Small crawler wheel wing combined type deformable land-air amphibious reconnaissance robot and control method thereof

Abstract

The invention discloses a small crawler wheel wing combined type deformable land-air amphibious reconnaissance robot and a control method thereof. The robot adopts a crawler wheel wing integrated miniaturization design and comprises a control cabin, a left crawler wheel wing system and a right crawler wheel wing system, the crawler wheel wing systems are connected with the control cabin through steering engines, and a flight mode and a ground mode can be stably switched under the driving of the steering engines. In the flight mode, a flight motor of the crawler wheel wing system drives propellers to rotate at a high speed, lift force is provided, and air flight and attitude stability are achieved. In the ground mode, the crawler wheel wing system is adjusted to the ground contact position, the propellers stop working, the ground motor drives the crawler and the driving wheels through the gear set to provide advancing power, the driven wheels assist in supporting, and the stability and obstacle crossing ability in the complex terrain are enhanced. A multi-mode control method based on model predictive control is adopted for the robot, and the multi-mode control method is achieved by establishing control models of a flight mode and a ground mode, constructing a unified state-space equation and a predictive equation, designing a cost function, dynamically switching the control models and coordinating constraint conditions and control targets. And the trajectory tracking precision and the system stability of the robot in the flight mode and the ground mode are ensured.

Description

Technical Field

[0001] The invention belongs to the technical field of reconnaissance robots, and in particular to a small track-wheel-wing composite deformable land-air amphibious reconnaissance robot and a control method thereof. Background Art

[0002] With the rapid development of robotics technology, intelligent robots are increasingly used in reconnaissance, rescue and other tasks. Especially in complex environments and disaster relief scenarios, the demand for robots that can quickly adapt to a variety of terrains and environments has increased significantly. However, existing reconnaissance robots are mainly divided into two categories: ground robots and flying robots, each with its own advantages and disadvantages, and it is difficult to fully meet the needs of complex tasks.

[0003] In order to solve the above problems, multimodal amphibious robots have gradually become a research hotspot in recent years. These robots integrate ground and flight capabilities, trying to combine the endurance advantages of ground robots with the leaping capabilities of flying robots. However, current multimodal robots have problems such as large size, insufficient mode switching stability, and high complexity of control methods. Summary of the invention

[0004] The purpose of the present invention is to provide a small wheel-wing composite deformable land-air amphibious reconnaissance robot and a control method thereof in order to address the problems existing in the above-mentioned prior art, so as to meet the requirements of multi-environment adaptability, functional diversity and efficient resource utilization in complex reconnaissance missions.

[0005] The technical solution for achieving the purpose of the present invention is: a small track-wheel-wing composite deformable land-air amphibious reconnaissance robot, the robot comprising a control cabin, a left track-wheel-wing system and a right track-wheel-wing system, wherein the left track-wheel-wing system and the right track-wheel-wing system are symmetrically installed on both sides of the control cabin, respectively, and connected to the control cabin through a steering gear; the left track-wheel-wing system and the right track-wheel-wing system are both integrated with a flight motor, a propeller, a ground motor, a driving wheel, a driven wheel and a track, and the switching between the flight mode and the ground mode is realized under the drive of the steering gear; in the flight mode, the flight motors of the left track-wheel-wing system and the right track-wheel-wing system drive the propeller to rotate at a high speed to provide lift, thereby realizing aerial flight and attitude stability; in the ground mode, the left track-wheel-wing system and the right track-wheel-wing system are rotated to the ground contact position by the steering gear, the propeller stops working, and the ground motor is driven by a gear set to realize ground travel.

[0006] Furthermore, the control cabin includes: a main control module, an integrated flight control and electric adjustment module, a GPS module, a camera and an image transmission module; the main control module is responsible for managing the operation logic of the whole machine, deformation control of the track wheel wing system, ground motion control and mode switching; the integrated flight control and electric adjustment module is responsible for the posture stability and power management of the robot in flight mode, and realizes high-precision control during flight by directly controlling the output of the flight motor; the GPS module is used to obtain position information in real time to provide support for navigation and path planning; the image transmission module is connected to the camera to transmit the real-time image to the remote monitoring terminal for environmental reconnaissance and mission command.

[0007] Furthermore, the left track wheel wing system and the right track wheel wing system both adopt an integrated design, the flight motor is directly connected to the propeller, and the propeller is installed concentrically with the drive wheel, so as to provide lift in the flight mode and realize aerial flight and attitude control; the ground motor is connected to the drive wheel through a gear set, driving the drive wheel to provide travel power in the ground mode; the track is arranged around the drive wheel and the driven wheel, and the driven wheel provides tension and support for the track; the left track wheel wing system and the right track wheel wing system are driven by a high-torque servo to rotate around a fixed axis, thereby realizing the morphological switching between the flight mode and the ground mode; the main controller coordinates the servo action in real time to ensure the smoothness and precision of the deformation process.

[0008] Furthermore, in flight mode, the flight motor of the track-wheel-wing system drives the propeller to rotate at high speed to provide thrust, enabling the robot to fly stably. At this time, the driving wheels and the driven wheels are in a folded state or a suspended state without contacting the ground, so as to reduce the air resistance and inertial burden in the overall flight mode; the servo drives the track-wheel-wing system to maintain a fixed angle in the flight mode, optimizing the aerodynamic performance and center of gravity distribution of the track-wheel-wing system; the integrated flight control and electric adjustment module monitors the robot's posture in real time, and achieves posture balance, position maintenance and heading control by adjusting the rotation speed of each propeller; the main control module works together with the integrated flight control and electric adjustment module to adjust the power output strategy according to environmental perception data.

[0009] Furthermore, in ground mode, the left and right track wheel wing systems independently adjust the angles through the servos to adapt to the requirements of different terrains and ground conditions. The track wheel wing systems are driven by the servos to adjust to a position suitable for ground driving. The ground motor drives the drive wheel through the gear set to provide the robot with forward, backward and turning movement power; the driven wheel cooperates with the track to provide stable support; the track is arranged around the drive wheel and the driven wheel, and the robot can pass through different complex terrains through the continuous grounding of the track; the main control module adjusts the speed and direction of the drive wheel in real time, and combines the path planning algorithm to achieve precise ground motion control.

[0010] Furthermore, when switching between flight mode and ground mode, the track-wheel-wing system, under the command of the main control module, drives the left and right track-wheel-wing systems to adjust their positions synchronously or sequentially through a high-torque servo, thereby achieving a smooth transition from flight mode to ground mode or from ground mode to flight mode; when switching from flight mode to ground mode, the servo drives the track-wheel-wing system to adjust to a position suitable for ground driving, and at the same time the ground motor and drive wheels enter working state to provide power support for ground travel; the tracks contact the ground, and the driven wheels provide additional stability to ensure stability and obstacle surmounting capability in ground mode; when switching from ground mode to flight mode, the servo drives the track-wheel-wing system to adjust to the flight mode position to ensure that the propeller has sufficient working space, the ground motor stops working, and the flight motor drives the propeller to rotate at high speed to provide lift and attitude control to achieve a smooth takeoff.

