A switching control method for temporarily lost target UAV with stability guarantee

Abstract

The invention discloses a switching control method for a temporarily lost target unmanned aerial vehicle with stability guarantee in the technical field of unmanned aerial vehicle automatic control. A switching system is used to model the temporary loss phenomenon of a target. Based on the low computing power controller design idea of ​​state feedback, the controller gain is adjusted in real time according to the actual tracking error to ensure the stability of the established switching tracking system. The gain adjustment calculation process is performed in the unmanned aerial vehicle onboard computing unit. The unmanned aerial vehicle adopts a dual-loop control framework of a position loop and an attitude loop. The established switching system is divided into an unmeasurable subsystem and a measurable subsystem according to whether the target is measurable. The unmanned aerial vehicle target tracking control can be realized in the case of temporary loss of the target. The controller gain is automatically adjusted during the tracking process to ensure the stability of the tracking system and improve the tracking performance in the case of temporary loss of the target.

Description

Technical Field

[0001] The invention relates to the technical field of automatic control of unmanned aerial vehicles, and in particular to a switching control method for a temporarily lost target unmanned aerial vehicle with guaranteed stability. Background Art

[0002] In recent years, drones have been widely used in military and civilian fields due to their flexibility, concealment, and economy. UAV target tracking refers to the process of a drone approaching a designated target and maintaining synchronous movement. It is an important control technology that is essential for drones to perform autonomous tasks such as aerial photography, vehicle collaborative operations, surveillance and tracking, and even target strikes. Usually, drones use airborne sensors such as visual cameras, laser radars, and ultrasonic rangefinders to capture and identify the tracked target using corresponding perception algorithms, thereby solving and obtaining reference signals for target tracking. However, in actual applications, due to the dynamic and changeable perception environment, the tracked target is sometimes unable to be captured or identified. For example, the image sensor is blocked by obstacles and interfered by natural noise, the laser radar has insufficient lines, resulting in insufficient available identification points on the tracked target, and the ultrasonic rangefinder has a small three-dimensional visual range, which makes it easy for the target to escape from the field of view. This results in the inability to obtain reference signals for drone target tracking at some times.

[0003] In the problem of UAV target tracking control, most methods have not yet considered the tracking control of UAVs in the case of lost tracked targets. Once the target is lost, the UAV will hover at the position at the time of target loss until the target is captured again. Obviously, this will greatly affect the tracking performance, and as the target loss time increases, this hovering scheme will lead to a smaller and smaller probability of target capture. Some methods focus on the position prediction after the target is lost to provide an estimated position to solve the reference signal of UAV tracking control, but this method is only applicable to some scenarios where the target loss time is short and the motion state change is less random. There are also methods that install a gimbal for the sensor so that the target can be locked, thereby reducing the probability of target loss, but this method is limited by the quality of the gimbal and the sensor and is difficult to apply to UAVs of all sizes. In recent years, there are also methods based on model predictive control, which plan the position and attitude of the UAV while predicting the target position, but the computing power required is too high. Therefore, there is an urgent need for a low-computing tracking control scheme suitable for UAVs in the case of temporary target loss.

[0004] In view of the aforementioned low computing power requirements, some methods are based on state feedback and use the switching system framework to model the target tracking problem of intelligent agents such as unmanned vehicles with temporary target loss. According to whether the target is lost, a measurable subsystem and an unmeasurable subsystem are established respectively, and the tracking problem of the temporarily lost target is converted into the stabilization problem of the modeled switching system. The stability of the system is guaranteed by determining the maximum target unmeasurable time and the minimum target measurable time. However, the above research is based on the pre-design of the state feedback controller of each subsystem. The derived upper and lower bound conditions of the subsystem time can only be used for the stability analysis of the system, and the obtained stability conditions have not been reasonably applied to the controller design. In addition, most of the existing methods based on the switching system framework are for full-drive control systems such as unmanned vehicles, and cannot be directly applied to under-actuated systems such as drones. For common drone control schemes, the target loss phenomenon has not been considered. The low-cost control method based on switching system modeling is only used in stability analysis. The research objects of the existing methods are mostly full-drive systems. The present invention is based on the switching system framework and proposes a switching control method for drone under-actuated systems with stability guarantee and low computational cost to solve the above problems. Summary of the invention

