A model predictive control method and device for a rotor unmanned aerial vehicle based on a non-inertial system

Abstract

The application discloses a kind of based on non-inertial system's rotor unmanned aerial vehicle model predictive control method and device, the method includes: according to the relationship between inertial system, non-inertial system and rotor unmanned aerial vehicle system, the dynamic model of rotor unmanned aerial vehicle under non-inertial system is established;According to the dynamic model, the nonlinear model predictive controller of rotor unmanned aerial vehicle is designed;Using sequential quadratic programming algorithm, the relative pose given by motion capture system or CREPES is combined with the state information collected by inertial measurement unit on non-inertial system to solve the optimization problem in the nonlinear model predictive controller, the total thrust of rotor unmanned aerial vehicle and three-axis angular velocity control quantity are obtained, so as to realize the flight control of rotor unmanned aerial vehicle under non-inertial system.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) control, and particularly relates to a predictive control method and device for a rotorcraft UAV model based on a non-inertial frame of reference. Background Technology

[0002] In recent years, rotary-wing drones have been widely used in aerial photography, monitoring, exploration, and rescue due to their advantages of small size, light weight, high maneuverability, and flexibility. This has rapidly driven the development of multi-robot systems, particularly in collaborative interaction tasks between rotary-wing drones and moving targets, such as point-to-point following, autonomous landing, and trajectory tracking. These interactive tasks often require accurate state estimation information of the rotary-wing drone relative to the target as feedback to control its motion in the inertial reference frame. The controller typically requires a complete system model to accurately control the rotary-wing drone's flight. Good trajectory planning and continuous replanning are crucial for the stable and efficient completion of these complex tasks. Common airborne relative state estimation methods often employ computer vision technology or are provided directly by external devices (motion capture systems, ultra-wideband positioning systems). However, both methods inevitably involve trajectory replanning, which significantly increases the computational load and complexity of the system. Furthermore, since the interactive target is usually in an unpredictable random movement process, it is difficult for the rotary-wing drone to obtain an accurate motion model of the target as prior knowledge. In summary, traditional rotorcraft UAV control in an inertial reference frame faces the following challenges: the need to accurately estimate the relative state between the rotorcraft UAV and the target; the inability to avoid frequent and complex trajectory replanning operations; and the difficulty in obtaining a target motion model for prediction.

[0003] Considering the different characteristics and challenges of rotary-wing UAVs completing interactive tasks such as point following, autonomous landing, and trajectory tracking during target movement, a large number of studies have been conducted on these tasks in recent years. Existing technology [1] uses a similar group control algorithm based on complex Laplace to achieve point following on a Leader-Follower network that actively estimates the relative position. Existing technology [2] proposes a vision-based autonomous landing method for a collaborative system between rotary-wing UAVs and unmanned vehicles. It uses multiple QR codes on the unmanned vehicle platform to estimate the relative distance, speed, and direction between the two. Based on this, it designs a speed controller based on the control barrier function (CBF) and control Lyapunov function (CLF) for the landing task of quadcopter UAVs on a mobile platform. Existing technology [3] also proposes an autonomous landing method. It uses extended Kalman filtering and visual inertial odometry for pose estimation and obtains the state estimation information of the rotary-wing UAV relative to the unmanned vehicle through a similar label recognition method. This technology also integrates a local trajectory planning algorithm into the system to ensure safer task completion. The prior art [4] proposed a solution for safe trajectory tracking of moving targets by rotary-wing UAVs. By predicting the target motion and combining it with a heuristic trajectory planning method, the purpose of planning a safe trajectory and tracking the rotary-wing UAV in a dense environment can be achieved. However, the above methods are all completed in the inertial frame to control and estimate the state of the rotary-wing UAV. The target motion model is usually only made simple assumptions (uniform speed / uniform acceleration), and lacks further dynamic information of the target (angular velocity and linear acceleration), which makes it difficult for the proposed system to control the rotary-wing UAV when the target is moving violently.

[0004] When it comes to specific control methods for rotary-wing UAVs, model predictive control has undoubtedly been a hot topic in recent years. Existing technology [5] adds perception terms and action terms to the optimization problem and designs a nonlinear model predictive controller suitable for quadcopter UAVs. In the experiment, the visual inertial odometry and the controller were running in real time on the onboard computer. The quadcopter UAV needed to keep the target point of interest in its field of vision while tracking a trajectory. However, this method also completes the control of the rotary-wing UAV in the inertial frame. For tasks such as fixed-point following of moving targets, autonomous landing and trajectory tracking, it is still impossible to avoid frequent and complex trajectory replanning operations, which not only increases the computational burden of the onboard computer, but also reduces the reliability of the system to a certain extent.