[0011] On the other hand, a control method for a small track-wheel-wing composite deformable land-air amphibious reconnaissance robot is provided, wherein the control method adapts to two modes by dynamically switching the dynamic model, and the robot position control uses a multi-modal control algorithm based on model prediction; the control method comprises the following steps:

[0012] Step 1, establish the robot dynamics model and establish the state space equation;

[0013] Step 2: construct the prediction equation, define the constraints and design the cost function;

[0014] Step 3: Construct an optimization problem, solve the optimal control sequence and output the control quantity.

[0015] Furthermore, the step 1 of establishing the robot dynamics model and establishing the state space equation specifically includes:

[0016] Step 1-1, establish the dynamic model of the amphibious reconnaissance robot in flight mode:

[0017] Select the state variable X as:

[0018] X=[xyzv x v y v z φ θ ψ ω φ ω θ ω ψ ] T

[0019] Among them, x, y, z are the position coordinates in the geodetic coordinate system, v x ,v y ,v z is the linear velocity component in the geodetic coordinate system, φ, θ, ψ are the roll angle, pitch angle, and yaw angle, ωφ ,ω θ ,ω ψ is the angular velocity component in the robot coordinate system;

[0020] The output is:

[0021] Y=[xyz φ θ ψ] T

[0022] The input is:

[0023] U=[ω1 ω2 ω3 ω4] T

[0024] Among them, ω i is the speed of the i-th flight motor; i=1,2,3,4;

[0025] The thrust and torque of the robot in flight mode are determined by the motor speed, and the total thrust is where k f is the thrust coefficient; the rolling moment is Where L' is the distance from the flight motor to the center of mass of the robot; the pitch moment is The yaw moment is Among them, k m is the moment coefficient;

[0026] Then the dynamic model of the amphibious reconnaissance robot in flight mode is:

[0027]

[0028] Among them, v x , v y , v z is the linear velocity component in the geodetic coordinate system, F is the total thrust, m is the robot mass, g is the gravitational acceleration, φ, θ, ψ are the roll angle, pitch angle, and yaw angle respectively, ω φ ,ω θ ,ω ψ is the angular velocity component in the robot coordinate system, J xx ,J yy ,J zz is the main diagonal element of the inertia matrix, m is the mass of the robot, g is the gravitational acceleration, τ φ ,τ θ ,τ ψ are the rolling moment, pitching moment and yaw moment;

[0029] Step 1-2, establish the discrete form state space equation of the amphibious reconnaissance robot in flight mode:

[0030] Assume the current time is k, and write the discrete state space equation according to the dynamic model in step 1-1 as:

[0031] X k+1 =A1X k +B1U k

[0032] Among them, X k+1 , X k are the state variables at time k+1 and k respectively, A1 is the state matrix, which reveals the change law of the state variable over time, and B1 is the input matrix, which reflects the response of the system to external input;

[0033] Steps 1-3, establish the kinematic model of the amphibious reconnaissance robot in ground mode:

[0034] Select the state variable X as:

[0035] X=[xyθ] T

[0036] Where x, y, θ are the position and orientation angle of the ground mode respectively;

[0037] The control input U is:

[0038]

[0039] Among them, v L is the left wheel speed, v R is the right wheel speed;

[0040] The kinematic model of the amphibious reconnaissance robot in ground mode is:

[0041]

[0042] Where r is the wheel radius, L is the wheel spacing in ground mode;

[0043] Steps 1-4: Establish the discrete state space equations in ground mode

[0044] Assume the current time is k, and write the discrete state space equation according to the kinematic model in the ground mode in steps 1-3 as:

[0045] X(k+1)=A2X(k)+B2U(k)

[0046] Among them, X(k+1) and X(k) are the state variables at time k+1 and k respectively, A2 is the state matrix, which reveals the change law of the state variable over time, B2 is the input matrix, which reflects the response of the system to external input, and U(k) is the control input at time k.

[0047] Furthermore, in step 2, a prediction equation is constructed based on the state space equation to describe the relationship between the state output and the control input in the prediction time domain; the constraints of the multimodal system are defined, including the input range and state constraints in the flight mode and the ground mode; and the cost function is designed to minimize the reference trajectory deviation and the change of the control increment; specifically, it includes:

[0048] Step 2-1, convert the linear discrete state space model into incremental form:

[0049]

[0050] Among them, Δx(k+1) and Δx(k) are the state increments at time k+1 and k respectively, Δu(k) is the input increment at time k, y(k+1) is the system output at time k+1, A is the state matrix, B is the input matrix, and C is the output matrix;

[0051] Step 2-2, construct the prediction equation:

[0052] Set the prediction horizon of the MPC controller to l and the control horizon to c;

[0053] Assume that the control quantity outside the control time domain does not change, that is, Δu(k+i)=0,i=c,c+1,…l-1,

[0054] Then the prediction equation is obtained to predict the system state increment from time k+1 through time k+c to time k+l:

[0055] Δx(k+1|k)=AΔx(k)+BΔu(k)

[0056]

[0057] In the formula, Δx(k+1|k) represents the predicted state at the next moment k+1, Δx(k+c|k) represents the state increment of the cth step in the prediction time domain, and A c represents the state transition of the system after step c, It describes the cumulative impact of multiple future input increments on the state, Δu(k+i) is the input increment at time k+i, Δx(k+l|k) represents the state increment at step l in the prediction time domain, and A l It represents the state transition of the system after l steps. Represents the cumulative effect of all future input increments on the state;

[0058] According to the output equation, the relationship between the output of the system from time k to time k+c and the controlled output is:

[0059] y(k+1|k)=CAΔx(k)+CBΔu(k)+y(k)

[0060]

[0061] In the formula, y(k+1|k) represents the predicted output at the next moment k+1, y(k+c|k) represents the predicted output at the moment k+c, and y(k) is the system output at the moment k. Indicates the cumulative effect of the current state increment on the output after c steps; represents the impact of the input increment of the future c steps on the output, Δu(k+cj) represents the input increment at time k+cj, and y(k+l|k) represents the predicted output at time k+l. It indicates the influence of the current state increment on the output after l steps. represents the impact of the input increment of the next l steps on the output, and Δu(k+lj) represents the input increment at the k+lj moment;

[0062] Define the predicted output vector Y at the current time k and time step l l (k+1|k), define the control input sequence of the system when the control time domain is c as ΔU(k), and obtain the future output prediction equation of the system at time k:

[0063] Y l (k+1|k)=Q x Δx(k)+Q u ΔU(k)+Py(k)

[0064] Among them, Y l (k+1|k) indicates the system output vector with a prediction time domain of l starting from time k+1, Q x Indicates the impact of the initial state increment on future output, Q u It represents the cumulative impact of input increments on future outputs, and Py(k) means incorporating the current output y(k) into the prediction;

[0065] in,

[0066] Step 2-3, constraints:

[0067] In view of the multi-modal control problem, the dynamic model and constraints are switched according to the current mode of the robot. In the flight mode, the dynamic model is described by six degrees of freedom, the control input is the propeller thrust, and the system constraints include the upper and lower limits of the thrust and angular torque; in the ground mode, the dynamic model is based on the differential chassis, the control input is the speed of the drive wheel, and the system constraints include the maximum speed, the speed difference, and the contact force limit between the track and the ground;

[0068] The constraints of the system control quantity are:

[0069] umin ≤u(k+i)≤u max ,i=0,1,…,c-1

[0070] Among them, u min ,u max They represent the minimum and maximum values ​​of the control input respectively, and u(k+i) represents the control input at the i-th prediction step at time k, i.e., at time k+i;

[0071] The constraints of the system control increment are:

[0072] Δu min ≤Δu(k+i)≤Δu max ,i=0,1,…,c-1

[0073] Among them, Δu min ,Δu max They represent the minimum and maximum values ​​of the control increment respectively, and Δu(k+i) represents the control increment at the i-th prediction step at time k, i.e., at time k+i;

[0074] Step 2-4, design the cost function:

[0075] The cost function for designing position control is:

[0076] J(x(k),ΔU(k))=||M y (Y l (k+1|k)-N(k+1))|| 2 +||M u ΔU(k)|| 2

[0077] Among them, M y Represents the weighting factor of the system output, Y l (k+1|k) represents the system prediction output, N(k+1) represents the reference trajectory, and M u represents the weighting factor of the control increment;

[0078] Write the cost function in matrix form:

[0079]

[0080] in, is the weight matrix of the system output with dimension l×l, M yi is the weight coefficient, i=1,2,...,l; is the weight matrix of the system control increment with dimension c×c, M uj is the weight coefficient, is the reference input sequence of the system at time k+1, and n(k+i) is the i-th reference point.

[0081] Furthermore, in step 3, the prediction equation and the cost function are combined to construct a quadratic programming problem, and the optimal control sequence is solved by the optimization algorithm; the current control amount is output in each control cycle, and the state is updated in a rolling manner to achieve accurate tracking of the target position by the robot; step 3 specifically includes:

[0082] Step 3-1, construct the optimization problem:

[0083] For the position control tracking problem, the MPC algorithm obtains the optimal control sequence by solving an optimization problem, which meets the system constraints while minimizing the cost function; the following quadratic optimization problem is constructed:

[0084] min z T Hz-g T z

[0085] stb≤C'z

[0086] Among them, H is the Hessian matrix, g is the gradient vector, z is the optimal independent variable, C' is the constraint matrix, and b is the right-hand side vector of the constraint;

[0087] Substituting the prediction equation into the objective function, we get:

[0088] J(x(k),ΔU(k))=||M y (Y(k+1|k)-N(k+1))|| 2 +||M u ΔU(k)|| 2

[0089] =||M y (Q x Δx(k)+Q u ΔU(k)+Py(k)-N(k+1))|| 2 +||M u ΔU(k)|| 2

[0090] Simplify the expression, let M y (N(k+1)-Py(k)-Q x Δx(k))=T(k+1), rewrite the above formula as:

[0091] J(x(k),ΔU(k))=ΔU(k) T HΔU(k)-G(k+1|k) T ΔU(k)

[0092] Among them, the Hessian matrix Gradient Vector

[0093] The matrix form of control increment constraint and control quantity constraint is:

[0094]

[0095] Among them, R1 and R2 both represent diagonal block matrices, I 4×4 represents the unit diagonal matrix; Δu min (*), Δu max (*) respectively represent the minimum and maximum values ​​of the control input at the time “*”, u min (*),u max (*) indicates the minimum and maximum values ​​of the control input at the time “*”;

[0096] Step 3-2, solve the optimal control sequence and output:

[0097] In each control cycle, the interior point method is used to obtain the optimal control sequence ΔU in the control time domain c. * (k)=[Δu * (k),Δu * (k+1),…,Δu * (k+cl)] T ; where Δu * (k+i) represents the optimal control increment at time k+i, i=0,1,....,cl;

[0098] According to the basic principle of MPC, the first step of the optimal control sequence is applied to the controlled system to obtain a new system output. The control amount of the current control cycle is determined based on the control amount input to the controlled system in the previous control cycle, that is, u(k) = u(k-1) + Δu * (k), where u(k) is the control output at the current moment k, u(k-1) is the control output at the previous moment k-1, Δu * (k) is the optimal control increment at the current moment; and in the next time step, a new quadratic programming problem is solved with the new system state. After the system state is updated, the quadratic programming problem is solved again, and the rolling optimization is performed until the reference position control is completed.

[0099] Compared with the prior art, the present invention has the following significant advantages:

[0100] (1) The present invention adopts an innovative integrated track-wheel-wing design, integrating the ground track driving system and the aerial flight propulsion system into a single module. Existing amphibious robots usually adopt a combined structure of a track chassis and a quad-rotor. Although they have certain multi-modal capabilities, they are large in size and heavy in weight, which limits their flexibility and adaptability in complex scenarios. Existing deformable robots are mostly wheeled designs, and the wheel-wing module has poor passability in complex terrains such as sand and mud in ground mode. The present invention innovatively proposes an integrated track-wheel-wing design, integrating the ground track driving system and the flight propulsion system into a single module, and realizing efficient switching between ground mode and aerial mode through a concentric shaft structure and a high-torque servo drive. The track module significantly improves the robot's adaptability to complex terrain (such as sand, mud, gravel roads, etc.) in ground mode, while the propeller provides high-view reconnaissance capabilities and flexible maneuverability in aerial mode.

[0101] (2) The present invention significantly reduces the overall size of the robot through the innovative integrated track-wheel-wing composite structure, and improves the robot's endurance in ground mode and air mode. Compared with existing deformable wheel-wing robots, the present invention achieves multi-modal motion while retaining excellent ground passability and flight flexibility, and is particularly suitable for long-endurance operations in narrow spaces and complex environments.

[0102] (3) A multimodal motion control method based on model predictive control is proposed. This method integrates the dynamic models of the ground mode and the air mode into a unified control model by constructing a unified control algorithm framework. By dynamically modeling the dynamic characteristics of the ground mode and the air mode, combined with optimized cost function design and constraint processing, the algorithm can achieve precise control and smooth transition of the robot motion during mode switching. Compared with traditional sub-mode control methods, the unified control algorithm of the present invention has significant resource utilization advantages in multimodal tasks. By simplifying the control logic and reducing the computational complexity, it effectively adapts to the limited computing resources of miniaturized robots. On resource-constrained embedded platforms, this control method effectively saves program storage space and running memory, significantly reduces the computing requirements of the system under resource-constrained conditions, and ensures the control accuracy and system stability of the deformation process by optimizing trajectory tracking and state control in real time.