[0005] The object of the present invention is to provide a switching control method for a temporarily lost target UAV with stability guarantee, so as to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a switching control method for a temporarily lost target drone with stability guarantee, comprising the following steps:

[0007] S1: Initialization and establishment of tracking coordinate system;

[0008] S2: UAV position control when the target is detectable;

[0009] S3: Gain determination when the target is unmeasurable;

[0010] S4: Position control when the target is undetectable;

[0011] S5: controller gain update during tracking; It should be noted that the above steps are all implemented in the airborne computing unit, the target measurability is determined by the sensors and measurement algorithms loaded on the UAV, and all other calculations are numerical operations, which ensures the tracking control of the UAV with a small amount of calculation in the event of temporary loss of the target.

[0012] Preferably, step S1 is specifically as follows: the UAV receives information provided by a ground operator, including the initial position and speed of the target, the maximum speed of the target, the relative position that the UAV needs to maintain with the target, a given minimum measurable time and a maximum unmeasurable time, and a preset position controller gain of a measurable subsystem; constructs a body coordinate system and a tracking coordinate system; and adjusts the attitude of the UAV so that the target is within the field of view of the sensor, so that the target position is measurable at the initial moment and within the subsequent minimum measurable time, to ensure that there is sufficient time to complete the determination of the controller gain of the adjacent unmeasurable subsystem.

[0013] Preferably, the step S2 is specifically as follows: according to the reference signal obtained by solving the step S1, the current tracking error of the UAV is calculated in the tracking coordinate system, thereby calculating and recording the tracking system energy; the current position tracking error of the UAV is brought into the state feedback controller of the position loop to obtain the required control input and the expected value of the attitude loop, and then the control input of the attitude loop is obtained; finally, the position control of the UAV is realized when the target is measurable, that is, the measurable subsystem control; wherein the UAV adopts a dual-loop control scheme of the position loop and the attitude loop, and the UAV dynamics model is

[0014]

[0015] Among them, u1~u4 are control inputs, which can be realized by adjusting the speed of the drone motor; x, y, z, θ, φ, ψ are the horizontal position of the corresponding coordinate axis and the attitude angle around the coordinate axis, and their reference signals are x, y, z, θ, φ, and ψ. d ,y d 、z d ,θ d ,φ d , d It indicates that in the controller design and stability analysis stage, according to the small angle theorem, the states z, θ, φ, ψ can be directly controlled by u1~u4, and the states x and y need to be controlled by changing the size of θ and φ; the dual-loop control scheme can be used to divide the control loop of the UAV into an inner attitude loop and an outer position loop, and the inner loop attitude desired angle is determined by the outer loop position error as the virtual control input;

[0016] For the inner loop attitude controller, in order to ensure its rapid response capability, the finite time sliding mode control method is adopted; taking the attitude θ as an example, the sliding mode surface is designed as where e θ =θ-θ d , α, β, p, q are all positive real numbers and p<q, the time derivative of the sliding surface can be calculated as

[0017]

[0018] According to the sliding condition The equivalent controller can be designed as

[0019]

[0020] Combined with the reaching law commonly used in sliding mode control, the attitude controller based on the finite time sliding mode control method is finally designed as follows:

[0021] u2=u θ,eq -Ksgn(s θ ) (4)

[0022] Where K is a positive real number, and the design of attitude controllers u3 and u4 is the same as u2;

[0023] In terms of position control, proportional-differential control is used to design the controller. Specifically, the height controller is first designed as Where: k vz , k pz is a positive real number. When the altitude reaches a steady state, u1 is always 1. At this point, the design of the four control inputs of the drone is completed. For the underactuated states x and y, the proportional-differential control idea is used to design the virtual control inputs.