[0005] In the existing literature, there is relatively little research on the modeling and control of rotary-wing UAVs in non-inertial frames. Such technology is more popular in the field of close-range operation of spacecraft. Existing technology [6] studies the dynamic model of quadcopter UAVs in non-inertial frames where no rotational transformation occurs. That is, the referenced non-inertial frame only undergoes translational transformation, and a sliding mode controller is used for trajectory tracking. Although this technology provides a new idea and direction for the control of rotary-wing UAVs in non-inertial frames, the application scope of the model is relatively limited because the dynamic model lacks the rotational transformation of the non-inertial frame.

[0006] References:

[0007] [1] Z.Han, K.Guo, L.Xie, and Z.Lin, “Integrated relative localization and leader–follower formation control,” IEEE Transactions on Automatic Control, vol.64, no.1, pp.20–34, 2018.

[0008] [2] G.Niu, Q.Yang, Y.Gao, and M.-O.Pun, "Vision-Based Autonomous Landing for Unmanned Aerial and Ground Vehicles Cooperative Systems," IEEE Roboticsand Automation Letters, vol.7, no.3, pp.6234–6241, Jul.2022, doi:10.1109 / LRA.2021.3101882.

[0009] [3] P.Wang, C.Wang, J.Wang, and MQ-H.Meng, "Quadrotor Autonomous Landingon Moving Platform." arXiv, Aug.10, 2022.doi:10.48550 / arXiv.2208.05201.

[0010] [4] Z.Han, R.Zhang, N.Pan, C.Xu, and F.Gao, "Fast-tracker: Arobust aerialsystem for tracking agile target in cluttered environments," in IEEEInternational Conference on Robotics and Automation (ICRA). IEEE, 2021, pp.328–334.

[0011] [5] D. Falanga, P. Foehn, P. Lu, and D. Scaramuzza, "PAMPC: PerceptionAwareModel Predictive Control for Quadrotors," in 2018 IEEE / RSJ International Conference on Intelligent Robots and Systems (IROS), Oct.2018, pp.1–8.doi:10.1109 / IROS.2018.8593739.

[0012] [6]Y.Marani, K.Telegenov, E.Feron, and M.-TLKirati, "Drone referencetracking in a non-inertial frame: control, design and experiment," in 2022IEEE / AIAA 41st Digital Avionics Systems Conference(DASC),Portsmouth, VA, USA, Sep.2022, pp.1–8.doi:10.1109 / DASC55683.2022.9925849. Summary of the Invention

[0013] To address the shortcomings of existing technologies, this invention proposes a model predictive control method and device for rotary-wing unmanned aerial vehicles (UAVs) based on a non-inertial frame of reference. This invention utilizes the attitude and velocity of the UAV relative to a reference target, as well as the angular velocity and acceleration information of the reference target in an inertial reference frame, to achieve stable flight of the UAV in a non-inertial frame of reference without relying on additional inertial positioning equipment and state estimation modules (such as GPS or SLAM), prior information on the target motion model, or complex trajectory replanning operations.

[0014] According to a first aspect of the embodiments of this application, a predictive control method for a rotorcraft unmanned aerial vehicle (UAV) model based on a non-inertial frame of reference is provided, comprising:

[0015] (1) Based on the relationship between the inertial frame, the non-inertial frame and the rotorcraft system, establish a dynamic model of the rotorcraft UAV in the non-inertial frame;

[0016] (2) Based on the dynamic model, design a nonlinear model predictive controller for the rotary-wing UAV;

[0017] (3) Using the sequential quadratic programming algorithm, combined with the relative pose given by the motion capture system or CREPES and the state information collected by the inertial measurement unit in the non-inertial frame, the optimization problem in the nonlinear model predictive controller is solved to obtain the total thrust and three-axis angular velocity control quantities of the rotor UAV, thereby realizing the flight control of the rotor UAV in the non-inertial frame.

[0018] Further, step (1) includes:

[0019] Based on the relationship between the inertial frame, the non-inertial frame, and the rotorcraft UAV system, the relative position of the rotorcraft UAV in the non-inertial frame can be obtained. in N R W Let represent the rotation matrix of the inertial frame W relative to the non-inertial frame N. Let N represent the relative displacement vector between the non-inertial frame N and the rotorcraft UAV system B in the inertial frame W.

[0020] By analyzing the above formula By taking the derivative, we can obtain the relative velocity of the rotary-wing UAV in the non-inertial frame. N v B With relative acceleration Represented as:

[0021]

[0022]

[0023] in,[ N Ω N ] × and[ N β N ] × Let N and N represent the angular velocities of the non-inertial frame of reference, respectively. N Ω N With angular acceleration N β N antisymmetric matrix, N R BLet represent the rotation matrix of the rotorcraft UAV system B relative to the non-inertial frame N, and let T represent the normalized total thrust of the rotorcraft UAV, where g = [0, 0, -g]. T Represents gravitational acceleration. N R W a N Let N be its linear acceleration in a non-inertial frame of reference.

[0024] By analyzing the relative velocity and relative acceleration... N R B By taking the derivative and converting it to quaternion form, we obtain the dynamic model of the rotating part:

[0025]

[0026]

[0027] in, B Ω B and[ B Ω B ] × Let represent the angular velocity and antisymmetry matrix of rotorcraft UAV system B, respectively. N q B Representing the rotation matrix N R B The quaternion form of ⊙ represents the Hamiltonian product of four arithmetic multiplications.

