Obstacle avoidance method for quadrotor UAV based on adaptive nonlinear model predictive control

Abstract

The present invention discloses a quadrotor unmanned aerial vehicle (UAV) obstacle avoidance system based on adaptive nonlinear model predictive control, the system comprising a ground mission management system, a sensing system, an adaptive nonlinear model predictive obstacle avoidance system, and a power system. The present invention also discloses a quadrotor unmanned aerial vehicle (UAV) obstacle avoidance method based on adaptive nonlinear model predictive control, the method comprising the following steps: step 1, obtaining UAV target trajectory information and obstacle information; step 2, obtaining the UAV's real-time status; step 3, obtaining control instructions based on the information in steps 1 and 2. The quadrotor unmanned aerial vehicle (UAV) obstacle avoidance method based on adaptive nonlinear model predictive control provided by the present invention can simultaneously solve the two major problems of fault handling and obstacle avoidance in quadrotor unmanned aerial vehicle flight control, and significantly improve flight performance.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicle (UAV) control and obstacle avoidance, and in particular relates to an obstacle avoidance method for a quad-rotor UAV based on adaptive nonlinear model predictive control. Background Art

[0002] Quadrotors have been widely used in disaster relief, geological surveys, agricultural insurance, military reconnaissance, film and television shooting and other fields. Due to their compact and light body, high maneuverability and low cost, they also have great application prospects in meteorological surveys, land resources, marine hydropower and urban planning.

[0003] As the application scope of quadrotor drones continues to expand and their operating environments become increasingly complex, the requirements for their flight performance are also increasing. The design of flight control algorithms has become a critical step in quadrotor R&D and design. If a quadrotor experiences environmental interference during operation, its fuselage structural parameters will change, affecting the efficiency of its power unit. Furthermore, obstacles in the flight environment can also affect the drone's flight control. Therefore, fault handling and obstacle avoidance are currently two major challenges in quadrotor flight control.

[0004] Common control methods in existing technologies include sliding mode control, backstepping control, PID control, and active disturbance rejection control. However, these methods all have drawbacks and limitations, and lack a flight control method that comprehensively considers fault handling and obstacle avoidance. Therefore, a control method for quadrotor drones is urgently needed to address these challenges. Summary of the Invention

[0005] In order to overcome the above problems, the inventors conducted intensive research and designed a quadcopter obstacle avoidance method based on adaptive nonlinear model predictive control. The method simultaneously sets a nonlinear model predictive obstacle avoidance loop and an adaptive fault handling loop. The nonlinear model predictive obstacle avoidance loop obtains the optimal control strategy through the target trajectory, obstacle information and the real-time status of the UAV to avoid obstacles; the adaptive fault handling loop adjusts the control strategy through the real-time status of the UAV and the reference instructions provided by the nonlinear model predictive obstacle avoidance loop, so that the actual flight state of the quadcopter is consistent with the state under ideal conditions, thereby significantly improving the flight performance of the UAV and making it suitable for flying in complex, changeable and obstacle-filled working environments, thereby completing the present invention.

[0006] Specifically, the purpose of the present invention is to provide the following aspects:

[0007] In a first aspect, a quadrotor unmanned aerial vehicle obstacle avoidance system based on adaptive nonlinear model predictive control is provided, the system comprising a ground mission management system, a sensing system, an adaptive nonlinear model predictive obstacle avoidance system, and a power system;

[0008] Among them, the ground mission management system is used to provide the target trajectory information and real-time obstacle information of the quadcopter drone.

[0009] The sensing system is used to obtain the real-time status of the quadrotor drone.

[0010] The adaptive nonlinear model predictive obstacle avoidance system is used to obtain the control instructions of the UAV;

[0011] The power system is used to execute control instructions.

[0012] The adaptive nonlinear model prediction obstacle avoidance system includes a nonlinear model prediction obstacle avoidance subsystem and an adaptive fault handling subsystem.

[0013] The nonlinear model predictive obstacle avoidance subsystem uses the target trajectory, real-time obstacle information and the real-time status of the quadrotor drone to obtain the optimal control strategy and provide reference instructions for the adaptive fault handling subsystem;

[0014] The adaptive fault handling subsystem adjusts the control strategy according to the reference instructions and the real-time status of the quadrotor drone and outputs control instructions to the power system.

