[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

WO2019085834A1 - 一种无人飞行器稳定飞行控制方法 - Google Patents

一种无人飞行器稳定飞行控制方法 Download PDF

Info

Publication number
WO2019085834A1
WO2019085834A1 PCT/CN2018/112114 CN2018112114W WO2019085834A1 WO 2019085834 A1 WO2019085834 A1 WO 2019085834A1 CN 2018112114 W CN2018112114 W CN 2018112114W WO 2019085834 A1 WO2019085834 A1 WO 2019085834A1
Authority
WO
WIPO (PCT)
Prior art keywords
aircraft
equation
aerial vehicle
unmanned aerial
controller
Prior art date
Application number
PCT/CN2018/112114
Other languages
English (en)
French (fr)
Inventor
张智军
郑陆楠
吉冬昱
Original Assignee
华南理工大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华南理工大学 filed Critical 华南理工大学
Priority to US16/652,457 priority Critical patent/US11721219B2/en
Publication of WO2019085834A1 publication Critical patent/WO2019085834A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05DSYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/08Control of attitude, i.e. control of roll, pitch, or yaw
    • G05D1/0808Control of attitude, i.e. control of roll, pitch, or yaw specially adapted for aircraft
    • G05D1/0858Control of attitude, i.e. control of roll, pitch, or yaw specially adapted for aircraft specially adapted for vertical take-off of aircraft
    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05DSYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/10Simultaneous control of position or course in three dimensions
    • G05D1/101Simultaneous control of position or course in three dimensions specially adapted for aircraft
    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G5/00Traffic control systems for aircraft
    • G08G5/30Flight plan management
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B64AIRCRAFT; AVIATION; COSMONAUTICS
    • B64UUNMANNED AERIAL VEHICLES [UAV]; EQUIPMENT THEREFOR
    • B64U10/00Type of UAV
    • B64U10/10Rotorcrafts
    • B64U10/13Flying platforms
    • B64U10/16Flying platforms with five or more distinct rotor axes, e.g. octocopters
    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05DSYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/0088Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots characterized by the autonomous decision making process, e.g. artificial intelligence, predefined behaviours
    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05DSYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/08Control of attitude, i.e. control of roll, pitch, or yaw
    • G05D1/0808Control of attitude, i.e. control of roll, pitch, or yaw specially adapted for aircraft
    • G05D1/0816Control of attitude, i.e. control of roll, pitch, or yaw specially adapted for aircraft to ensure stability
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B64AIRCRAFT; AVIATION; COSMONAUTICS
    • B64UUNMANNED AERIAL VEHICLES [UAV]; EQUIPMENT THEREFOR
    • B64U2201/00UAVs characterised by their flight controls
    • B64U2201/10UAVs characterised by their flight controls autonomous, i.e. by navigating independently from ground or air stations, e.g. by using inertial navigation systems [INS]

Definitions

  • the invention relates to a flight control method, in particular to a method for stable flight control of an unmanned aerial vehicle.
  • the PID controller and the feedback closed-loop control system are simple in design and have good control effects, they are widely used in the controller design of the aircraft. Although the PID controller is easy to use, the PID controller and the power distribution scheme obtained according to the PID controller do not achieve the desired stability of the unmanned aerial vehicle.
  • the object of the present invention is to overcome the deficiencies of the prior art and provide a design method for a stable flight controller and a power distribution scheme.
  • the method uses a sensor to acquire real-time flight operation data of an aircraft, and solves the output of the aircraft motor through a multi-layer zero-transformation neural network. The amount of control is controlled, and the corresponding power distribution scheme is obtained to achieve stable flight control of the unmanned aerial vehicle.
  • An unmanned aerial vehicle stable flight control method comprising:
  • step 2) According to the flight real-time running data and target attitude data obtained in step 1), construct a deviation function; construct a neural dynamics equation based on the deviation function by using multi-layer zero-transformation dynamics method; the deviation function-based nerve corresponding to all parameters
  • the kinetic equations together constitute the controller of the unmanned aerial vehicle, and the output of the differential equation of the controller is the output control amount of the aircraft motor;
  • the control amount obtained by the controller has the following relationship with the power of the multi-rotor UAV motor:
  • W is the power distribution matrix of the unmanned aerial vehicle.