[0028] Furthermore, the nonlinear model predicts the state variables of the controller. Input in N p B The relative position of the rotary-wing UAV in a non-inertial frame of reference. N v B Let be the relative velocity of the rotary-wing UAV in a non-inertial frame of reference. N q B Rotation matrix N R B The quaternion form represents the relative attitude of a rotary-wing UAV in a non-inertial frame. and These are the angular velocity and linear acceleration values ​​of the non-inertial frame N, obtained directly from the IMU. N β N Let N be the angular acceleration in the non-inertial frame, and T represent the normalized total thrust of the rotary-wing UAV. These are the three-axis angular velocities of rotorcraft system B.

[0029] Furthermore, the nonlinear model predictive controller is:

[0030]

[0031] stx(0)=x0

[0032] x(k+1)=f d (x(k),u(k))

[0033] T min ≤T≤T max

[0034]

[0035]

[0036]

[0037] Where, ‖x‖ Q =x T Qx, x(k) and u(k) represent the state and input variables at time k, respectively. ref With x(n) ref U represents the reference state quantity at time k. h Represents the reference input quantity (u) h = [g; 0; 0; 0], meaning the input value when the rotorcraft is hovering is always used as the reference input value), Q = diag(Q x ), Q x With Q n Let f be the positive definite weight matrix for each system state variable, and R be the positive definite weight matrix for the system input variables; for the constraint part, f d (x(k),u(k)) is the discretized representation of the differential equation form f(x,u) of the dynamic model, T min With T max Ω represents the minimum and maximum thrust limits, respectively. rp With Ω yaw These represent the angular velocity limits in the roll, pitch, and yaw directions, respectively.

[0038] Further, in step (3), based on the relative pose given by the motion capture system or CREPES and the non-inertial frame angular velocity and linear acceleration directly obtained from the inertial measurement unit, and setting the estimated and reference values ​​of its angular acceleration to 0, that is, assuming that the non-inertial frame is always moving on the plane, the controller is solved.

[0039] According to a second aspect of the embodiments of this application, a predictive control device for a rotorcraft unmanned aerial vehicle model based on a non-inertial frame of reference is provided, comprising:

[0040] The modeling module is used to establish a dynamic model of a rotary-wing UAV in a non-inertial frame based on the relationship between the inertial frame, the non-inertial frame, and the rotary-wing UAV system.

[0041] The design module is used to design a nonlinear model predictive controller for the rotary-wing UAV based on the dynamic model.

[0042] The solution module is used to solve the optimization problem in the nonlinear model predictive controller by using a sequential quadratic programming algorithm, combined with the relative pose given by the motion capture system or CREPES and the state information collected by the inertial measurement unit in the non-inertial frame, to obtain the total thrust and three-axis angular velocity control quantities of the rotary-wing UAV, thereby realizing the flight control of the rotary-wing UAV in the non-inertial frame.

[0043] According to a third aspect of the embodiments of this application, an electronic device is provided, comprising:

[0044] One or more processors;

[0045] Memory, used to store one or more programs;

[0046] When the one or more programs are executed by the one or more processors, the one or more processors perform the method as described in the first aspect.

[0047] According to a fourth aspect of the embodiments of this application, a computer-readable storage medium is provided that stores computer instructions thereon, which, when executed by a processor, implement the steps of the method as described in the first aspect.

[0048] The technical solutions provided by the embodiments of this application may include the following beneficial effects:

[0049] As can be seen from the above embodiments, the rotorcraft UAV dynamic model established in this application under a non-inertial frame eliminates the dependence on inertial frame state estimation in traditional modeling methods. Since the non-inertial frame is directly built on the moving target's mechanical system, complex collaborative interaction tasks between the rotorcraft UAV and the moving target can be achieved using only the angular velocity and linear acceleration information of the moving target's inertial measurement unit, without requiring difficult-to-obtain information such as the target motion model as prior knowledge. This avoids the impact of frequent trajectory replanning in previous inertial frame systems. The designed nonlinear model predictive control scheme adds the angular velocity, linear acceleration, and angular acceleration information of the non-inertial frame to the system's state variables, achieving high-precision, high-stability, and high-adaptability non-inertial frame rotorcraft UAV flight motion control effects, and facilitating future system improvements. This application solves the dependence of traditional inertial frame rotorcraft UAV flight control on additional state estimation, target motion models, and frequent trajectory replanning, expands the application of interactive tasks such as fixed-point following, autonomous landing, and trajectory tracking, and verifies the system's adaptability to existing relative pose estimation algorithms.

[0050] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0051] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0052] Figure 1 This is an overall framework diagram illustrating a model predictive control method for a rotorcraft unmanned aerial vehicle based on a non-inertial frame, according to an exemplary embodiment.