[0015] In a second aspect, a quadrotor drone obstacle avoidance method based on adaptive nonlinear model predictive control is provided, the method comprising the following steps:

[0016] Step 1: Obtain the target trajectory information and obstacle information of the UAV;

[0017] Step 2: Get the real-time status of the drone;

[0018] Step 3: Get control instructions based on the information from steps 1 and 2.

[0019] Wherein, step 3 includes the following sub-steps:

[0020] Step 3-1: Based on the target trajectory, obstacle information, and the real-time status of the drone, the drone performs obstacle avoidance control and obtains reference instructions;

[0021] In step 3-2, based on the reference instructions and the real-time status of the UAV, the UAV performs adaptive control and obtains the control instructions.

[0022] Among them, in step 3-1,

[0023] The obstacle avoidance control is performed by a nonlinear model predictive obstacle avoidance subsystem, and the cost function of the model predictive control method adopted is shown in the following formula:

[0024]

[0025] Among them, x * (t) is the target state of the UAV, x(t) is the real-time state information output by the sensor system; u r (t) is the reference instruction; Q and R are weight matrices; J AC is an additional term to the cost function, a is the error coefficient, b is the additional cost weight, x o is the obstacle position, and x is the state of the quadrotor drone.

[0026] Wherein, step 3-2 includes the following sub-steps:

[0027] Step 3-2-1, obtain the basic controller output instruction;

[0028] Step 3-2-2: perform gain adjustment on the basic controller output command to obtain the control command.

[0029] In step 3-2-2, the gain of the basic controller output command is adjusted by the following adaptive law:

[0030]

[0031]

[0032] Among them, Θ x ,Θ r are adaptive gains, representing state feedback gain and command gain respectively; Γ x , Γ r are the adaptive update rates, representing the state feedback gain adaptive rate and the command gain adaptive rate respectively; x is the state of the quadrotor drone; u r is the reference instruction; e is the error between the state of the controlled object and the reference model, B is the control matrix of the state space model of the controlled object, P l is a stable matrix, which is obtained by the following formula:

[0033] Preferably, Q2=I, where I is the unit matrix.

[0034] The control law of the adaptive fault handling is obtained by the following formula:

[0035]

[0036] The beneficial effects of the present invention include:

[0037] (1) The quadrotor UAV obstacle avoidance system based on adaptive nonlinear model predictive control provided by the present invention can solve the fault problems and obstacle avoidance problems of the quadrotor UAV during flight by setting up a nonlinear model predictive obstacle avoidance subsystem and an adaptive fault handling subsystem, and simultaneously forming a sub-adaptive fault handling loop and an obstacle avoidance constraint handling loop, so as to enable the quadrotor UAV to adapt to complex and changeable working environments with numerous obstacles;

[0038] (2) The quadrotor UAV obstacle avoidance method based on adaptive nonlinear model predictive control provided by the present invention combines nonlinear model predictive control with adaptive control, and transforms obstacle inequality constraints into nonlinear constraints, appends them to the cost function, and utilizes the ability of nonlinear model predictive algorithms to solve nonlinear problems and calculate optimal control, so that the actual flight state of the quadrotor UAV is consistent with the state under ideal conditions;

[0039] (3) The quadrotor UAV obstacle avoidance method based on adaptive nonlinear model predictive control provided by the present invention can simultaneously solve the two major problems of fault handling and obstacle avoidance in the flight control of quadrotor UAVs, and significantly improve the flight performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A schematic diagram showing a quadrotor drone obstacle avoidance method based on adaptive nonlinear model predictive control according to a preferred embodiment of the present invention is shown;

[0041] Figure 2 A schematic diagram illustrating an adaptive fault handling method according to a preferred embodiment of the present invention is shown;

[0042] Figure 3 The simulation experiment results of Example 1 are shown. DETAILED DESCRIPTION

[0043] The present invention will be further described in detail below through preferred embodiments and examples. Through these descriptions, the characteristics and advantages of the present invention will become more clear and distinct.

[0044] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0045] The present invention provides a quadrotor UAV obstacle avoidance system based on adaptive nonlinear model predictive control, such as Figure 1 As shown, the system includes a ground mission management system, a sensing system, an adaptive nonlinear model prediction obstacle avoidance system and a power system;

[0046] Among them, the ground mission management system is used to provide the target trajectory information and real-time obstacle information of the quadcopter drone.

[0047] The sensing system is used to obtain the real-time status of the quadrotor drone.