  • the matrix W has different forms according to the different structures and the number of rotors, and needs to be determined according to its structure and the number of rotors;
  • the corresponding motor power F is obtained by matrix inversion or pseudo-inverse, namely:
  • W -1 is obtained by inverse operation. If W is not a square matrix, W -1 is solved by the corresponding pseudo inverse operation; finally, the required motor power F is obtained and according to the motor voltage and The relationship of power controls the input voltage of the motor to control the motor speed, and finally realizes the control of the motor power and completes the stable flight control of the drone.
  • parsing of the kinematics problem of the aircraft by the carried processor includes:
  • ground coordinate system E and the body coordinate system B.
  • is the roll angle
  • is the pitch angle
  • is the yaw angle
  • I 3 ⁇ 3 is the unit matrix
  • I is the inertia matrix
  • V is the linear velocity in the body coordinate system
  • is the angular velocity in the body coordinate system
  • F is the combined external force
  • is the combined torque
  • the establishing the aircraft dynamics model includes:
  • l is the arm length
  • g is the gravitational acceleration
  • x, y, and z are the position coordinates of the aircraft in the ground coordinate system
  • I x , I y , and I z are the moments of inertia of the aircraft on the X, Y, and Z axes, respectively.
  • u x cos ⁇ sin ⁇ cos ⁇ +sin ⁇ sin ⁇
  • u y cos ⁇ sin ⁇ sin ⁇ -sin ⁇ cos ⁇
  • u 1 , u 2 , u 3 , u 4 are output control quantities.
  • controller for designing the UAV includes:
  • the deviation function of the design regarding the output control quantity u 1 and the corresponding drone height controller specifically include:
  • the deviation function can be defined according to the target height value z T and the actual height value z in the Z-axis direction.
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (4) is generally not valid in the initial situation and does not contain the relevant information of the output control amount, the solution of the control amount cannot be realized, so further design is needed, so the definition is
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • the deviation function of the design regarding u x , u y , and the corresponding drone position controller specifically include:
  • the deviation function can be defined according to the target value x T and the actual value x in the X-axis direction.
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (17) is generally not valid in the initial situation and does not contain the relevant information of the output control amount, the solution of the control amount cannot be realized, so further design is needed, so the definition is
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • equation (22) can be reduced to
  • the deviation function can be defined according to the target value y T and the actual value y in the Y-axis direction.
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (30) is generally not valid in the initial situation and does not contain the relevant information of the output control amount, the solution amount cannot be solved, so further design is needed, so
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • equation (35) can be reduced to
  • the target attitude angles ⁇ T and ⁇ T are inversely solved, and the calculation method is
  • the method is characterized in that the deviation function of the output control quantity u 2 ⁇ u 4 is designed, and the corresponding UAV attitude controller includes:
  • the deviation function can be defined as
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (44) is generally not valid in the initial situation and does not contain information about the output control amount, the solution to the control amount cannot be solved, so further design is required, so
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • equation (49) can be simplified to
  • the deviation function can be defined as
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (57) is generally not valid in the initial situation and does not contain the relevant information of the output control amount, the solution of the control amount cannot be realized, so further design is needed, so the definition is
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • equation (62) can be simplified to
  • the deviation function can be defined as
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (70) is generally not valid in the initial situation and does not contain the relevant information of the output control amount, the solution amount cannot be solved, so further design is needed, so definition
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • equation (75) can be simplified to
  • the height controller designed according to the height variable z, the position controller designed according to the position variables x and y, and the attitude controller designed according to the attitude control quantities ⁇ , ⁇ , and ⁇ together constitute a multi-rotor unmanned aerial vehicle Stable flight controller, including:
  • an unmanned aerial vehicle controller can be obtained, which can be realized by network structure; the unmanned aerial vehicle controller can control the unmanned aerial vehicle to stably fly
  • the controller can be written in the following form:
  • the zeroing neural network is constructed by the differential equation of the controller, and the unmanned aerial vehicle control amount is solved by the zeroing neural network.
