV using torque, resulting in a singular and unconstrained six-degree-of-freedom kinematic and dynamic model. Based on the integrated pose model of the UAV, a cascade control structure of outer loop position and inner loop attitude is constructed by combining the underactuated characteristics of the quadrotor; the disturbance is uniformly represented as a disturbance force helix, and robust control design is carried out in the torque space. In the outer loop position control loop, a PD controller based on the Latin hypercube sampling-seagull optimization algorithm is designed to generate the desired attitude command based on the position command and feedback. In the inner loop attitude control loop, a hyperbolic sliding mode controller with preset performance is designed to track the desired attitude command and generate a control force spiral including thrust and torque to act on the UAV, thereby realizing closed-loop cascade control of attitude.

2. The underactuated unmanned aerial vehicle cascade pose control method based on torque according to claim 1, characterized in that, The dynamic and kinematic model satisfies the following formula: Where m is the mass of the drone. Acceleration of the drone, gravity term , It is the acceleration due to gravity. For the control of drones, For external disturbance force, For the rotational inertia of the drone, For the angular velocity of the drone, For the angular acceleration of the drone, For the control torque of the drone, This refers to the external disturbance torque.

3. The underactuated unmanned aerial vehicle cascade pose control method based on torque according to claim 1, characterized in that, Torque and its dual vector Satisfies the following definition: in, These are the real and dual parts of the torque, respectively. As a dual unit, for antisymmetric form, E represents the drone's position, and E represents the unit array. For dual quaternions, satisfying: in, and They are respectively The real part and the dual part, for The derivative, For the dual quaternion form of velocity; remember for Conjugate form, dual quaternion Normalization constraints must be met: Will express Substituting into the above equation, we get: 。 4. The underactuated unmanned aerial vehicle cascade pose control method based on torque according to claim 1, characterized in that, Define the body coordinate system B, the inertial coordinate system I, and the desired coordinate system D. The torque-based UAV pose integration model satisfies the following formula: in, This represents the variable representing the body coordinate system B used in the current description relative to the desired coordinate system D. This represents the variable representing the body coordinate system B used in the current description relative to the inertial coordinate system I. For torque, The first derivative of the torsion, for vector form, For the mass operator of the drone, To control the force spiral, For the disturbance force spiral, It is a gravity-driven spiral.

5. The underactuated unmanned aerial vehicle cascade pose control method based on torque according to claim 1, characterized in that, The cascaded control structure is specifically implemented under the torque framework as follows: The outer-loop position controller processes position commands and feedback information in torque space. First, it generates the desired control force helix. Then, it introduces thrust direction constraints and projects the desired control force helix into a subspace to satisfy the physical constraint that the UAV thrust is applied only along the fixed axis of the fuselage. Based on the projected force direction, it constructs the desired attitude command and total thrust. The inner-loop attitude controller tracks the desired attitude command in the torque space and synthesizes a unified control force spiral that includes both thrust and torque, which directly acts on the UAV attitude integrated model.

6. The underactuated unmanned aerial vehicle cascade pose control method based on torque according to claim 1, characterized in that, The design process of the PD controller based on the Latin hypercube sampling-seagull optimization algorithm, namely the DSPD controller, includes: Latin hypercube sampling was used to initialize the initial population of the seagull optimization algorithm, with each individual corresponding to a set of PD control parameters; The fitness function H is the absolute error of the integration time. in, For time, For pose error, This is the maximum score. The population is iteratively updated by simulating the migration and spiral attack behaviors of seagulls to find the PD parameter combination that minimizes the absolute error of the integration time. Based on the obtained PD parameters, a position control law is generated and substituted into the UAV attitude integrated model to obtain the UAV's center of mass acceleration relative to the desired coordinate system: in, This indicates the position of the UAV's center of mass relative to the desired coordinate system. This represents the linear velocity of the UAV relative to the desired coordinate system from the body coordinate system. The required control force in an inertial frame, derived from Newton's equations, is: in, Let be the unit vector of the direction of gravity in the inertial frame. It is the acceleration due to gravity; The desired z-axis direction of the machine body is determined by the desired control force vector direction: The desired z-axis direction of the machine body; Combined with the given desired yaw angle Calculate the desired x-axis direction of the organism: in, In the inertial coordinate system, the desired yaw angle is... Defined horizontal reference direction vector; This leads to the desired y-axis direction of the organism: Therefore, the desired rotation matrix is: in, It is a three-dimensional special orthogonal group; Therefore, the desired attitude command can be obtained. and total thrust : in, This represents the corrected Rodrigues parameter of the body coordinate system relative to the desired coordinate system.

7. The underactuated unmanned aerial vehicle cascade pose control method based on torque according to claim 1, characterized in that, The design method of the preset performance hyperbolic sliding mode controller is as follows: Design sliding surface With the law of convergence for: in, Let be the angular velocity of the body coordinate system B relative to the desired coordinate system D. , These are the sliding mode control parameters. This represents the corrected Rodrigues parameter of the body coordinate system relative to the desired coordinate system; The corresponding control torque satisfies: in, This represents the control torque of the body coordinate system B relative to the desired coordinate system D. This represents the disturbance moment of the body coordinate system B relative to the desired coordinate system D. The moment of inertia of the drone.

8. The underactuated unmanned aerial vehicle cascade pose control method based on torque according to claim 7, characterized in that, The preset performance hyperbolic sliding mode controller also includes a preset performance function. It satisfies the following formula: in, It is the overshoot suppression parameter. For sliding mode 3D unfolding, These are the initial constraint parameters. These are the final constraint parameters, which satisfy... , These are the convergence parameters.

9. The underactuated unmanned aerial vehicle cascade pose control method based on torque according to claim 8, characterized in that, The preset performance hyperbolic sliding mode controller also transforms the attitude error based on a preset performance function: definition Then there is ; The derivatives of the unconstrained transformation variables satisfy: in, For sliding surface The derivative, It is a real number; The adjusted control torque satisfies the following formula: in, E is the identity matrix.

10. An electronic device, characterized in that, It includes a memory, a processor, and computer program instructions stored in the memory and executable by the processor, wherein when the processor executes the computer program instructions, it can implement the method as described in any one of claims 1-9.

Images