[0024]

[0025] where k vx , k px , k vy , k py are all positive real numbers, thus indirectly realizing the position control of the UAV by adjusting its attitude.

[0026] Preferably, the step S3 is specifically as follows: according to the target measurable subsystem energy initial value recorded in step S2, the system energy of the adjacent target unmeasurable subsystem is ensured to be less than the current measurable subsystem energy initial value at the end of operation, that is, the stability of each switching stage is ensured, the controller gain of the unmeasurable subsystem is calculated, and the controller gain that can ensure that the drone has a large acceleration is selected and recorded for backup; it is determined whether the running time of the measurable subsystem exceeds a given minimum measurable time. Once the given minimum measurable time is not reached, it is considered that the target measurability during the drone tracking process is extremely poor, and human intervention control is required;

[0027] When the target is unmeasurable, the predictor is used to predict the target state, and the predicted reference signal is obtained according to the predicted target position, which is recorded as At this time, formula (5) will be transformed into

[0028]

[0029] Define the systematic error as e xx d [xx d v x -vxd yy d v y -v yd ] T According to (5) and (6), the closed-loop control system of the UAV in the x and y directions can be established based on the switching system framework:

[0030]

[0031] Where A and A′ are composed of the controller gains in (5) and (6) respectively, and M is composed of the prediction error and the controller gain, which is the cause of system instability. The energy function of the tracking system is constructed as V = 0.5e T e. The energy functions of the measurable subsystem and the unmeasurable subsystem are represented by V(t) and V′(t), respectively, and we have Determine the upper bound of M, then obtain the error coefficient δ so that e T M≤δ||e||, then the tracking system energy function satisfies the following inequality

[0032]

[0033] Among them, λ and λ′ are the maximum eigenvalues ​​of the system matrices A and A′ respectively. According to the stability criterion of the switching system, an adjacent measurable subsystem and an unmeasurable subsystem are defined as a switching stage. The condition for the energy drop of the switching tracking system in a switching stage is solved as follows:

[0034]

[0035] It can be seen that when the controller gain satisfies (9), the system energy decreases in the current switching stage; the initial value of the switching stage energy V(0) is determined by step S2. It can be seen that for any λ, λ′, as the switching stage continues to advance, the initial value of the switching stage will be equal to the final value; then, the switching system will be ultimately bounded, and its final bound can be determined by the following formula:

[0036]

[0037] Obviously, as long as the initial value of the energy in the switching stage is greater than the final bound of the system, the convergence of the switching system in the current switching stage can be guaranteed, that is,

[0038]

[0039] The convergence of the dynamic process of the system can be guaranteed. Note that the right side of the equation can also be multiplied by an attenuation coefficient to make the convergence faster. In this step, the maximum controller gain that satisfies (11) is selected to ensure that when the target is far away, that is, when the initial energy value is large, the UAV has a large acceleration, so that it can quickly approach the target.

[0040] Preferably, in step S4, once the target is unmeasurable, the position and speed in the unmeasurable time period are predicted according to the motion state of the target in the previous measurable time period, a reference signal is generated according to the predicted position and speed, and the controller gain determined in step S3 is used to perform the state feedback-based UAV position control consistent with step S2, and it is determined whether the running time of the unmeasurable subsystem reaches a given maximum unmeasurable time. Once the given maximum unmeasurable time is exceeded, it is regarded that the target measurability during the UAV tracking process is extremely poor, and human intervention control is required.

[0041] Preferably, in step S5, when the target is measurable again, the workflow in steps S2, S3 and S4 is repeated; since the initial energy value of the target measurable subsystem is constantly changing as the target tracking proceeds, the controller gain of the unmeasurable subsystem will be continuously updated as the target tracking proceeds, and ultimately the stability of each switching stage can be guaranteed until the energy of the switching system reaches the final energy upper bound.