  • the invention has the following beneficial effects:
  • the multi-layered zero neural network has better convergence characteristics and can realize the real-time response of the aircraft. At the same time, the multi-layered zero neural network has strong robustness.
  • the controller system designed according to the neural network is stable and controlled. The effect is good.
  • the invention is based on a multi-layered zero-dynamic neurodynamics method, which is described by a ubiquitous implicit dynamic model, which can fully utilize the derivative information of each time-varying parameter from the method and system level, and has a certain prediction for solving the problem.
  • Capabilities which can quickly and accurately approximate the correct solution in real time, can solve a variety of time-varying problems such as matrix, vector, algebra and optimization.
  • FIG. 1 is a flow chart of a method for controlling a stable flight of a multi-rotor aircraft according to an embodiment of the present invention
  • Figure 2 is a side view showing the structure of the multi-rotor aircraft of the present invention.
  • Figure 3 is a plan view showing the structure of the multi-rotor aircraft of the present invention.
  • Figure 4 is a three-dimensional view of the structure of the multi-rotor aircraft of the present invention.
  • Figure 5 is a coordinate diagram of a multi-rotor aircraft body.
  • the embodiment provides a method for stable flight control of an unmanned aerial vehicle, and the method includes the following steps:
  • S1 Obtain the flight real-time running data of the aircraft by using the attitude sensor, the position sensor and the height sensor mounted on the unmanned aerial vehicle, and analyze the kinematics of the aircraft by the equipped processor to establish an aircraft dynamics model. ;
  • FIGS. 3 and 4 One type of rotor flight structure in a multi-rotor aircraft is shown in Figures 2, 3 and 4.
  • the structure is a six-rotor aircraft mechanism model consisting of a multi-rotor aircraft propeller, a brushless motor, a rotor arm and a fuselage.
  • the directions of the arrows in FIGS. 3 and 4 indicate the direction of rotation of the motor, and the combination of the clockwise and counterclockwise directions of the illustrated rotation direction is to achieve mutual cancellation of the motor torque to achieve stable steering control.
  • the invention utilizes an algorithm such as a four-element algorithm and a Kalman filter to obtain real-time attitude data ⁇ (t), ⁇ (t) and ⁇ (t) of the aircraft by using sensors such as a gyroscope and an accelerometer mounted on the multi-rotor aircraft.
  • the position data x(t), y(t) and z(t) of the aircraft in three-dimensional space are obtained by using the height sensor and the position sensor.
  • the multi-rotor aircraft in Figure 5 is defined as follows according to the body coordinate system:
  • the six motors of the six-rotor aircraft are defined in a clockwise direction to be No. 1 to No. 6;
  • the X-axis is oriented in the direction of the No. 1 rotor arm, and is directed toward the forward direction of the aircraft through the weight of the machine;
  • the Y-axis is along the axis of symmetry of the rotor arms of No. 2 and No. 3, and points to the right direction of movement of the aircraft through the weight of the machine;
  • the Z-axis is perpendicular to the plane of the six-rotor, pointing to the direction of climb of the aircraft through the weight of the machine;
  • the pitch angle ⁇ is the angle between the X-axis of the body and the horizontal plane of the earth, and is set to be positive when the body is down;
  • the roll angle ⁇ is the angle between the Z-axis of the body and the vertical plane of the earth passing through the X-axis of the body, and is set to be positive when the body is turned to the right;
  • the yaw angle ⁇ is the angle between the projection of the X-axis of the body on the horizontal plane and the X-axis in the geodetic coordinate system, and is set to be positive when the head is to the left.
  • ground coordinate system E and the body coordinate system B.
  • is the roll angle
  • is the pitch angle
  • is the yaw angle
  • I 3 ⁇ 3 is the unit matrix
  • I is the inertia matrix
  • V is the linear velocity in the body coordinate system
  • is the angular velocity in the body coordinate system
  • F is the combined external force
  • is the combined torque
  • l is the arm length
  • g is the gravitational acceleration
  • x, y, and z are the position coordinates of the aircraft in the ground coordinate system
  • I x , I y , and I z are the moments of inertia of the aircraft on the X, Y, and Z axes, respectively.