[0053] Figure 2 This is a schematic diagram illustrating the relationship between an inertial frame of reference, a non-inertial frame of reference, and a mechanical frame of reference, according to an exemplary embodiment.

[0054] Figure 3 This is a schematic diagram of the dimensions of a rotary-wing drone and an unmanned vehicle according to an exemplary embodiment, where a is a schematic diagram of the dimensions of the rotary-wing drone and b is a schematic diagram of the dimensions of the unmanned vehicle.

[0055] Figure 4 This is a schematic diagram of a rotorcraft drone trajectory tracking task (loop flight) according to an exemplary embodiment, where a is a schematic diagram of the actual test process, b is a schematic diagram of the rotorcraft drone's fixed loop figure-eight trajectory, and c is the actual flight trajectory of the rotorcraft drone during the loop process.

[0056] Figure 5This is a schematic diagram of a rotorcraft drone trajectory tracking task (surround aerial photography) according to an exemplary embodiment, where a represents the rotorcraft drone flight trajectory observed from the first-person perspective of the unmanned vehicle in a non-inertial frame, and b represents the movement trajectory of the rotorcraft drone and the unmanned vehicle observed from a fixed third-person perspective.

[0057] Figure 6 This is a block diagram illustrating a model predictive control device for a rotary-wing unmanned aerial vehicle based on a non-inertial frame, according to an exemplary embodiment.

[0058] Figure 7 This is a schematic diagram of an electronic device according to an exemplary embodiment. Detailed Implementation

[0059] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0060] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0061] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0062] Figure 1 This is a flowchart illustrating a model predictive control method for a rotorcraft unmanned aerial vehicle based on a non-inertial frame, according to an exemplary embodiment. Figure 1 As shown, this method, when applied to the terminal of a rotary-wing UAV, may include the following steps:

[0063] (1) Based on the relationship between the inertial frame, the non-inertial frame and the rotorcraft system, establish a dynamic model of the rotorcraft UAV in the non-inertial frame;

[0064] (2) Based on the dynamic model, design a nonlinear model predictive controller for the rotary-wing UAV;

[0065] (3) Using the sequential quadratic programming algorithm, combined with the relative pose given by the motion capture system or CREPES (《Z.Xun, J.Huang, Z.Li, C.Xu, F.Gao, and Y.Cao, "CREPES: Cooperative RElative Pose EStimation towards Real-World Multi-Robot Systems." arXiv, Feb.02, 2023. doi:10.48550 / arXiv.2302.01036.》) and the state information collected by the inertial measurement unit in the non-inertial frame, the optimization problem in the nonlinear model predictive controller is solved to obtain the total thrust and three-axis angular velocity control quantities of the rotorcraft UAV, thereby realizing the flight control of the rotorcraft UAV in the non-inertial frame.

[0066] This invention utilizes the attitude and velocity of a rotary-wing UAV relative to a reference target, as well as the angular velocity and acceleration information of the reference target in an inertial reference frame, to achieve stable flight of a rotary-wing UAV in a non-inertial frame without relying on additional inertial positioning equipment and state estimation modules (such as GPS, SLAM), prior information of the target motion model, or complex trajectory replanning operations.

[0067] The overall framework of the method proposed in this invention is as follows: Figure 1 As shown, the main steps are as follows: (1) Establish a dynamic model of a non-inertial rotorcraft; (2) Design a nonlinear model predictive controller for the rotorcraft based on the model; (3) Use this nonlinear model predictive controller to control the flight of the rotorcraft and complete the interactive task.

[0068] In the specific implementation of step (1), a dynamic model of the rotor-wing UAV in the non-inertial frame is established based on the relationship between the inertial frame, the non-inertial frame and the rotor-wing UAV system.

[0069] like Figure 2 As shown, the inertial frame, the non-inertial frame, and the rotorcraft UAV system are represented by W, N, and B, respectively. The relative position of the rotorcraft UAV in the non-inertial frame is then determined. N p B Represented as:

[0070]

[0071] in N R W Let represent the rotation matrix of the inertial frame W relative to the non-inertial frame N. Let N represent the relative displacement vector between the non-inertial frame N and the rotorcraft UAV system B, and the inertial frame W. Taking the first and second derivatives of the above equation, we obtain the relative velocity of the rotorcraft UAV in the non-inertial frame. N v B With relative acceleration Represented as:

[0072]

[0073]

[0074] in,[ N Ω N ] × and[ N β N ] × Let N and N represent the angular velocities of the non-inertial frame of reference, respectively. N Ω N With angular acceleration N β N antisymmetric matrix, N R B Let represent the rotation matrix of the rotorcraft UAV system B relative to the non-inertial frame N, and let T represent the normalized total thrust of the rotorcraft UAV, where g = [0, 0, -g]. T Represents gravitational acceleration. N R W a N Let be the linear acceleration of in the non-inertial frame N. For where N R B After differentiation and conversion to quaternion form, the dynamic model of the rotational part is expressed as:

[0075]

[0076]

[0077] in, B Ω B and[ B Ω B ] × Let represent the angular velocity and antisymmetry matrix of rotorcraft UAV system B, respectively. N q B Representing the rotation matrix N R B The quaternion form of ⊙ represents the Hamiltonian product of four arithmetic multiplications.