[0048] The adaptive nonlinear model predictive obstacle avoidance system is used to obtain the control instructions of the UAV;

[0049] The power system is used to execute control instructions.

[0050] In a further preferred embodiment, the ground mission management system may be a desktop computer or a laptop computer, and further includes a remote control.

[0051] In a further preferred embodiment, the sensing system includes a gyroscope, an accelerometer, a GPS signal sensor and an ultrasonic sensor.

[0052] According to a preferred embodiment of the present invention, the adaptive nonlinear model prediction obstacle avoidance system includes a nonlinear model prediction obstacle avoidance subsystem and an adaptive fault handling subsystem.

[0053] The nonlinear model predictive obstacle avoidance subsystem uses the target trajectory, real-time obstacle information and the real-time status of the quadrotor drone to obtain the optimal control strategy and provide reference instructions for the adaptive fault handling subsystem;

[0054] The adaptive fault handling subsystem adjusts the control strategy according to the reference instructions and the real-time status of the quadrotor drone and outputs control instructions to the power system.

[0055] Preferably, the nonlinear model prediction obstacle avoidance subsystem and the adaptive fault handling subsystem can both be implemented by an onboard microprocessor.

[0056] In a further preferred embodiment, the power system is a control system of a drone, which maps control instructions to the speed of the motor.

[0057] In the present invention, by setting up a nonlinear model prediction obstacle avoidance subsystem and an adaptive fault handling subsystem, and at the same time forming a sub-adaptive fault handling loop and an obstacle avoidance constraint processing loop, it is possible to solve the fault problems and obstacle avoidance problems during the flight of the quadcopter UAV, so that it can adapt to a complex and changeable working environment with obstacles everywhere.

[0058] The present invention also provides a quadrotor UAV obstacle avoidance method based on adaptive nonlinear model predictive control, such as Figure 1 As shown, the method includes the following steps:

[0059] Step 1: Obtain the drone's target trajectory information and obstacle information.

[0060] According to a preferred embodiment of the present invention, the UAV target trajectory information and obstacle information are obtained through a ground mission management system.

[0061] The ground mission management system may be a desktop computer or a laptop computer, and further includes a remote control.

[0062] Wherein, the obstacle information is real-time obstacle information.

[0063] Step 2: Get the real-time status of the drone.

[0064] According to a preferred embodiment of the present invention, the real-time status of the drone is obtained through a sensing system, which includes a gyroscope, an accelerometer, a GPS signal sensor and an ultrasonic sensor.

[0065] Preferably, the real-time status of the drone includes real-time position and real-time attitude information.

[0066] Among them, the sensor reads the environmental data in real time, and obtains the real-time status of the quadrotor drone after filtering and attitude estimation.

[0067] Step 3: Get control instructions based on the information from steps 1 and 2.

[0068] Wherein, step 3 includes the following sub-steps:

[0069] In step 3-1, based on the target trajectory, obstacle information, and the real-time status of the UAV, the UAV performs obstacle avoidance control and obtains reference instructions.

[0070] According to a preferred embodiment of the present invention, the obstacle avoidance control is performed by a nonlinear model predictive obstacle avoidance subsystem, and the cost function of the model predictive control method adopted is shown as follows:

[0071]

[0072] Among them, x * (t) is the target state of the UAV, x(t) is the real-time state information output by the sensor system; u r (t) is the reference instruction; Q and R are weight matrices; J AC Add a term to the cost function;

[0073] a is the error coefficient, b is the additional cost weight, x o is the obstacle position, and x is the state of the quadrotor drone.

[0074] The inventors have found that in the obstacle avoidance problem, it is generally desired that the difference between the current position of the quadrotor drone and the distance to the obstacle be greater than a certain value, as follows:

[0075]

[0076] Where c is a number less than or equal to zero, d des is the desired distance between the quadrotor drone and the obstacle, x o is the obstacle position, and x is the state of the quadrotor drone.

[0077] In the actual solution process, inequality constraints will greatly increase the amount of calculation and increase the computing load of the microprocessor. Therefore, in the present invention, it is preferred to convert the inequality constraints into equality constraints and integrate the obstacle avoidance problem into an additional term of the cost function in the optimal control. AC , in order to transform the linear constraints in the obstacle avoidance problem into nonlinear constraints.