  • u x cos ⁇ sin ⁇ cos ⁇ +sin ⁇ sin ⁇
  • u y cos ⁇ sin ⁇ sin ⁇ -sin ⁇ cos ⁇
  • u 1 , u 2 , u 3 , u 4 are output control quantities.
  • the deviation function of the output control quantity u 1 is designed.
  • the multi-rotor unmanned aerial vehicle height controller is designed to solve for u 1 , and then the horizontal position x, y is set, and the design is about u x , the deviation function of u y and the corresponding multi-rotor unmanned aerial vehicle position controller, inversely solve the target attitude angles ⁇ T and ⁇ T , according to the target attitude angle, respectively, the roll angle ⁇ , the pitch angle ⁇ and the yaw angle ⁇
  • the specific steps are as follows:
  • the deviation function can be defined according to the target height value z T and the actual height value z in the Z-axis direction.
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (4) is generally not valid in the initial situation and does not contain the relevant information of the output control amount, the solution of the control amount cannot be realized, so further design is needed, so the definition is
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • the deviation function can be defined according to the target value x T and the actual value x in the X-axis direction.
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (17) is generally not valid in the initial situation and does not contain the relevant information of the output control amount, the solution of the control amount cannot be realized, so further design is needed, so the definition is
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • equation (22) can be reduced to
  • the deviation function can be defined according to the target value y T and the actual value y in the Y-axis direction.
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (30) is generally not valid in the initial situation and does not contain the relevant information of the output control amount, the solution amount cannot be solved, so further design is needed, so
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • equation (35) can be reduced to
  • the deviation function can be defined as
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (44) is generally not valid in the initial situation and does not contain information about the output control amount, the solution to the control amount cannot be solved, so further design is required, so
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • equation (49) can be simplified to
  • the deviation function can be defined as
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (57) is generally not valid in the initial situation and does not contain the relevant information of the output control amount, the solution of the control amount cannot be realized, so further design is needed, so the definition is
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • equation (62) can be simplified to
  • the deviation function can be defined as
  • the neurodynamic equation based on the deviation function can be designed as
  • equation (70) is generally not valid in the initial situation and does not contain the relevant information of the output control amount, the solution amount cannot be solved, so further design is needed, so definition
  • the neurodynamic equation based on the deviation function can be designed as
  • the deviation function can be defined
  • equation (75) can be simplified to
  • S3 realizing the real-time operation data and the target attitude data of the acquired aircraft, and solving the output control amount of the aircraft motor through the designed multi-layer zero-transformation neural network controller;
  • an unmanned aerial vehicle controller can be obtained, which can be realized by network structure; the UAV controller can Control the unmanned aerial vehicle to stabilize flight, where the controller can be written in the following form:
  • the zeroing neural network is constructed by the differential equation of the controller, and the unmanned aerial vehicle control amount is solved by the zeroing neural network.
  • step 4 Passing the solution result of step 4) to the aircraft motor governor, and controlling the motor power according to the relationship between the control quantity obtained by the controller and the motor power of the multi-rotor unmanned aerial vehicle, and controlling the movement of the unmanned aerial vehicle ;
  • the control amount obtained by the controller has the following relationship with the power of the multi-rotor UAV motor:
  • U [u 1 u 2 u 3 u 4 ] T is the control quantity of the unmanned aerial vehicle
  • F [F 1 ... F j ] T is the motor power of the unmanned aerial vehicle
  • j is the motor of the multi-rotor drone Number
  • W is the power distribution matrix of the unmanned aerial vehicle.
  • the corresponding motor power F can be obtained by matrix inversion or pseudo-inverse, ie
  • the matrix W is square and reversible, then W -1 is obtained by inverse operation. If W is not a square matrix, W -1 is solved by the corresponding pseudo inverse operation; finally, the required motor power F is obtained and according to the motor voltage and The relationship of power controls the input voltage of the motor to control the motor speed, and finally realizes the control of the motor power and completes the stable flight control of the drone. Since the number and structure of different rotors will affect the control mode of the multi-rotor UAV, the matrix W will have different forms depending on the structure and the number of rotors.