[0078] relative acceleration The expression contains N R W g and N R W aN Two items, of which N R W The rotation matrix representing the inertial frame W relative to the non-inertial frame N is difficult to obtain directly in practice. This invention solves this problem by using an inertial measurement unit (IMU) mounted on the non-inertial frame N. The linear acceleration of the non-inertial frame N itself is then expressed as... N R W a N Represented as N a N It can be obtained by applying a negative gravity force to the IMU accelerometer, and is expressed as:

[0079]

[0080] in, and It is the linear acceleration value obtained directly from the IMU. Express it in terms of relative acceleration. Substitute the above equation into the expression for relative acceleration. In the expression, It can be reformulated as:

[0081]

[0082] in, and This refers to the angular velocity and linear acceleration of the non-inertial frame N obtained from the IMU. This completes the establishment of the dynamic model of the rotary-wing UAV in the non-inertial frame.

[0083] To address the reliance on inertial frame state estimation in traditional modeling methods, this application essentially eliminates variables related to the inertial frame in the model by establishing a dynamic model of the rotorcraft UAV in a non-inertial frame. Since the non-inertial frame is directly established on the moving target's mechanical system, complex interactive tasks between the rotorcraft UAV and the moving target can be achieved using only the angular velocity and linear acceleration information of the moving target's inertial measurement unit, without requiring prior knowledge such as the target motion model. Furthermore, relative acceleration is eliminated. middle N R W While impacting the system, it also avoids the frequent trajectory replanning operations previously required in inertial frames, reducing the system's computational load and complexity.

[0084] In the specific implementation of step (2), a nonlinear model predictive controller for the rotary-wing UAV is designed based on the dynamic model.

[0085] Since the aforementioned non-inertial frame rotorcraft UAV dynamic model is a highly nonlinear system, controller design is quite complex. Traditional inertial frame rotorcraft UAV control methods typically linearize the nonlinear dynamic equations and utilize the differential flatness of the rotorcraft UAV to design a cascaded PID (proportional-integral-derivative) controller with feedforward. This approach is clearly no longer suitable for the highly nonlinear system established in this application. Therefore, this application treats the rotorcraft UAV as a completely nonlinear dynamic system and uses nonlinear model predictive control to establish the controller, comprehensively considering the limitations of the normalized total thrust and three-axis angular velocities of the rotorcraft UAV in the non-inertial frame.

[0086] Nonlinear model predictive control generates control commands by solving the finite-time optimal control problem (OCP). Its loss function is the error between the predicted state and the reference state within a fixed time range, which means that multiple reference states within that time range are needed for the solution. For numerical optimization, the state and input variables are discretized into n equal intervals over a time range δ∈[t,t+h] of size dt = h / n, thus constructing the final nonlinear optimization problem, where h represents the iteration time range.

[0087] This invention is based on and These are respectively used as the state variables and input variables of the predictive controller in the nonlinear model. These represent the three-axis angular velocities of rotorcraft system B. N p B , N v B , N q B It can be obtained through a motion capture system or by using the CREPES relative pose estimation algorithm. and Obtained from an inertial measurement unit (IMU) in a non-inertial frame, due to angular acceleration N β N Since it is difficult to estimate accurately, this application sets it to 0, that is, it is assumed that the non-inertial frame N rotates at a constant angular velocity in each iteration.

[0088] The last three terms of the state variables and N β N In reality, it does not possess predictive properties; this application treats it as a state variable for the convenience of subsequent implementation and future system improvements. The aforementioned non-inertial frame rotorcraft UAV dynamics model is then expressed using differential equations. The loss function is described in the form of C(x,u) = ||x(t) - x(t) between the current / final state and the reference state. ref ||Q +‖u(t)-u h || R and The final nonlinear optimization problem under the constraints can then be expressed as:

[0089]

[0090] stx(0)=x0

[0091] x(k+1)=f d (x(k),u(k))

[0092] T min ≤T≤T max

[0093]

[0094]

[0095]

[0096] Where, ‖x‖ Q =x T Qx, x(k) and u(k) represent the state and input variables at time k, respectively. ref With x(n) ref U represents the reference state quantity at time k. h Represents the reference input quantity, Q = diag(Q x ), Q x With Q n R is the positive definite weighting matrix for each system state variable, and R is the positive definite weighting matrix for the system input variables. In this application, the reference input variable u h = [g; 0; 0; 0], meaning the input value when the rotorcraft UAV is hovering is always used as the reference input value. For the constraint part, f d (x(k),u(k)) is the discretized representation of the differential equation f(x,u), T min With T max Ω represents the minimum and maximum thrust limits, respectively. rp With Ω yaw These represent the angular velocity limits in the maximum roll, pitch, and yaw directions, respectively. This is the designed nonlinear model predictive controller for the rotary-wing UAV.