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention has the following advantages in the tracking process when the target is temporarily lost: (1) The UAV approaches the target quickly. Under the premise of ensuring the stability of the tracking system, a larger controller gain is selected, so that the UAV has a large acceleration, ensuring the rapid convergence of the dynamic process of the tracking system; (2) The steady-state error of the tracking system is small. Under the switching system framework, the switching controller designed with stability guarantee can ensure that the initial energy value of each switching stage of the system decreases until the steady-state error of the tracking system is minimized; (3) The computational complexity is low. During the entire UAV temporarily lost target tracking process, the designed switching controller is based on state feedback, and the designed parameter solution method only requires simple numerical calculations. Compared with trajectory planning methods based on model predictive control, the computing power saving effect is obvious. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0044] Figure 1 This is a UAV operation framework diagram of the temporarily lost target tracking process of the present invention;

[0045] Figure 2 It is the dual-loop control framework of the drone of the present invention;

[0046] Figure 3 A schematic diagram of the final boundedness of the energy of the switching system of the present invention;

[0047] Figure 4 The state response diagram of the TSMC controller designed for the present invention;

[0048] Figure 5 The error diagram of the TSMC controller designed by the present invention tracking the transition reference signal;

[0049] Figure 6 A switching signal diagram describing the measurable and unmeasurable situations in the comparative experiment of the present invention;

[0050] Figure 7 A comparison diagram of the switching controller with stability guarantee designed by the present invention and the traditional controller;

[0051] Figure 8 A trajectory diagram of a UAV using a switching controller designed for application of the present invention to track a temporarily lost target. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] In the description of the present invention, it is necessary to understand that the terms "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0054] The present invention proposes a method for switching control of a temporarily lost target UAV with stability guarantee. Figure 1 The operation process shown in the figure mainly includes five steps: initialization and establishment of tracking coordinate system, drone position control when the target is measurable, gain determination when the target is unmeasurable, position control when the target is unmeasurable, and controller gain update during tracking. The tracking control method for executing the above steps can be implemented based on the drone onboard computing unit, and the calculated controller input is implemented by adjusting the drone motor speed. The specific implementation methods of each step in the scheme are as follows:

[0055] S1: Initialization and establishment of tracking coordinate system; specifically: the drone receives the given information of the target's initial position / speed and maximum speed, minimum measurable time τ, and maximum unmeasurable time T, and adjusts its own posture to make the target enter the sensor's field of view; during the operation of the drone, its body coordinate system F b The center of mass of the drone is fixed to itself, and the origin of the body coordinate system is defined as the left, front, and top directions of the drone are defined as the x, y, and z axes, respectively. The tracking coordinate system in the tracking process is established according to the origin and direction of the drone body coordinate system at the initial moment, that is, the tracking coordinate system F t The origin is at the initial position of the drone's center of mass, and the directions of the coordinate axes correspond to the left, front, and top directions of the drone at the initial moment.

[0056] S2: UAV position control when the target is measurable; obtain the reference signal according to step S1, calculate the current tracking error of the UAV in the tracking coordinate system, and then calculate and record the tracking system energy; bring the current position tracking error of the UAV into the state feedback controller of the position loop, obtain the required control input and the expected value of the attitude loop, and then obtain the control input of the attitude loop; finally, realize the position control of the UAV when the target is measurable, that is, the control of the measurable subsystem; the UAV adopts a dual-loop control scheme of the position loop and the attitude loop, such as Figure 2 As shown in Figure 2, the UAV dynamics model is

[0057]