  • W- 1 can be obtained by pseudo-reverse, that is,
  • the power distribution of the six-rotor UAV can be obtained and the corresponding actual motor control can be used to control the motor operation.

Landscapes

  • Engineering & Computer Science (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Automation & Control Theory (AREA)
  • Mechanical Engineering (AREA)
  • Business, Economics & Management (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Game Theory and Decision Science (AREA)
  • Medical Informatics (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
  • Measuring Magnetic Variables (AREA)
  • Feedback Control In General (AREA)

Abstract

一种无人飞行器稳定飞行控制方法,包括如下步骤:通过无人飞行器上所搭载的姿态传感器、位置传感器以及高度传感器获取飞行器自身的飞行实时运行数据,通过所搭载的处理器对飞行器的运动学问题进行相应的解析处理,建立飞行器动力学模型(S1);根据多层零化神经动力学方法,设计无人飞行器的控制器(S2);利用获取的飞行器实时运行数据与目标姿态数据,通过设计的多层零化神经网络控制器求解飞行器电机的输出控制量(S3);将求解结果传递给飞行器电机调速器,根据控制器求解的得到的控制量与多旋翼无人机电机动力的关系,实现对电机动力的控制,控制无人飞行器的运动(S4)。基于多层零化神经动力学方法,可快速、准确、实时地逼近问题正确解,能够很好地解决时变问题。

Description

一种无人飞行器稳定飞行控制方法 技术领域
本发明涉及一种飞行控制方法,尤其是一种无人飞行器稳定飞行控制方法。
技术背景
近年来,世界无人飞行器技术得到迅猛的发展,具有垂直起降、稳定悬停、无线传输、远程航拍和自主巡航能力的多旋翼飞行器在军事及民事领域具有广阔的应用前景。小型旋翼式无人机由于具有优异的机动性能、简单的机械结构、方便部署与维护简单等特点,被广泛应用于航拍摄影、电力巡检、环境监测、森林防火,灾情巡查、防恐救生、军事侦察及战场评估等领域。伴随着无人飞行器的广泛应用,稳定以及反应快速的无人机控制器设计引起了众多研究者的关注,而传统的无人机控制器都是基于PID闭环控制算法以及相应的改进控制算法进行设计的。由于PID控制器以及反馈闭环控制系统设计简单,且已具备较好的控制效果,在飞行器的控制器设计上被广泛利用。尽管PID控制器使用简便,但PID控制器以及根据PID控制器所得到的动力分配方案并不能使无人飞行器达到理想中的稳定性。
发明内容
本发明的目的在于克服现有技术的不足,提供一种稳定飞行控制器及动力分配方案的设计方法,该方法利用传感器获取飞行器实时飞行运行数据,通过多层零化神经网络求解飞行器电机的输出控制量,并得到相应的动力分配方案,实现控制无人飞行器的稳定飞行。
本发明的目的可以通过如下技术方案实现:
一种无人飞行器稳定飞行控制方法,所述方法包括:
1)获取飞行器自身的飞行实时运行数据,对飞行器的运动学问题进行解析处理,建立飞行器动力学模型;
2)根据步骤1)获取的飞行实时运行数据与目标姿态数据,构建偏差函数;利用多层零化神经动力学方法,构建基于偏差函数的神经动力学方程;所有参数对应的基于偏差函数的神经动力学方程,共同构成无人飞行器的控制器,控制器的微分方程解算的输出量为飞行器电机的输出控制量;
3)根据步骤2)求解得到的输出控制量与多旋翼无人机电机动力的关 系,控制电机动力,完成无人飞行器运动的控制,具体步骤为:
根据无人机动力分配方案,控制器求解得到的控制量与多旋翼无人机电机动力存在如下关系:
U=WF
其中U=[u 1 u 2 u 3 u 4] T为无人飞行器的输出控制量,F=[F 1 … F j] T为无人飞行器的电机动力,j为多旋翼无人机的电机个数,W为无人飞行器动力分配矩阵,矩阵W根据不同结构与旋翼数目会有不同的形式,需要根据其结构和旋翼数确定;
通过矩阵求逆或求伪逆的形式得到相应的电机动力F,也即:
F=W -1U
若矩阵W为方阵且可逆,则W -1通过求逆运算得到,若W不为方阵,则通过相应的伪逆运算求解W -1;最终得到所需电机动力F并根据电机电压与动力的关系控制电机输入电压以控制电机转速,最终实现对电机动力的控制,完成无人机稳定飞行控制。
进一步的,所述的通过所搭载的处理器对飞行器的运动学问题进行相应的解析处理,具体包括:
定义地面坐标系E和机体坐标系B,地面坐标系和机体坐标系可通过转换矩阵R建立联系:E=RB,R可表示为
Figure PCTCN2018112114-appb-000001