[0103] (4) The present invention combines the integrated track-wheel-wing structural design with a unified multi-modal control algorithm, and has significant advantages in terms of structural compactness, environmental adaptability, control accuracy, and resource utilization efficiency. The innovative track-wheel-wing module enables the robot to have both ground obstacle crossing capabilities and aerial maneuverability, while achieving miniaturization and lightweight through structural optimization; the unified control algorithm effectively solves the problem of decentralized control logic in multi-modal systems, and achieves efficient motion control and stable mode switching under resource-constrained conditions. The present invention is particularly suitable for complex mission scenarios such as disaster relief, environmental reconnaissance, and security patrols, and has important technical value and broad application prospects.

[0104] The present invention is further described in detail below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0105] Figure 1 This is the overall structure diagram of the transformable land and air amphibious reconnaissance robot.

[0106] Figure 2 Schematic diagram of the ground mode of the transformable land-air amphibious reconnaissance robot.

[0107] Figure 3 This is the front view of the transformable land and air amphibious reconnaissance robot.

[0108] Figure 4 Schematic diagram of the track, wheel and wing system of the transformable land and air amphibious reconnaissance robot.

[0109] Reference numerals: 1, control cabin; 2, left wheel wing system; 3, right wheel wing system;

[0110] 21. Flight motor; 22. Propeller; 23. Driving wheel; 24. Ground motor; 25. Driven wheel; 26. Track. DETAILED DESCRIPTION

[0111] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0112] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0113] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0114] In one embodiment, in combination Figure 1 to Figure 2 , provides a small track-wheel-wing composite deformable land-air amphibious reconnaissance robot, the robot comprises a control cabin, a left track-wheel-wing system and a right track-wheel-wing system, wherein the left track-wheel-wing system and the right track-wheel-wing system are symmetrically installed on both sides of the control cabin, respectively, and connected to the control cabin through a steering gear; the left track-wheel-wing system and the right track-wheel-wing system are both integrated with a flight motor, a propeller, a ground motor, a driving wheel, a driven wheel and a track, and the switching between the flight mode and the ground mode is realized under the drive of the steering gear; in the flight mode, the flight motors of the left track-wheel-wing system and the right track-wheel-wing system drive the propeller to rotate at a high speed to provide lift, so as to realize air flight and attitude stability; in the ground mode, the left track-wheel-wing system and the right track-wheel-wing system are rotated to the ground contact position through the steering gear, the propeller stops working, and the ground motor is driven by a gear set to realize ground travel.

[0115] Furthermore, in one of the embodiments, the control cabin includes: a main control module, an integrated flight control and electric adjustment module, a GPS module, a camera and an image transmission module; the main control module is responsible for managing the operation logic of the whole machine, deformation control of the track wheel wing system, ground motion control and mode switching; the integrated flight control and electric adjustment module is responsible for the attitude stability and power management of the robot in flight mode, and achieves high-precision control during flight by directly controlling the output of the flight motor; the GPS module is used to obtain position information in real time to provide support for navigation and path planning; the image transmission module is connected to the camera to transmit real-time images to the remote monitoring terminal for environmental reconnaissance and mission command.

[0116] Further, in one embodiment, in combination Figure 3 and Figure 4The left track wheel wing system and the right track wheel wing system are both integrated and designed. The flight motor 21 is directly connected to the propeller 22, and the propeller 22 is installed concentrically with the drive wheel 23 to provide lift in the flight mode to achieve air flight and attitude control; the ground motor 24 is connected to the drive wheel 23 through a gear set to drive the drive wheel 23 to provide travel power in the ground mode; the track 26 is arranged around the drive wheel 23 and the driven wheel 25 to improve the adaptability to complex terrain and obstacle crossing performance in the ground mode; the driven wheel 25 provides tension and support for the track 26 to ensure the stability and efficient operation of the robot in the ground mode; the flight motor and the ground motor are respectively integrated into the left-right symmetrical track wheel wing module, and the propeller and the drive wheel adopt a concentric shaft structure design. The left track wheel wing system and the right track wheel wing system are driven by a large torque steering gear to rotate around a fixed axis, thereby realizing the morphological switching between the flight mode and the ground mode; the main controller coordinates the steering gear action in real time to ensure the smoothness and accuracy of the deformation process.

[0117] Specifically, in flight mode, the flight motor of the crawler wheel wing system drives the propeller to rotate at high speed to provide thrust, so that the robot can fly stably. At this time, the driving wheel and the driven wheel are in a folded state or a suspended state without contacting the ground to reduce the air resistance and inertial burden in the overall flight mode; the servo drives the crawler wheel wing system to maintain a fixed angle in the flight mode, optimizes the aerodynamic performance and center of gravity distribution of the crawler wheel wing system, and ensures stability during flight; the integrated flight control and electric adjustment module monitors the robot's attitude in real time, and achieves attitude balance, position maintenance and heading control by adjusting the speed of each propeller; the main control module and the integrated flight control and electric adjustment module work together to adjust the power output strategy according to the environmental perception data to improve the adaptability and flexibility in the flight mode. The flight mode is particularly suitable for crossing complex terrain or performing high-angle reconnaissance missions, and can achieve precise flight trajectory control and mission response.

[0118] Specifically, in ground mode, the left and right track wheel wing systems independently adjust the angles through the servos to adapt to the needs of different terrains and ground conditions. The track wheel wing system is driven by the servos to adjust to a position suitable for ground driving. The ground motor drives the drive wheel through the gear set to provide the robot with forward, backward and turning movement power; the driven wheel cooperates with the track to provide stable support, enhancing the robot's stability and obstacle crossing ability in ground mode; the track is arranged around the drive wheel and the driven wheel, and the continuous grounding of the track enables the robot to pass through different complex terrains (such as sand, mud, and gravel roads); the main control module adjusts the speed and direction of the drive wheel in real time, and combines the path planning algorithm to achieve precise ground motion control.

[0119] Here, the left and right track wheel wing systems adjust the angles through independent high-torque servos to ensure the optimal contact area between the track and the ground, improve the adaptability to complex terrain, and maintain the overall stability of the robot. Improved flexibility. The independent control characteristics of the servos allow the left and right track wheel wing systems to adapt to local differences in different terrains, such as providing better ground support on uneven terrain. The high-torque servos of the track wheel wing module ensure that the deformation process from flight mode to ground mode is smooth and accurate. The main controller and the servos are linked to monitor the module status in real time, providing comprehensive support for movement and task execution in ground mode.

[0120] Specifically, when switching between flight mode and ground mode, the track-wheel-wing system, under the command of the main control module, drives the left and right track-wheel-wing systems to adjust their positions synchronously or sequentially through a high-torque servo (here, the two sides are independently controlled, and when switching modes, they can be adjusted synchronously like the design of common deformable structures, or independently adjusted sequentially), to achieve a smooth transition from flight mode to ground mode or from ground mode to flight mode; when switching from flight mode to ground mode, the servo drives the track-wheel-wing system to adjust to a position suitable for ground driving, and at the same time the ground motor and drive wheels enter working state to provide power support for ground travel; the tracks contact the ground, and the driven wheels provide additional stability to ensure stability and obstacle surmounting capability in ground mode; when switching from ground mode to flight mode, the servo drives the track-wheel-wing system to adjust to the flight mode position to ensure that the propeller has sufficient working space, the ground motor stops working, and the flight motor drives the propeller to rotate at high speed to provide lift and attitude control to achieve a smooth takeoff.