[0058] Among them, u1~u4 are control inputs, which can be realized by adjusting the speed of the drone motor; x, y, z, θ, φ, ψ are the horizontal position of the corresponding coordinate axis and the attitude angle around the coordinate axis, and their reference signals are x, y, z, θ, φ, and ψ. d ,y d 、z d ,θ d ,φ d , d It indicates that in the controller design and stability analysis stage, according to the small angle theorem, the states z, θ, φ, ψ can be directly controlled by u1~u4, and the states x and y need to be controlled by changing the size of θ and φ; the dual-loop control scheme can be used to divide the control loop of the UAV into an inner attitude loop and an outer position loop, and the inner loop attitude desired angle is determined by the outer loop position error as the virtual control input;

[0059] For the inner loop attitude controller, in order to ensure its rapid response capability, the finite time sliding mode control method is adopted; taking the attitude θ as an example, the sliding mode surface is designed as where e θ =θ-θ d , α, β, p, q are all positive real numbers and p<q, the time derivative of the sliding surface can be calculated as

[0060]

[0061] According to the sliding condition The equivalent controller can be designed as

[0062]

[0063] Combined with the reaching law commonly used in sliding mode control, the attitude controller based on the finite time sliding mode control method is finally designed as follows:

[0064] u2=u θ,eq -Ksgn(s θ ) (4)

[0065] Where K is a positive real number, and the design of attitude controllers u3 and u4 is the same as u2;

[0066] In terms of position control, proportional-differential control is used to design the controller. Specifically, the height controller is first designed as Where: k vz , k pz is a positive real number. When the altitude reaches a steady state, u1 is always 1. At this point, the design of the four control inputs of the drone is completed. For the underactuated states x and y, the proportional-differential control idea is used to design the virtual control inputs.

[0067]

[0068] where k vx , k px , k vy , k py are all positive real numbers, thus indirectly realizing the position control of the UAV by adjusting its attitude.

[0069] S3: Determine the controller gain when the target is unmeasurable; according to the target measurable subsystem energy initial value recorded in step S2, the controller gain of the unmeasurable subsystem is calculated to ensure that the system energy of the adjacent target unmeasurable subsystem is less than the current measurable subsystem energy initial value at the end of operation, that is, to ensure the stability of each switching stage, and select the controller gain that can ensure that the drone has a large acceleration to record for backup; determine whether the measurable subsystem operation time exceeds the given minimum measurable time. Once the given minimum measurable time is not reached, it is considered that the target measurability during the drone tracking process is extremely poor, and human intervention control is required;

[0070] When the target is unmeasurable, the predictor is used to predict the target state, and the predicted reference signal is obtained according to the predicted target position, which is recorded as At this time, formula (5) will be transformed into

[0071]

[0072] Define the systematic error as e xx d [xx d v x -v xd yy d v y -v yd ] T According to (5) and (6), the closed-loop control system of the UAV in the x and y directions can be established based on the switching system framework:

[0073]

[0074] Where A and A′ are composed of the controller gains in (5) and (6) respectively, and M is composed of the prediction error and the controller gain, which is the cause of system instability. The energy function of the tracking system is constructed as V = 0.5e T e. The energy functions of the measurable subsystem and the unmeasurable subsystem are represented by V(t) and V′(t), respectively, and we have Determine the upper bound of M, then obtain the error coefficient δ so that e T M≤δ||e||, then the tracking system energy function satisfies the following inequality

[0075]

[0076] Among them, λ and λ′ are the maximum eigenvalues ​​of the system matrices A and A′ respectively. According to the stability criterion of the switching system, an adjacent measurable subsystem and an unmeasurable subsystem are defined as a switching stage. The condition for the energy drop of the switching tracking system in a switching stage is solved as follows:

[0077]

[0078] It can be seen that when the controller gain satisfies (9), the system energy decreases in the current switching stage; the initial value of the switching stage energy V(0) is determined by step S2. It can be seen that for any λ, λ′, as the switching stage continues to advance, the initial value of the switching stage will be equal to the final value; then, the switching system will be ultimately bounded, and its final bound can be determined by the following formula:

[0079]

[0080] Obviously, as long as the initial value of the energy in the switching stage is greater than the final bound of the system, the convergence of the switching system in the current switching stage can be guaranteed, that is,

[0081]

[0082] The convergence of the dynamic process of the system can be guaranteed. Note that the right side of the equation can also be multiplied by an attenuation coefficient to make the convergence faster. In this step, the maximum controller gain that satisfies (11) is selected to ensure that when the target is far away, that is, when the initial energy value is large, the UAV has a large acceleration, so that it can quickly approach the target.