[0097] To eliminate the influence of model accuracy, measurement errors, and environmental noise on the system, this application adopts a nonlinear model predictive control framework, adding the angular velocity, linear acceleration, and angular acceleration information of the non-inertial frame to the system's state variables. This achieves high-precision, high-stability, and high-adaptability flight motion control for non-inertial rotorcraft UAVs, and facilitates future system improvements.

[0098] In the specific implementation of step (3), the sequential quadratic programming algorithm is used to solve the optimization problem in the nonlinear model predictive controller by combining the relative pose given by the motion capture system or CREPES and the state information collected by the inertial measurement unit in the non-inertial frame, so as to obtain the total thrust and three-axis angular velocity control quantities of the rotor drone, thereby realizing the flight control of the rotor drone in the non-inertial frame.

[0099] To address the quadratic nonlinear optimization problem of the aforementioned model predictive control framework, this invention solves it by executing the Sequential Quadratic Programming (SQP) algorithm within a real-time iterative framework. This algorithm is implemented using the ACADO toolkit and the qpOASES solver. The discretization of the rotorcraft UAV dynamics model in the optimization problem is achieved using the Multiple Shooting Technique and the Runge-Kutta Integration Scheme. The time range and step size of each controller iteration are set to h = 2s and dt = 0.1s, respectively.

[0100] Solving the model using the above method yields the total thrust and three-axis angular velocity control variables u = This enables flight control of rotary-wing UAVs in a non-inertial frame.

[0101] This application proposes a model predictive control method for a rotary-wing unmanned aerial vehicle (UAV) based on a non-inertial frame of reference. By utilizing the UAV's attitude and velocity relative to a reference target, as well as the reference target's angular velocity and acceleration in an inertial reference frame, stable flight of the UAV in a non-inertial frame can be achieved. Compared with traditional control methods in an inertial frame, this method eliminates the need for any state estimation operations in the inertial frame. Furthermore, since the reference frame is directly established on the moving target's mechanical system, it avoids the need for difficult-to-obtain information such as the target's motion model as prior knowledge for prediction. It also avoids frequent and complex trajectory replanning operations in the inertial frame.

[0102] This method can be easily extended to complex cooperative interactive tasks between rotorcraft UAVs and moving targets, such as point-to-point following, autonomous landing, and trajectory tracking. For point-to-point following tasks, this method only requires a fixed point containing the rotorcraft UAV's state variables as feedback; for more complex tasks such as autonomous landing and trajectory tracking, it only requires a pre-planned fixed trajectory as a reference, without any complex trajectory replanning operations. To verify the feasibility and stability of this method, this invention conducted extensive tests on the aforementioned interactive tasks.

[0103] Fixed-point following test: A non-inertial frame (unmanned vehicle) is set to travel at different speeds v and angular velocities ω. A rotary-wing UAV follows behind the unmanned vehicle at different distances r. Its flight altitude is fixed at 1m above the unmanned vehicle. The tracking error is shown in the table below.

[0104] Table 1 Tracking Error in Fixed-Point Following Test

[0105] (r,v,ω) setting values (1.0,1.0,0.31) (1.0,1.0,0.71) (1.0,0.3,1.0) (1.0,0.7,1.0) (0.3,1.0,1.0) (0.7,1.0,1.0) Tracking error / m 0.13 0.23 0.13 0.14 0.11 0.16

[0106] Autonomous landing test: The non-inertial frame (unmanned vehicle) was set to travel at different speeds v and angular velocities ω. The rotary-wing UAV started the autonomous landing task from different distances r behind the unmanned vehicle and at a fixed height of 1m above the unmanned vehicle. The tracking error is shown in the table below.

[0107] Table 2 Tracking Error of Autonomous Landing Test

[0108] (r,v,ω) setting values (4.0,1.0,0.31) (4.0,1.0,0.71) (4.0,0.3,1.0) (4.0,0.7,1.0) (3.3,1.0,1.0) (3.7,1.0,1.0) Tracking error / m 0.15 0.19 0.21 0.24 0.18 0.23

[0109] The maximum tracking errors in the above-mentioned fixed-point following and autonomous landing tests were 0.23m and 0.24m, respectively, indicating that this method still has good tracking performance even during periods of rapid target movement, ensuring the successful completion of the mission. The state estimation information of the rotary-wing UAV relative to the unmanned vehicle is provided by an indoor motion capture system.