[0121] In one embodiment, a control method for a small track-wheel-wing composite deformable land-air amphibious reconnaissance robot is provided, wherein the control method adapts to two modes by dynamically switching the dynamic model, and the robot position control uses a multi-modal control algorithm based on model prediction; the control method comprises the following steps:

[0122] Step 1, establish the robot dynamics model and establish the state space equation;

[0123] Step 2: construct the prediction equation, define the constraints and design the cost function;

[0124] Step 3: Construct an optimization problem, solve the optimal control sequence and output the control quantity.

[0125] The main controller switches the control model according to the current mode of the robot, coordinates the constraints and control objectives, and ensures the trajectory tracking accuracy and system stability of the robot in flight mode and ground mode.

[0126] Further, in one embodiment, the step 1 of establishing a robot dynamics model and establishing a state space equation specifically includes:

[0127] Step 1-1, establish the dynamic model of the amphibious reconnaissance robot in flight mode:

[0128] Select the state variable X as:

[0129] X=[xyzv x v y v z φ θ ψ ω φ ω θ ω ψ ] T

[0130] Among them, x, y, z are the position coordinates in the geodetic coordinate system, v x ,v y ,v z is the linear velocity component in the geodetic coordinate system, φ, θ, ψ are the roll angle, pitch angle, and yaw angle, ω φ ,ω θ ,ω ψ is the angular velocity component in the robot coordinate system;

[0131] The output is:

[0132] Y=[xyz φ θ ψ T

[0133] The input is:

[0134] U=[ω1 ω2 ω3 ω4] T

[0135] Among them, ω i is the speed of the i-th flight motor; i=1,2,3,4;

[0136] The thrust and torque of the robot in flight mode are determined by the motor speed, and the total thrust is where k f is the thrust coefficient; the rolling moment is Where L' is the distance from the flight motor to the center of mass of the robot; the pitch moment is The yaw moment is Among them, k m is the moment coefficient;

[0137] Then the dynamic model of the amphibious reconnaissance robot in flight mode is:

[0138]

[0139] Among them, v x , v y , v zis the linear velocity component in the geodetic coordinate system, F is the total thrust, m is the robot mass, g is the gravitational acceleration, φ, θ, ψ are the roll angle, pitch angle, and yaw angle respectively, ω φ ,ω θ ,ω ψ is the angular velocity component in the robot coordinate system, J xx ,J yy ,J zz is the main diagonal element of the inertia matrix, m is the mass of the robot, g is the gravitational acceleration, τ φ ,τ θ ,τ ψ are the rolling moment, pitching moment and yaw moment;

[0140] Step 1-2, establish the discrete form state space equation of the amphibious reconnaissance robot in flight mode:

[0141] Assume the current time is k, and write the discrete state space equation according to the dynamic model in step 1-1 as:

[0142] X k+1 =A1X k +B1U k

[0143] Among them, X k+1 , X k are the state variables at time k+1 and k respectively, A1 is the state matrix, which reveals the change law of the state variable over time, and B1 is the input matrix, which reflects the response of the system to external input;

[0144] Steps 1-3, establish the kinematic model of the amphibious reconnaissance robot in ground mode:

[0145] Select the state variable X as:

[0146] X=[xyθ] T

[0147] Where x, y, θ are the position and orientation angle of the ground mode respectively;

[0148] The control input U is:

[0149] U=[v L v R ] T

[0150] Among them, v L is the left wheel speed, v R is the right wheel speed;

[0151] The kinematic model of the amphibious reconnaissance robot in ground mode is:

[0152]

[0153] Where r is the wheel radius, L is the wheel spacing in ground mode;

[0154] Steps 1-4: Establish the discrete state space equations in ground mode

[0155] Assume the current time is k, and write the discrete state space equation according to the kinematic model in the ground mode in steps 1-3 as:

[0156] X(k+1)=A2X(k)+B2U(k)

[0157] Among them, X(k+1) and X(k) are the state variables at time k+1 and k respectively, A2 is the state matrix, which reveals the change law of the state variable over time, B2 is the input matrix, which reflects the response of the system to external input, and U(k) is the control input at time k.

[0158] Further, in one of the embodiments, in step 2, a prediction equation is constructed based on the state space equation to describe the relationship between the state output and the control input in the prediction time domain; the constraints of the multimodal system are defined, including the input range and state constraints in the flight mode and the ground mode; and a cost function is designed to minimize the reference trajectory deviation and the change of the control increment; specifically including:

[0159] Step 2-1, convert the linear discrete state space model into incremental form:

[0160]

[0161] Among them, Δx(k+1) and Δx(k) are the state increments at time k+1 and k respectively, Δu(k) is the input increment at time k, y(k+1) is the system output at time k+1, A is the state matrix, B is the input matrix, and C is the output matrix;

[0162] Step 2-2, construct the prediction equation:

[0163] Set the prediction horizon of the MPC controller to l and the control horizon to c;

[0164] Assume that the control quantity outside the control time domain does not change, that is, Δu(k+i)=0,i=c,c+1,…l-1,

[0165] Then the prediction equation is obtained to predict the system state increment from time k+1 through time k+c to time k+l:

[0166] Δx(k+1|k)=AΔx(k)+BΔu(k)

[0167]

[0168] In the formula, Δx(k+1|k) represents the predicted state at the next moment k+1, Δx(k+c|k) represents the state increment of the cth step in the prediction time domain, and A c represents the state transition of the system after step c, It describes the cumulative impact of multiple future input increments on the state, Δu(k+i) is the input increment at time k+i, Δx(k+l|k) represents the state increment at step l in the prediction time domain, and A l It represents the state transition of the system after l steps. Represents the cumulative effect of all future input increments on the state;

[0169] According to the output equation, the relationship between the output of the system from time k to time k+c and the controlled output is:

[0170] y(k+1|k)=CAΔx(k)+CBΔu(k)+y(k)

[0171]

[0172] In the formula, y(k+1|k) represents the predicted output at the next moment k+1, y(k+c|k) represents the predicted output at the moment k+c, and y(k) is the system output at the moment k. Indicates the cumulative effect of the current state increment on the output after c steps; represents the impact of the input increment of the future c steps on the output, Δu(k+cj) represents the input increment at time k+cj, and y(k+l|k) represents the predicted output at time k+l. It indicates the influence of the current state increment on the output after l steps. represents the impact of the input increment of the next l steps on the output, and Δu(k+lj) represents the input increment at the k+lj moment;

[0173] Define the predicted output vector Y at the current time k and time step l l (k+1|k), define the control input sequence of the system when the control time domain is c as ΔU(k), and obtain the future output prediction equation of the system at time k:

[0174] Y l (k+1|k)=Q x Δx(k)+Q u ΔU(k)+Py(k)

[0175] Among them, Y l (k+1|k) indicates the system output vector with a prediction time domain of l starting from time k+1, Q x Indicates the impact of the initial state increment on future output, Qu It represents the cumulative impact of input increments on future outputs, and Py(k) means incorporating the current output y(k) into the prediction;

[0176] in,

[0177] Step 2-3, constraints:

[0178] In view of the multi-modal control problem, the dynamic model and constraints are switched according to the current mode of the robot. In the flight mode, the dynamic model is described by six degrees of freedom, the control input is the propeller thrust, and the system constraints include the upper and lower limits of the thrust and angular torque; in the ground mode, the dynamic model is based on the differential chassis, the control input is the speed of the drive wheel, and the system constraints include the maximum speed, the speed difference, and the contact force limit between the track and the ground;

[0179] The constraints of the system control quantity are:

[0180] u min ≤u(k+i)≤u max ,i=0,1,…,c-1

[0181] Among them, u min ,u max They represent the minimum and maximum values ​​of the control input respectively, and u(k+i) represents the control input at the i-th prediction step at time k, i.e., at time k+i;

[0182] The constraints of the system control increment are:

[0183] Δu min ≤Δu(k+i)≤Δu max ,i=0,1,…,c-1

[0184] Among them, Δu min ,Δu max They represent the minimum and maximum values ​​of the control increment respectively, and Δu(k+i) represents the control increment at the i-th prediction step at time k, i.e., at time k+i;

[0185] Step 2-4, design the cost function:

[0186] The cost function for designing position control is:

[0187] J(x(k),ΔU(k))=||M y (Y l (k+1|k)-N(k+1))|| 2 +||M u ΔU(k)|| 2

[0188] The first term is the degree to which the penalty deviates from the reference state, and the second term is the magnitude of the change in the penalty control increment;

[0189] Among them, M y Represents the weighting factor of the system output, Y l (k+1|k) represents the system prediction output, N(k+1) represents the reference trajectory, and M u represents the weighting factor of the control increment;

[0190] Write the cost function in matrix form:

[0191]

[0192] in, is the weight matrix of the system output with dimension l×l, M yi is the weight coefficient, i=1,2,...,l; is the weight matrix of the system control increment with dimension c×c, M uj is the weight coefficient, j=1,2,...,c, is the reference input sequence of the system at time k+1, and n(k+i) is the i-th reference point.

[0193] Furthermore, in one of the embodiments, in step 3, the prediction equation and the cost function are combined to construct a quadratic programming problem, and the optimal control sequence is solved by an optimization algorithm; the current control amount is output in each control cycle, and the state is updated in a rolling manner to achieve accurate tracking of the target position by the robot; step 3 specifically includes:

[0194] Step 3-1, construct the optimization problem:

[0195] For the position control tracking problem, the MPC algorithm obtains the optimal control sequence by solving an optimization problem, which meets the system constraints while minimizing the cost function; the following quadratic optimization problem is constructed:

[0196] min z T Hz-g T z

[0197] stb≤C'z

[0198] Among them, H is the Hessian matrix, g is the gradient vector, z is the optimal independent variable, C' is the constraint matrix, and b is the right-hand side vector of the constraint;

[0199] Substituting the prediction equation into the objective function, we get:

[0200] J(x(k),ΔU(k))=||M y (Y(k+1|k)-N(k+1))|| 2 +||Mu ΔU(k)|| 2

[0201] =||M y (Q x Δx(k)+Q u ΔU(k)+Py(k)-N(k+1))|| 2 +||M u ΔU(k)|| 2

[0202] Simplify the expression, let M y (N(k+1)-Py(k)-Q x Δx(k))=T(k+1), rewrite the above formula as:

[0203] J(x(k),ΔU(k))=ΔU(k) T HΔU(k)-G(k+1|k) T ΔU(k)

[0204] Among them, the Hessian matrix Gradient Vector

[0205] The matrix form of control increment constraint and control quantity constraint is:

[0206]

[0207] Among them, R1 and R2 both represent diagonal block matrices, I 4×4 represents the unit diagonal matrix; Δu min (*), Δu max (*) respectively represent the minimum and maximum values ​​of the control input at the time “*”, u min (*),u max (*) indicates the minimum and maximum values ​​of the control input at the time “*”;

[0208] Step 3-2, solve the optimal control sequence and output:

[0209] In each control cycle, the interior point method is used to obtain the optimal control sequence ΔU in the control time domain c. * (k)=[Δu * (k),Δu * (k+1),…,Δu * (k+cl)] T ; where Δu * (k+i) represents the optimal control increment at time k+i, i=0,1,....,cl;

[0210] According to the basic principle of MPC, the first step of the optimal control sequence is applied to the controlled system to obtain a new system output. The control amount of the current control cycle is determined based on the control amount input to the controlled system in the previous control cycle, that is, u(k) = u(k-1) + Δu * (k), where u(k) is the control output at the current moment k, u(k-1) is the control output at the previous moment k-1, Δu * (k) is the optimal control increment at the current moment; and in the next time step, a new quadratic programming problem is solved with the new system state. After the system state is updated, the quadratic programming problem is solved again, and the rolling optimization is performed until the reference position control is completed.

[0211] In the above control method, the control models of flight mode and ground mode are modeled in a unified form, and the same control method is used to uniformly control ground mode and air mode, making the switching smoother. The two modes use the same algorithm to save computing resources, such as program storage space and runtime memory usage.

[0212] The present invention solves the key problems in the prior art by comprehensively optimizing the structure and control method, and provides new technical support and solutions for tasks such as disaster relief, environmental reconnaissance and security patrols. It has broad application prospects and important technical value.

[0213] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A small track-wheel-wing composite deformable land and air amphibious reconnaissance robot, characterized in that: The robot comprises a control cabin, a left track wheel wing system and a right track wheel wing system, wherein the left track wheel wing system and the right track wheel wing system are symmetrically installed on both sides of the control cabin, respectively, and are connected to the control cabin through a steering gear; the left track wheel wing system and the right track wheel wing system are both integrated with a flight motor, a propeller, a ground motor, a driving wheel, a driven wheel and a track, and the switching between the flight mode and the ground mode is realized under the drive of the steering gear; in the flight mode, the flight motors of the left track wheel wing system and the right track wheel wing system drive the propeller to rotate at a high speed to provide lift, so as to realize aerial flight and posture stability; in the ground mode, the left track wheel wing system and the right track wheel wing system are rotated to the ground contact position through the steering gear, the propeller stops working, and the ground motor is driven by a gear set to realize ground driving.

2. The small track-wheel-wing composite deformable land-air amphibious reconnaissance robot according to claim 1 is characterized in that: The control cabin includes: a main control module, an integrated flight control and electric adjustment module, a GPS module, a camera and an image transmission module; the main control module is responsible for managing the operation logic of the whole machine, deformation control of the track wheel wing system, ground motion control and mode switching; the integrated flight control and electric adjustment module is responsible for the attitude stability and power management of the robot in flight mode, and achieves high-precision control during flight by directly controlling the output of the flight motor; the GPS module is used to obtain location information in real time to provide support for navigation and path planning; the image transmission module is connected to the camera to transmit real-time images to the remote monitoring terminal for environmental reconnaissance and mission command.