[0083] S4: Position control when the target is unmeasurable; once the target is unmeasurable, the position and speed of the target in the unmeasurable time period are predicted according to the motion state of the target in the previous measurable time period, and a reference signal is generated according to the predicted position and speed. The controller gain determined in step S3 is used to perform the drone position control based on state feedback consistent with step S2, and it is judged whether the running time of the unmeasurable subsystem reaches the given maximum unmeasurable time. Once the given maximum unmeasurable time is exceeded, it is regarded that the target measurability during the drone tracking process is extremely poor, and human intervention control is required. In this scheme, the predictor used is a first-order retainer, that is, it is assumed that the target maintains its speed at the time of loss, and the generated predicted position is tracked. The virtual control input used is shown in formula (6), and its controller gain is determined by formula (11); in this step, it is judged whether the running time of the unmeasurable subsystem exceeds the given maximum unmeasurable time. Once the target is unmeasurable for too long, it is regarded that the visual condition is extremely poor at this time. The drone will feedback the information of poor measurable condition and wait for the next instruction.

[0084] S5: controller gain update during tracking; when the target is measurable again, repeat the workflow in steps S2, S3 and S4; since the initial energy value of the target measurable subsystem changes continuously with the progress of target tracking, the controller gain of the unmeasurable subsystem will be continuously updated with the progress of target tracking, and finally the stability of each switching stage can be guaranteed until the energy of the switching system reaches the final energy upper bound, wherein, as the tracking process continues, the measurable subsystem and the unmeasurable subsystem are continuously switched, and the initial value of each switching stage will continue to decrease until equation (11) has no solution, that is, the initial value of the switching stage is equal to the final value, and the system energy reaches the final bound; note that if the designed controller does not change the gain, that is, λ′ and λ in equation (9) will be constants, the system energy will reach a fixed non-minimum final bound; since the proposed controller with stability guarantee automatically solves the controller gain that can ensure that the final energy value is less than the initial value in each switching stage, the final bound reached by the tracking system energy is the smallest when the controller is applied, and the best steady-state tracking effect can be achieved.

[0085] One use state of the present invention is: using the nonlinear model formula (1) of the drone, constructing Figure 2The UAV dual-loop control system shown in the figure uses attitude controller parameters set to α = 3, β = 0.5, p = 9, q = 10, K = 1. The state response and tracking switching reference signal are shown in Figure 4 , Figure 5 As shown, it can be seen that it can quickly converge to the reference value; the duration of the target measurable and unmeasurable situations is randomly generated: the maximum unmeasurable time T = 4, the minimum measurable time τ = 3 (see Figure 6 , where the switching signal values ​​corresponding to the measurable and unmeasurable situations are 0 and 1 respectively); the UAV maintains the initial heading and altitude to track the target, and the target moves at a variable speed (see Figure 8 );

[0086] In this simulation, three position controllers are used, which are denoted as controllers ①, ②, and ③. When the target is measurable, the gains of the three controllers are the same, set to k px =k py =0.015, k vx =k vy =0.45. When the target is unmeasurable, controller ① is the designed switching controller with stability guarantee, and its gain is determined by formula (11); controller ② is a switching controller with one-time switching, and its gain is 0 when the target is unmeasurable; controller ③ is a non-switching controller without stability guarantee, and its gain is the same as when the target is measurable.