其中φ为横滚角,θ为俯仰角,ψ为偏航角;
忽略飞行器所受空气阻力作用,在机体坐标系下,飞行器系统受力分析(Newton-Euler形式)如下
Figure PCTCN2018112114-appb-000002
其中m为飞行器的总质量,I 3×3为单位矩阵,I为惯性矩阵,V为机体坐标系下的线速度,ω为机体坐标系下的角速度,F为合外力,τ为合力矩;
进一步的,所述的建立飞行器动力学模型,具体包括:
根据定义的地面坐标系E和机体坐标系B、两者通过转换矩阵R建立的联系:E=RB以及在机体坐标系下,飞行器系统的受力分析,得到多旋翼飞行器的动力学方程为
Figure PCTCN2018112114-appb-000003
其中l为臂长,g为重力加速度,x、y、z分别为飞行器在地面坐标系下的位置坐标,I x、I y、I z分别为飞行器在X、Y、Z轴上的转动惯量,u x=cosφsinθcosψ+sinφsinψ,u y=cosφsinθsinψ-sinφcosψ,u 1、u 2、u 3、u 4为输出控制量。
进一步的,所述设计无人飞行器的控制器,具体包括:
(1)通过多层零化神经动力学方法,由垂直高度z出发,设计关于输出控制量u 1的偏差函数,根据该偏差函数,设计无人机高度控制器;
(2)通过多层零化神经动力学方法,由水平位置x、y出发,设计关于u x、u y的偏差函数,根据该偏差函数,设计无人机位置控制器,再反解出目标姿态角度φ T和θ T
(3)通过多层零化神经动力学方法,分别由横滚角φ、俯仰角θ以及偏航角ψ出发,设计关于输出控制量u 2~u 4的偏差函数,根据该偏差函数,设计姿态控制器。
进一步的,所述设计关于输出控制量u 1的偏差函数,以及相应的无人机高度控制器,具体包括:
针对垂直高度z,根据Z轴方向上的目标高度值z T以及实际高度值z,可以定义偏差函数为
e z1=z-z T (1)
并可以得到其导数为
Figure PCTCN2018112114-appb-000004
为了使实际值z能够收敛到目标值z T,根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000005
其中γ为一常数。
将等式(1)和(2)代入等式(3),整理得
Figure PCTCN2018112114-appb-000006
因为等式(4)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000007
并可以得到其导数为
Figure PCTCN2018112114-appb-000008
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000009
将等式(5)和(6)代入等式(7),整理得
Figure PCTCN2018112114-appb-000010
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000011
根据飞行器动力学方程,(9)可简化为
E z=a zu 1+b z (10)
其中
Figure PCTCN2018112114-appb-000012
并可以得到其导数为
Figure PCTCN2018112114-appb-000013
利用多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000014
将等式(10)和(11)代入等式(12)整理得
Figure PCTCN2018112114-appb-000015
进一步的,所述设计关于u x、u y的偏差函数,以及相应的无人机位置控制器,具体包括:
针对水平位置x,根据X轴方向上的目标值x T以及实际值x,可以定义偏差函数
e x1=x-x T (14)
并可以得到其导数为
Figure PCTCN2018112114-appb-000016
为了使实际值x能够收敛到目标值x T,根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000017
将等式(14)和(15)代入等式(16)整理得
Figure PCTCN2018112114-appb-000018
因为等式(17)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000019
并可以得到其导数为
Figure PCTCN2018112114-appb-000020
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000021
将等式(18)和(19)代入等式(20)整理得
Figure PCTCN2018112114-appb-000022
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000023
根据飞行器动力学方程,等式(22)可化简为
E x--a xu x+b x (23)
其中
Figure PCTCN2018112114-appb-000024
并可以得到其导数为
Figure PCTCN2018112114-appb-000025