3. The small track-wheel-wing composite deformable land-air amphibious reconnaissance robot according to claim 2 is characterized in that: The left track wheel wing system and the right track wheel wing system both adopt an integrated design. The flight motor is directly connected to the propeller, and the propeller is installed concentrically with the drive wheel to provide lift in the flight mode to achieve aerial flight and attitude control; the ground motor is connected to the drive wheel through a gear set, driving the drive wheel to provide travel power in the ground mode; the track is arranged around the drive wheel and the driven wheel, and the driven wheel provides tension and support for the track; the left track wheel wing system and the right track wheel wing system are driven by a large-torque servo to rotate around a fixed axis, thereby achieving morphological switching between the flight mode and the ground mode; the main controller coordinates the servo action in real time to ensure the smoothness and accuracy of the deformation process.

4. The small track-wheel-wing composite deformable land-air amphibious reconnaissance robot according to claim 2 is characterized in that: In flight mode, the flight motor of the track-wheel-wing system drives the propeller to rotate at high speed to provide thrust, enabling the robot to fly stably. At this time, the driving wheels and driven wheels are in a folded state or a suspended state without contacting the ground, so as to reduce the air resistance and inertial burden in the overall flight mode; the servo drives the track-wheel-wing system to maintain a fixed angle in the flight mode, optimizing the aerodynamic performance and center of gravity distribution of the track-wheel-wing system; the integrated flight control and electric adjustment module monitors the robot's posture in real time, and achieves posture balance, position maintenance and heading control by adjusting the speed of each propeller; the main control module works together with the integrated flight control and electric adjustment module to adjust the power output strategy according to environmental perception data.

5. The small track-wheel-wing composite deformable land-air amphibious reconnaissance robot according to claim 2 is characterized in that: In ground mode, the left and right track wheel wing systems independently adjust the angles through the servos to adapt to the needs of different terrains and ground conditions. The track wheel wing system is driven by the servos to adjust to a position suitable for ground driving. The ground motor drives the drive wheel through the gear set to provide the robot with moving power for forward, backward and turning; the driven wheel cooperates with the track to provide stable support; the track is arranged around the drive wheel and the driven wheel, and the robot can pass through different complex terrains through the continuous grounding of the track; the main control module adjusts the speed and direction of the drive wheel in real time, and combines the path planning algorithm to achieve precise ground motion control.

6. The small track-wheel-wing composite deformable land-air amphibious reconnaissance robot according to claim 2 is characterized in that: When switching between flight mode and ground mode, the track-wheel-wing system, under the command of the main control module, drives the left and right track-wheel-wing systems to adjust their positions synchronously or sequentially through the high-torque servo, so as to achieve a smooth transition from flight mode to ground mode or from ground mode to flight mode; when switching from flight mode to ground mode, the servo drives the track-wheel-wing system to adjust to a position suitable for ground driving, and at the same time the ground motor and drive wheels enter working state to provide power support for ground travel; the tracks contact the ground, and the driven wheels provide additional stability to ensure stability and obstacle surmounting capability in ground mode; when switching from ground mode to flight mode, the servo drives the track-wheel-wing system to adjust to the flight mode position to ensure that the propeller has sufficient working space, the ground motor stops working, and the flight motor drives the propeller to rotate at high speed to provide lift and attitude control to achieve a smooth takeoff.

7. A control method for a small track-wheel-wing composite deformable land-air amphibious reconnaissance robot according to any one of claims 1 to 6, characterized in that: The control method adapts to the two modes by dynamically switching the dynamic model, and the robot position control uses a multi-modal control algorithm based on model prediction; the control method includes the following steps: Step 1, establish the robot dynamics model and establish the state space equation; Step 2: construct the prediction equation, define the constraints and design the cost function; Step 3: Construct an optimization problem, solve the optimal control sequence and output the control quantity.

8. The control method according to claim 7, characterized in that: Step 1 establishes the robot dynamics model and the state space equation, specifically including: Step 1-1, establish the dynamic model of the amphibious reconnaissance robot in flight mode: Select the state variable X as: X=[xyzv x v y v z f t s y φ oh θ oh ψ T Among them, x, y, z are the position coordinates in the geodetic coordinate system, v x ,v y ,v z is the linear velocity component in the geodetic coordinate system, φ, θ, ψ are the roll angle, pitch angle, and yaw angle, ω φ ,ω θ ,ω ψ is the angular velocity component in the robot coordinate system; The output is: Y=[xyz φ θ ψ] T The input is: U=[ω1 ω2 ω3 ω4] T Among them, ω i is the speed of the i-th flight motor; i=1,2,3,4; The thrust and torque of the robot in flight mode are determined by the motor speed, and the total thrust is where k f is the thrust coefficient; the rolling moment is Where L' is the distance from the flight motor to the center of mass of the robot; the pitch moment is The yaw moment is Among them, k m is the moment coefficient; Then the dynamic model of the amphibious reconnaissance robot in flight mode is: Among them, v x , v y , v z is the linear velocity component in the geodetic coordinate system, F is the total thrust, m is the robot mass, g is the gravitational acceleration, φ, θ, ψ are the roll angle, pitch angle, and yaw angle respectively, ω φ ,ω θ ,ω ψ is the angular velocity component in the robot coordinate system, J xx ,J yy ,J zz is the main diagonal element of the inertia matrix, m is the mass of the robot, g is the gravitational acceleration, τ φ ,τ θ ,τ ψ are the rolling moment, pitching moment and yaw moment; Step 1-2, establish the discrete form state space equation of the amphibious reconnaissance robot in flight mode: Assume the current time is k, and write the discrete state space equation according to the dynamic model in step 1-1 as: X k+1 =A1X k +B1U k Among them, X k+1 , X k are the state variables at time k+1 and k respectively, A1 is the state matrix, which reveals the change law of the state variable over time, and B1 is the input matrix, which reflects the response of the system to external input; Steps 1-3, establish the kinematic model of the amphibious reconnaissance robot in ground mode: Select the state variable X as: X=[xy θ] T Where x, y, θ are the position and orientation angle of the ground mode respectively; The control input U is: U=[v L v R ] T Among them, v L is the left wheel speed, v R is the right wheel speed; The kinematic model of the amphibious reconnaissance robot in ground mode is: Where r is the wheel radius, L is the wheel spacing in ground mode; Steps 1-4: Establish the discrete state space equations in ground mode Assume the current time is k, and write the discrete state space equation according to the kinematic model in the ground mode in steps 1-3: X(k+1)=A2X(k)+B2U(k) Among them, X(k+1) and X(k) are the state variables at time k+1 and k respectively, A2 is the state matrix, which reveals the change law of the state variable over time, B2 is the input matrix, which reflects the response of the system to external input, and U(k) is the control input at time k.