[0087] The system energy changes obtained by the drone using the above controller to track the temporarily lost target are as follows: Figure 7 As shown in the figure, it can be seen that compared with the switching controller with one switching, the switching controller with stability guarantee can make the tracking system converge faster in the transient process; compared with the non-switching controller without stability guarantee, it can make the steady-state error of the tracking system smaller; the UAV trajectory and target trajectory of the switching controller with stability guarantee are shown in Figure 8 As shown, by applying the proposed controller, the UAV can effectively track the target when the target is temporarily lost.

[0088] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0089] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for switching control of a temporarily lost target UAV with stability guarantee, characterized in that: The steps include: S1: Initialization and establishment of tracking coordinate system; S2: UAV position control when the target is detectable; The step S2 is specifically as follows: according to the reference signal obtained by solving the step S1, the current tracking error of the UAV is calculated in the tracking coordinate system, thereby calculating and recording the tracking system energy; the current position tracking error of the UAV is brought into the state feedback controller of the position loop to obtain the required control input and the expected value of the attitude loop, and then the control input of the attitude loop is obtained; finally, the position control of the UAV is realized when the target is measurable, that is, the control of the measurable subsystem; wherein the UAV adopts a dual-loop control scheme of the position loop and the attitude loop, and the UAV dynamics model is Among them, u1~u4 are control inputs, which can be realized by adjusting the speed of the drone motor; x, y, z, θ, φ, ψ are the horizontal position of the corresponding coordinate axis and the attitude angle around the coordinate axis, and their reference signals are x, y, z, θ, φ, and ψ. d ,y d 、z d ,θ d ,φ d , d It indicates that in the controller design and stability analysis stage, according to the small angle theorem, the states z, θ, φ, ψ can be directly controlled by u1~u4, and the states x and y need to be controlled by changing the size of θ and φ; the dual-loop control scheme can be used to divide the control loop of the UAV into an inner attitude loop and an outer position loop, and the inner loop attitude desired angle is determined by the outer loop position error as the virtual control input; For the inner loop attitude controller, in order to ensure its rapid response capability, the finite time sliding mode control method is adopted; taking the attitude θ as an example, the sliding mode surface is designed as where e θ =θ-θ d , α, β, p, q are all positive real numbers and p<q, the time derivative of the sliding surface can be calculated as According to the sliding condition The equivalent controller can be designed as Combined with the reaching law commonly used in sliding mode control, the attitude controller based on the finite time sliding mode control method is finally designed as follows: u2=u θ,eq -Ksgn(s θ ) (4) Where K is a positive real number, and the design of attitude controllers u3 and u4 is the same as u2; In terms of position control, proportional-differential control is used to design the controller. Specifically, the height controller is first designed as Where: k vz , k pz is a positive real number. When the altitude reaches a steady state, u1 is always 1. At this point, the design of the four control inputs of the drone is completed. For the underactuated states x and y, the proportional-differential control idea is used to design the virtual control inputs. where k vx , k px , k vy , k py are all positive real numbers, so the position control of the UAV can be indirectly achieved by adjusting its attitude; S3: Gain determination when the target is unmeasurable; The step S3 specifically includes: according to the target measurable subsystem energy initial value recorded in step S2, the system energy of the adjacent target unmeasurable subsystem is ensured to be less than the current measurable subsystem energy initial value at the end of operation, that is, to ensure the stability of each switching stage, the controller gain of the unmeasurable subsystem is calculated, and the controller gain that can ensure that the UAV has a large acceleration is selected and recorded for backup; it is determined whether the running time of the measurable subsystem exceeds a given minimum measurable time. Once the given minimum measurable time is not reached, it is considered that the target measurability during the UAV tracking process is extremely poor, and human intervention control is required; When the target is unmeasurable, the predictor is used to predict the target state, and the predicted reference signal is obtained according to the predicted target position, which is recorded as At this time, formula (5) will be transformed into Define the systematic error as e xx d [xx d v x -v xd yy d v y -v yd ] T According to (5) and (6), the closed-loop control system of the UAV in the x and y directions can be established based on the switching system framework: Where A and A′ are composed of the controller gains in (5) and (6), respectively, and M is composed of the prediction error and the controller gain, which is the cause of system instability. The energy function of the tracking system is constructed as V = 0.5e T e. The energy functions of the measurable subsystem and the unmeasurable subsystem are represented by V(t) and V′(t), respectively, and we have Determine the upper bound of M, then obtain the error coefficient δ so that e T M≤δ||e||, then the tracking system energy function satisfies the following inequality Among them, λ and λ′ are the maximum eigenvalues ​​of the system matrices A and A′ respectively. According to the stability criterion of the switching system, an adjacent measurable subsystem and an unmeasurable subsystem are defined as a switching stage. The condition for the energy drop of the switching tracking system in a switching stage is solved as follows: It can be seen that when the controller gain satisfies (9), the system energy decreases in the current switching stage; the initial value of the switching stage energy V(0) is determined by step S2. It can be seen that for any λ, λ′, as the switching stage continues to advance, the initial value of the switching stage will be equal to the final value; then, the switching system will be ultimately bounded, and its final bound can be determined by the following formula: Obviously, as long as the initial value of the energy in the switching stage is greater than the final bound of the system, the convergence of the switching system in the current switching stage can be guaranteed, that is, It can ensure the convergence of the dynamic process of the system. Note that the right side of the equation can also be multiplied by a decay coefficient to make the convergence faster. In this step, the maximum controller gain that satisfies (11) is selected to ensure that when the target is far away, that is, when the initial energy value is large, the UAV has a large acceleration, so that it can quickly approach the target. S4: Position control when the target is undetectable; S5: controller gain update during tracking; It should be noted that the above steps are all implemented in the airborne computing unit, the target measurability is determined by the sensors and measurement algorithms loaded on the UAV, and all other calculations are numerical operations, which ensures the tracking control of the UAV with a small amount of calculation in the event of temporary loss of the target.