[0078] In a further preferred embodiment, the drone performs obstacle avoidance control by the following formula to obtain a reference instruction:

[0079]

[0080]

[0081] u r ≤u max ;

[0082] x(0)=x0;

[0083] in, is the state space model of the UAV, A is the state transfer matrix of the quadrotor UAV, B is the control matrix, obtained by the system identification method, x * (t) represents the target state of the UAV, x(0) is the initial value of the state of the controlled object, x0 represents the state of the quadrotor UAV at each sampling point, u max Limits the controller output.

[0084] In the present invention, the obstacle avoidance system is predicted by a nonlinear model, and according to the above control method, an optimal control strategy is obtained to provide reference instructions for subsequent processes.

[0085] In step 3-2, based on the reference instructions and the real-time status of the UAV, the UAV performs adaptive control and obtains the control instructions.

[0086] Among them, such as Figure 2 As shown, step 3-2 includes the following sub-steps:

[0087] Step 3-2-1, obtain the basic controller output instruction.

[0088] In the present invention, Figure 2 As shown, u r For reference instructions, ubl is the basic controller output instruction, u is the adaptive control output (control instruction), x is the state of the quadrotor drone, and x ref is the reference model state, e r is the difference between the reference model and the state of the controlled object; the reference model should be a closed-loop model composed of the nominal model under ideal conditions of the controlled object, that is, A ref =AB*K,B ref =B, K is the basic controller gain.

[0089] According to a preferred embodiment of the present invention, the basic controller obtains the basic controller output instruction through the following formula:

[0090] u bl =-R -1 B T K(t)X(t).

[0091] In the present invention, the A(t) and B(t) matrices of the quadrotor drone state space model have been obtained through the system identification method (as described in step 3-1); the weight matrices Q(t), R(t), and P are designed according to the state space model of the quadrotor drone.

[0092] According to the above parameters, the state feedback matrix K(t) can be obtained as shown below for the design of the basic controller:

[0093]

[0094] According to the boundary condition K(t f )=P, then the final value of K(t) can be obtained, and the expression of K(t) can be obtained by integrating it against time.

[0095] Step 3-2-2: perform gain adjustment on the basic controller output command to obtain the control command.

[0096] According to a preferred embodiment of the present invention, the gain of the basic controller output command is adjusted by the following adaptive law:

[0097]

[0098]

[0099] Among them, Θ x ,Θ r are adaptive gains, representing state feedback gain and command gain respectively; Γ x , Γ r is the adaptive update rate, which represents the state feedback gain adaptive rate and the command gain adaptive rate respectively. The actual value needs to be determined through simulation experiments. The actual value is a constant matrix; x is the state of the quadrotor drone; ur is the reference instruction; e is the error between the state of the controlled object and the reference model, B is the control matrix of the state space model of the controlled object, P l is a stable matrix used to ensure the stability of the system and can be obtained by the following formula:

[0100] It is preferred to take Q2=I, where I is the unit matrix.

[0101] In the present invention, adaptive fault handling is achieved by adjusting controller gain online through an adaptive law.

[0102] In a further preferred embodiment, the control law for adaptive fault handling is obtained by the following formula:

[0103]

[0104] The quadrotor drone obstacle avoidance method based on adaptive nonlinear model predictive control described in the present invention combines nonlinear model predictive control with adaptive control, converts obstacle inequality constraints into nonlinear constraints, and appends them to the cost function. It utilizes the ability of nonlinear model predictive algorithms to solve nonlinear problems and calculates optimal control, so that the actual flight state of the quadrotor drone is consistent with the state under ideal conditions.

[0105] Example

[0106] The present invention is further described below through specific examples. However, these examples are merely exemplary and do not constitute any limitation to the scope of protection of the present invention.

[0107] Example 1

[0108] In this embodiment, obstacle avoidance control is performed by a nonlinear model predictive obstacle avoidance subsystem, and the cost function of the model predictive control method adopted is shown as follows:

[0109]

[0110] Among them, x * (t) is the target state of the UAV, x(t) is the real-time state information output by the sensor system; u r (t) is the reference instruction; Q and R are weight matrices; J AC Adding terms to the cost function to transform the linear constraints in the obstacle avoidance problem into nonlinear constraints;

[0111]

[0112] Among them, a is the error coefficient and b is the additional cost weight.