9. The control method according to claim 7, characterized in that: In step 2, a prediction equation is constructed based on the state space equation to describe the relationship between the state output and the control input in the prediction time domain; the constraints of the multimodal system are defined, including the input range and state constraints in the flight mode and the ground mode; and the cost function is designed to minimize the reference trajectory deviation and the change of the control increment. Specifically, it includes: Step 2-1, convert the linear discrete state space model into incremental form: Among them, Δx(k+1) and Δx(k) are the state increments at time k+1 and k respectively, Δu(k) is the input increment at time k, y(k+1) is the system output at time k+1, A is the state matrix, B is the input matrix, and C is the output matrix; Step 2-2, construct the prediction equation: Set the prediction horizon of the MPC controller to l and the control horizon to c; Assume that the control quantity outside the control time domain does not change, that is, Δu(k+i)=0,i=c,c+1,…l-1, Then the prediction equation is obtained to predict the system state increment from time k+1 through time k+c to time k+l: Δx(k+1|k)=AΔx(k)+BΔu(k) In the formula, Δx(k+1|k) represents the predicted state at the next moment k+1, Δx(k+c|k) represents the state increment of the cth step in the prediction time domain, and A c represents the state transition of the system after step c, It describes the cumulative impact of multiple future input increments on the state, Δu(k+i) is the input increment at time k+i, Δx(k+l|k) represents the state increment at step l in the prediction time domain, and A l It represents the state transition of the system after l steps. Represents the cumulative effect of all future input increments on the state; According to the output equation, the relationship between the output of the system from time k to time k+c and the controlled output is: y(k+1|k)=CAΔx(k)+CBΔu(k)+y(k) In the formula, y(k+1|k) represents the predicted output at the next moment k+1, y(k+c|k) represents the predicted output at the moment k+c, and y(k) is the system output at the moment k. Indicates the cumulative effect of the current state increment on the output after c steps; represents the impact of the input increment of the future c steps on the output, Δu(k+cj) represents the input increment at time k+cj, and y(k+l|k) represents the predicted output at time k+l. It indicates the influence of the current state increment on the output after l steps. represents the impact of the input increment of the next l steps on the output, and Δu(k+lj) represents the input increment at the k+lj moment; Define the predicted output vector Y at the current time k and time step l l (k+1|k), define the control input sequence of the system when the control time domain is c as ΔU(k), and obtain the future output prediction equation of the system at time k: Y l (k+1|k)=Q x Δx(k)+Q u ΔU(k)+Py(k) Among them, Y l (k+1|k) indicates the system output vector with a prediction time domain of l starting from time k+1, Q x Indicates the impact of the initial state increment on future output, Q u It represents the cumulative impact of input increments on future outputs, and Py(k) means incorporating the current output y(k) into the prediction; in, Step 2-3, constraints: In view of the multi-modal control problem, the dynamic model and constraints are switched according to the current mode of the robot. In the flight mode, the dynamic model is described by six degrees of freedom, the control input is the propeller thrust, and the system constraints include the upper and lower limits of the thrust and angular torque; in the ground mode, the dynamic model is based on the differential chassis, the control input is the speed of the drive wheel, and the system constraints include the maximum speed, the speed difference, and the contact force limit between the track and the ground; The constraints of the system control quantity are: u min ≤u(k+i)≤u max ,i=0,1,…,c-1 Among them, u min ,u max They represent the minimum and maximum values ​​of the control input respectively, and u(k+i) represents the control input at the i-th prediction step at time k, i.e., at time k+i; The constraints of the system control increment are: Δu min ≤Δu(k+i)≤Δu max ,i=0,1,…,c-1 Among them, Δu min ,Δu max They represent the minimum and maximum values ​​of the control increment respectively, and Δu(k+i) represents the control increment at the i-th prediction step at time k, i.e., at time k+i; Step 2-4, design the cost function: The cost function for designing position control is: J(x(k),ΔU(k))=||M y (Y l (k+1|k)-N(k+1))|| 2 +||M u ΔU(k)|| 2 Among them, M y Represents the weighting factor of the system output, Y l (k+1|k) represents the system prediction output, N(k+1) represents the reference trajectory, and M u represents the weighting factor of the control increment; Write the cost function in matrix form: in, is the weight matrix of the system output with dimension l×l, M yi is the weight coefficient, i=1,2,...,l; is the weight matrix of the system control increment with dimension c×c, M uj is the weight coefficient, j=1,2,...,c, is the reference input sequence of the system at time k+1, and n(k+i) is the i-th reference point.

10. The control method according to claim 7, characterized in that: In step 3, the prediction equation and the cost function are combined to construct a quadratic programming problem, and the optimal control sequence is solved by the optimization algorithm; the current control amount is output in each control cycle, and the state is updated in a rolling manner to achieve accurate tracking of the target position by the robot; step 3 specifically includes: Step 3-1, construct the optimization problem: For the position control tracking problem, the MPC algorithm obtains the optimal control sequence by solving an optimization problem, which meets the system constraints while minimizing the cost function; the following quadratic optimization problem is constructed: stb≤C'z Among them, H is the Hessian matrix, g is the gradient vector, z is the optimal independent variable, C' is the constraint matrix, and b is the right-hand side vector of the constraint; Substituting the prediction equation into the objective function, we get: J(x(k),ΔU(k))=||M y (Y(k+1|k)-N(k+1))|| 2 +||M u ΔU(k)|| 2 =||M y (Q x Δx(k)+Q u ΔU(k)+Py(k)-N(k+1))|| 2 +||M u ΔU(k)|| 2 Simplify the expression, let M y (N(k+1)-Py(k)-Q x Δx(k))=T(k+1), rewrite the above formula as: Among them, the Hessian matrix Gradient Vector The matrix form of control increment constraint and control quantity constraint is: Among them, R1 and R2 both represent diagonal block matrices, I 4×4 represents the unit diagonal matrix; Δu min (*), Δu max (*) respectively represent the minimum and maximum values ​​of the control input at the time "*", u min (*),u max (*) respectively represent the minimum and maximum values ​​of the control input at the time "*"; Step 3-2, solve the optimal control sequence and output: In each control cycle, the interior point method is used to obtain the optimal control sequence in the control time domain c. Among them, Δu * (k+i) represents the optimal control increment at time k+i, i=0,1,....,cl; According to the basic principle of MPC, the first step of the optimal control sequence is applied to the controlled system to obtain a new system output. The control amount of the current control cycle is determined based on the control amount input to the controlled system in the previous control cycle, that is, u(k) = u(k-1) + Δu * (k), where u(k) is the control output at the current moment k, u(k-1) is the control output at the previous moment k-1, Δu * (k) is the optimal control increment at the current moment; and in the next time step, a new quadratic programming problem is solved with the new system state. After the system state is updated, the quadratic programming problem is solved again, and the rolling optimization is performed until the reference position control is completed.

Images