[0110] Trajectory tracking task: A non-inertial frame (unmanned vehicle) is set to travel along an S-shaped trajectory with a maximum speed of 0.38 m / s. A rotary-wing UAV tracks a pre-planned fixed trajectory while the unmanned vehicle moves. To further test the stability of this method, this application designs a highly challenging trajectory tracking task (looping flight). Figure 3 As shown, the unmanned vehicle is equipped with four rings with an inner diameter of 420mm, forming a square with a side length of 600mm. The rotary-wing drone needs to continuously traverse these rings during trajectory tracking. The rotary-wing drone measures 197.5mm × 203.9mm. The test results are as follows... Figure 4As shown in the figure, a is a schematic diagram of the actual test process, b is a schematic diagram of the fixed loop-and-figure-eight trajectory of the rotary-wing UAV, and c is the actual flight trajectory of the rotary-wing UAV during the loop-and-figure process in a non-inertial frame. The test results shown in the figure demonstrate that this method has good performance capabilities for complex trajectory tracking tasks, in which the relative state estimation information is also provided by the indoor motion capture system.

[0111] To further verify the applicability of this method to other relative state estimation methods, this invention utilizes methods and devices in CREPES to provide state estimation information of a rotary-wing UAV relative to a non-inertial frame of reference, and conducts an outdoor trajectory tracking task (circling aerial photography) test. In this task, the unmanned vehicle moves randomly, and the rotary-wing UAV tracks a fixed circular trajectory (radius of 1.5m) during this process. The test results are as follows. Figure 5 As shown in the figure, 'a' represents the flight trajectory of the rotorcraft UAV as observed from the first-person perspective (non-inertial frame) of the autonomous vehicle, and 'b' represents the motion trajectories of the rotorcraft UAV and the autonomous vehicle as observed from a fixed third-person perspective. This result demonstrates that this method exhibits strong adaptability and robustness to traditional state estimation methods, and possesses broad practical application value and prospects.

[0112] Furthermore, this method can be easily extended to other multi-robot systems such as vehicle-machine collaboration, multi-robot formation, and collaborative SLAM.

[0113] Corresponding to the aforementioned embodiments of the model predictive control method for rotorcraft unmanned aerial vehicles based on non-inertial frames of reference, this application also provides embodiments of a model predictive control device for rotorcraft unmanned aerial vehicles based on non-inertial frames of reference.

[0114] Figure 6 This is a block diagram illustrating a predictive control device for a rotorcraft unmanned aerial vehicle (UAV) model based on a non-inertial frame of reference, according to an exemplary embodiment. (Refer to...) Figure 6 The device may include:

[0115] Modeling module 21 is used to establish a dynamic model of a rotary-wing UAV in a non-inertial frame based on the relationship between the inertial frame, the non-inertial frame and the rotary-wing UAV system.

[0116] Design module 22 is used to design a nonlinear model predictive controller for a rotary-wing UAV based on the dynamic model.

[0117] The solution module 23 is used to solve the optimization problem in the nonlinear model predictive controller by using a sequential quadratic programming algorithm, combined with the relative pose given by the motion capture system or CREPES and the state information collected by the inertial measurement unit in the non-inertial frame, to obtain the total thrust and three-axis angular velocity control quantities of the rotor-wing UAV, thereby realizing the flight control of the rotor-wing UAV in the non-inertial frame.

[0118] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0119] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0120] Accordingly, this application also provides an electronic device, including: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the above-described predictive control method for a rotorcraft unmanned aerial vehicle based on a non-inertial frame of reference. Figure 7 The diagram shown is a hardware structure diagram of any device with data processing capabilities for a model predictive control method for a rotary-wing unmanned aerial vehicle based on a non-inertial frame of reference provided in an embodiment of the present invention, except for... Figure 7 In addition to the processor, memory, and network interface shown, any data processing device in the embodiment may also include other hardware depending on the actual function of the data processing device, which will not be described in detail here.

[0121] Accordingly, this application also provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the aforementioned predictive control method for a rotorcraft unmanned aerial vehicle based on a non-inertial frame of reference. The computer-readable storage medium can be an internal storage unit of any data-processing device as described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., equipped on the device. Furthermore, the computer-readable storage medium can include both internal storage units of any data-processing device and external storage devices. The computer-readable storage medium is used to store the computer program and other programs and data required by the data-processing device, and can also be used to temporarily store data that has been output or will be output.