利用多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000026
将等式(23)和(24)代入等式(25)整理得
Figure PCTCN2018112114-appb-000027
针对水平位置y,根据Y轴方向上的目标值y T以及实际值y,可以定义偏差函数
e y1=y-y T (27)
并可以得到其导数为
Figure PCTCN2018112114-appb-000028
为了使实际值y能够收敛到目标值y T,根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000029
将等式(27)和(28)代入等式(29)整理得
Figure PCTCN2018112114-appb-000030
因为等式(30)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000031
并可以得到其导数为
Figure PCTCN2018112114-appb-000032
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000033
将等式(31)和(32)代入等式(33)整理得
Figure PCTCN2018112114-appb-000034
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000035
根据飞行器动力学方程,等式(35)可化简为
E y=a yu y+b y (36)
其中
Figure PCTCN2018112114-appb-000036
并可以得到其导数为
Figure PCTCN2018112114-appb-000037
利用多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000038
将等式(36)和(37)代入等式(38)整理得
Figure PCTCN2018112114-appb-000039
进一步的,所述的根据设计的位置控制器,反解目标姿态角度φ T和θ T,其计算方法为
由等式(26)和(39)求解出的u x、u y
Figure PCTCN2018112114-appb-000040
从而反解出的目标角度值φ T和θ T
Figure PCTCN2018112114-appb-000041
进一步的,所述其特征在于,设计关于输出控制量u 2~u 4的偏差函数,以及相应的无人机姿态控制器,具体包括:
针对横滚角φ,根据(40)中求解出的目标角度φ T以及实际角度φ,可以定义偏差函数为
e φ1=φ-φ T (41)
并可以得到其导数为
Figure PCTCN2018112114-appb-000042
为了使实际值φ能够收敛到目标值φ T,根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000043
将等式(41)和(42)代入等式(43)整理得
Figure PCTCN2018112114-appb-000044
因为等式(44)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000045
并可以得到其导数为
Figure PCTCN2018112114-appb-000046
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000047
将等式(45)和(46)代入等式(47)整理得
Figure PCTCN2018112114-appb-000048
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000049
根据飞行器动力学方程,等式(49)可简化为
E φ=a φu 2+b φ (50)
其中
Figure PCTCN2018112114-appb-000050
并可以得到其导数为
Figure PCTCN2018112114-appb-000051
根据多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000052
将等式(50)和(51)代入等式(52)整理得
Figure PCTCN2018112114-appb-000053
针对俯仰角θ,根据等式(40)中求解出的目标角度θ T以及实际角度θ,可以定义偏差函数为
e θ1=θ-θ T (54)
并可以得到其导数为
Figure PCTCN2018112114-appb-000054
为了使实际值θ能够收敛到目标值θ T,根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000055
将等式(54)和(55)代入等式(56)整理得
Figure PCTCN2018112114-appb-000056
因为等式(57)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000057
并可以得到其导数为
Figure PCTCN2018112114-appb-000058
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000059