[0113] The drone performs obstacle avoidance control and obtains reference instructions through the following formula:

[0114]

[0115]

[0116] u r ≤u max ;

[0117] x(0)=x0;

[0118] in, is the state space model of the UAV, A is the state transfer matrix of the quadrotor UAV, B is the control matrix, obtained by the system identification method, x * (t) represents the target state of the UAV, x(0) is the initial value of the state of the controlled object, x0 represents the state of the quadrotor UAV at each sampling point, u max Limits the controller output.

[0119] The basic controller output command is obtained by the following formula:

[0120] u bl =-R -1 B T K(t)X(t).

[0121] The gain of the basic controller output command is adjusted by the following adaptive law:

[0122]

[0123]

[0124] Among them, Θ x ,Θ r are adaptive gains, representing state feedback gain and command gain respectively; Γ x , Γ r is the adaptive update rate, which represents the state feedback gain adaptive rate and the command gain adaptive rate respectively. The actual value needs to be determined through simulation experiments. The actual value is a constant matrix; x is the state of the quadrotor drone; u r is the reference instruction; e is the error between the state of the controlled object and the reference model, B is the control matrix of the state space model of the controlled object, P l is a stable matrix, which is obtained by the following formula:

[0125]

[0126] A simulation experiment was used to verify the application effect of the method described in this embodiment in the position control of a quadrotor. In this embodiment, a quadrotor UAV position kinematic point model was used, as follows:

[0127]

[0128] Among them, x, y, z represent the position of the quadrotor drone, v x ,v y ,v z Indicates the speed of the quadrotor drone in three directions, u x ,u y ,u z It is the control quantity, which represents the acceleration of the quadrotor drone in three directions.

[0129] In the simulation experiment, Figure 3 As shown, v x =20,v y =0,v z =0, that is, the speed in the x-axis direction is 20m / s, the speeds in the y- and z-axis directions are zero, and the simulation time is 5s.

[0130] The obstacle is a rectangular area, and the coordinates of the four points are (47, -4), (60, -4), (47, 4), and (60, 4).

[0131] Depend on Figure 3 It can be seen from the simulation results that the method of the present invention can enable the quadrotor drone to avoid obstacles.

[0132] The present invention has been described in detail above with reference to specific embodiments and exemplary examples. However, these descriptions are not to be construed as limiting the present invention. Those skilled in the art will appreciate that, without departing from the spirit and scope of the present invention, various equivalent substitutions, modifications, or improvements may be made to the technical solutions and implementations of the present invention, all of which fall within the scope of the present invention.

Claims

1. A quadrotor UAV obstacle avoidance method based on adaptive nonlinear model predictive control, characterized in that: The method comprises the following steps: Step 1: Obtain the target trajectory information and obstacle information of the UAV; Step 2: Get the real-time status of the drone; Step 3: Obtain control instructions based on the information in steps 1 and 2; Step 3 includes the following sub-steps: Step 3-1: Based on the target trajectory, obstacle information, and the real-time status of the drone, the drone performs obstacle avoidance control and obtains reference instructions; Step 3-2: Based on the reference command and the real-time status of the UAV, the UAV performs adaptive control and obtains the control command; In step 3-1, The obstacle avoidance control is performed by a nonlinear model predictive obstacle avoidance subsystem, and the cost function of the model predictive control method adopted is shown in the following formula: Among them, x * (t) is the target state of the UAV, x(t) is the real-time state information output by the sensor system; u r (t) is the reference instruction; Q and R are weight matrices; J AC is an additional term to the cost function, a is the error coefficient, b is the additional cost weight, x o is the obstacle position, and x is the state of the quadrotor drone.

2. The quadrotor UAV obstacle avoidance method based on adaptive nonlinear model predictive control according to claim 1, characterized in that: Step 3-2 includes the following sub-steps: Step 3-2-1, obtain the basic controller output instruction; Step 3-2-2: perform gain adjustment on the basic controller output command to obtain the control command.

3. The quadrotor UAV obstacle avoidance method based on adaptive nonlinear model predictive control according to claim 2, characterized in that: In step 3-2-2, the gain of the basic controller output command is adjusted by the following adaptive law: Among them, Θ x ,Θ r are adaptive gains, representing state feedback gain and command gain respectively; Γ x , Γ r are the adaptive update rates, representing the state feedback gain adaptive rate and the command gain adaptive rate respectively, x is the state of the quadrotor drone; u r is the reference instruction; e is the error between the state of the controlled object and the reference model, B is the control matrix of the state space model of the controlled object, P l is a stable matrix, which is obtained by the following formula: Q2=I, where I is the unit matrix.

Images