[0122] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0123] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A predictive control method for a rotorcraft unmanned aerial vehicle (UAV) model based on a non-inertial frame of reference, characterized in that, include: (1) Based on the relationship between the inertial frame, the non-inertial frame and the rotorcraft system, establish a dynamic model of the rotorcraft UAV in the non-inertial frame; (2) Based on the aforementioned dynamic model, design a nonlinear model predictive controller for the rotary-wing UAV; (3) Using the sequential quadratic programming algorithm, combined with the relative pose given by the motion capture system or CREPES and the state information collected by the inertial measurement unit in the non-inertial frame, the optimization problem in the nonlinear model predictive controller is solved to obtain the total thrust and three-axis angular velocity control quantities of the rotor UAV, thereby realizing the flight control of the rotor UAV in the non-inertial frame. Step (1) includes: Based on the relationship between the inertial frame, the non-inertial frame, and the rotorcraft UAV system, the relative position of the rotorcraft UAV in the non-inertial frame can be obtained. ,in Inertial frame of reference Relative to non-inertial frames of reference The rotation matrix, Representing non-inertial frames of reference Rotary-wing UAV system With inertial frame The relative displacement vector below; By analyzing the above formula By taking the derivative, we can obtain the relative velocity of the rotary-wing UAV in the non-inertial frame. With relative acceleration Represented as: ， ， in, and Representing non-inertial frames of reference respectively angular velocity With angular acceleration antisymmetric matrix, Indicating a rotary-wing unmanned aerial vehicle system Relative to non-inertial frames of reference The rotation matrix, This represents the normalized total thrust of a rotary-wing unmanned aerial vehicle. Represents gravitational acceleration. Non-inertial frame of reference The following represents its own linear acceleration; By analyzing the relative velocity and relative acceleration... By taking the derivative and converting it to quaternion form, we obtain the dynamic model of the rotating part: ， ， in, and These respectively represent the rotorcraft unmanned aerial vehicle (UAV) system. Angular velocity and its antisymmetric matrix, Representing the rotation matrix The quaternion form, Represents the Hamiltonian product of four arithmetic multiplications.

2. The method according to claim 1, characterized in that, The nonlinear model predicts the state variables of the controller. Input quantity ,in The relative position of the rotary-wing UAV in a non-inertial frame of reference. Let be the relative velocity of the rotary-wing UAV in a non-inertial frame of reference. Rotation matrix The quaternion form represents the relative attitude of a rotary-wing UAV in a non-inertial frame. and Non-inertial frames of reference obtained directly from the IMU are respectively Angular velocity and linear acceleration values, Non-inertial frame of reference angular acceleration, This represents the normalized total thrust of a rotary-wing unmanned aerial vehicle. These are rotary-wing unmanned aerial vehicle systems. The three-axis angular velocity.

3. The method according to claim 1, characterized in that, The nonlinear model prediction controller is: ， ， in, , and express State and input quantities at any given time. and express Reference state quantity at time t, Indicates the reference input quantity. That is, the input value when the rotorcraft is hovering is always used as the reference input value. , and Here is the positive definite weight matrix for each system state variable. The system input is a positive definite weight matrix; for the constraint part, It is the differential equation form of the dynamic model. Discretized representation, and These represent the minimum and maximum thrust limits, respectively. and These represent the angular velocity limits in the roll, pitch, and yaw directions, respectively.

4. The method according to claim 2, characterized in that, In step (3), the controller is solved based on the relative pose given by the motion capture system or CREPES and the non-inertial frame angular velocity and linear acceleration directly obtained from the inertial measurement unit, and the estimated and reference values ​​of its angular acceleration are set to 0, that is, it is assumed that the non-inertial frame always moves on the plane.

5. A predictive control device for a rotorcraft unmanned aerial vehicle model based on a non-inertial frame of reference, characterized in that, include: The modeling module is used to establish a dynamic model of a rotary-wing UAV in a non-inertial frame based on the relationship between the inertial frame, the non-inertial frame, and the rotary-wing UAV system. The design module is used to design a nonlinear model predictive controller for the rotary-wing UAV based on the dynamic model. The solution module is used to solve the optimization problem in the nonlinear model predictive controller by using a sequential quadratic programming algorithm, combined with the relative pose given by the motion capture system or CREPES and the state information collected by the inertial measurement unit in the non-inertial frame, to obtain the total thrust and three-axis angular velocity control quantities of the rotary-wing UAV, thereby realizing the flight control of the rotary-wing UAV in the non-inertial frame. In the modeling module: Based on the relationship between the inertial frame, the non-inertial frame, and the rotorcraft UAV system, the relative position of the rotorcraft UAV in the non-inertial frame can be obtained. ,in Inertial frame of reference Relative to non-inertial frames of reference The rotation matrix, Representing non-inertial frames of reference Rotary-wing UAV system With inertial frame The relative displacement vector below; By analyzing the above formula By taking the derivative, we can obtain the relative velocity of the rotary-wing UAV in the non-inertial frame. With relative acceleration Represented as: ， ， in, and Representing non-inertial frames of reference respectively angular velocity With angular acceleration antisymmetric matrix, Indicating a rotary-wing unmanned aerial vehicle system Relative to non-inertial frames of reference The rotation matrix, This represents the normalized total thrust of a rotary-wing unmanned aerial vehicle. Represents gravitational acceleration. Non-inertial frame of reference The following represents its own linear acceleration; By analyzing the relative velocity and relative acceleration... By taking the derivative and converting it to quaternion form, we obtain the dynamic model of the rotating part: ， ， in, and These respectively represent the rotorcraft unmanned aerial vehicle (UAV) system. Angular velocity and its antisymmetric matrix, Representing the rotation matrix The quaternion form, Represents the Hamiltonian product of four arithmetic multiplications.

6. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-4.

7. A computer-readable storage medium storing computer instructions thereon, characterized in that, When executed by the processor, this instruction implements the steps of the method as described in any one of claims 1-4.

Images