2. A method for switching control of a temporarily lost target UAV with stability guarantee according to claim 1, characterized in that: The step S1 is specifically as follows: the UAV receives information provided by the ground operator, including the initial position and speed of the target, the maximum speed of the target, the relative position that the UAV needs to maintain with the target, the given minimum measurable time and the maximum unmeasurable time, and the position controller gain of the preset measurable subsystem; constructs a body coordinate system and a tracking coordinate system; adjusts the attitude of the UAV so that the target is within the field of view of the sensor, so that the target position is measurable at the initial moment and within the subsequent minimum measurable time, so as to ensure that there is enough time to complete the determination of the controller gain of the adjacent unmeasurable subsystem.

3. The method for switching control of a temporarily lost target UAV with stability guarantee according to claim 1, characterized in that: In the step S4, once the target is unmeasurable, the position and speed in the unmeasurable time period are predicted according to the motion state of the target in the previous measurable time period, a reference signal is generated according to the predicted position and speed, and the controller gain determined in the step S3 is used to perform the drone position control based on state feedback consistent with the step S2, and determine whether the running time of the unmeasurable subsystem reaches the given maximum unmeasurable time. Once the given maximum unmeasurable time is exceeded, it is regarded that the target measurability during the drone tracking process is extremely poor, and human intervention control is required.

4. The method for switching control of a temporarily lost target UAV with stability guarantee according to claim 1, characterized in that: When the target is measurable again in step S5, the workflow in steps S2, S3 and S4 is repeated; since the initial energy value of the target measurable subsystem is constantly changing as the target tracking proceeds, the controller gain of the unmeasurable subsystem will be continuously updated as the target tracking proceeds, and ultimately the stability of each switching stage can be guaranteed until the energy of the switching system reaches the final energy upper bound.

Images