将等式(58)和(59)代入等式(60)整理得
Figure PCTCN2018112114-appb-000060
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000061
根据飞行器动力学方程,等式(62)可简化为
E θ=a θu 3+b θ (63)
其中
Figure PCTCN2018112114-appb-000062
并可以得到其导数为
Figure PCTCN2018112114-appb-000063
根据多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000064
将等式(63)和(64)代入等式(65)整理得
Figure PCTCN2018112114-appb-000065
针对偏航角ψ,根据人为设定的角度ψ T以及实际角度ψ,可以定义偏差函数为
e ψ1=ψ-ψ T (67)
并可以得到其导数为
Figure PCTCN2018112114-appb-000066
为了使实际值ψ能够收敛到目标值ψ T,根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000067
将等式(67)和(68)代入等式(69)整理得
Figure PCTCN2018112114-appb-000068
因为等式(70)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000069
并可以得到其导数为
Figure PCTCN2018112114-appb-000070
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000071
将等式(71)和(72)代入等式(73)整理得
Figure PCTCN2018112114-appb-000072
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000073
根据飞行器动力学方程,等式(75)可简化为
E ψ=a ψu 4+b ψ (76)
其中
Figure PCTCN2018112114-appb-000074
并可以得到其导数为
Figure PCTCN2018112114-appb-000075
根据多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000076
将等式(76)和(77)代入等式(78)整理得
Figure PCTCN2018112114-appb-000077
进一步的,所述根据高度变量z设计的高度控制器,根据位置变量x和y设计的位置控制器以及根据姿态控制量φ、θ和ψ设计的姿态控制器,共同构成多旋翼无人飞行器的稳定飞行控制器,具体包括:
根据等式(13)、(53)、(66)和(79),可得无人飞行器控制器,该控制器能够通过网络结构化实现;该无人飞行器控制器能够控制无人飞行器稳定飞行;其中该控制器可以写成如下的形式:
Figure PCTCN2018112114-appb-000078
通过该控制器微分方程构建零化神经网络,通过零化神经网络解算无人飞行器控制量。
本发明相对现有技术,具有如下有益效果:
1、多层零化神经网络具有较好的收敛特性,能够实现飞行器的实时响应,同时,多层零化神经网络具有较强的鲁棒性,根据该神经网络设计的控制器系统稳定且控制效果良好。
2、本发明基于多层零化神经动力学方法,该方法采用普遍存在的隐 动力学模型进行描述,可从方法和系统层面上充分利用各时变参数的导数信息,对问题求解具有一定预测能力,可快速、准确、实时地逼近问题正确解,可以很好地解决矩阵、向量、代数及优化等多种时变问题。
附图说明
图1为本发明实施的多旋翼飞行器稳定飞行控制方法流程图;
图2为本发明的多旋翼飞行器结构侧视图;
图3为本发明的多旋翼飞行器结构俯视图;
图4为本发明的多旋翼飞行器结构三维视图;
图5为多旋翼飞行器机体坐标系图。
具体实施方式
下面结合实施例和附图对本发明做进一步详细的描述,但本发明的实施方式不限于此。
实施例:
如图1所示,本实施例提供了一种无人飞行器稳定飞行控制方法,该方法包括如下步骤:
S1:通过无人飞行器上所搭载的姿态传感器、位置传感器以及高度传感器获取飞行器自身的飞行实时运行数据,通过所搭载的处理器对飞行器的运动学问题进行相应的解析处理,建立飞行器动力学模型;
多旋翼飞行器中的一种旋翼飞行结构如图2、图3以及图4所示。该结构为六旋翼飞行器机构模型,该机构模型由多旋翼飞行器螺旋桨、无刷电机、旋翼臂与机身组成。图3和图4中的箭头方向指示电机的旋转方向,而图示旋转方向顺时针与逆时针组合目的为实现电机转矩的相互抵消,实现稳定的转向控制。
本发明利用四元素算法以及卡尔曼滤波等算法,可以实现利用多旋翼飞行器上搭载的陀螺仪与加速度计等传感器获取飞行器的实时姿态数据θ(t),φ(t)以及ψ(t),利用高度传感器以及位置传感器获取飞行器在三维空间中的位置数据x(t),y(t)与z(t)。
飞行器姿态变量的定义如图5所示。
图5中的多旋翼飞行器根据机体坐标系做出如下定义:
(1)按照顺时针方向定义六旋翼飞行器六个电机分别为①号到⑥号;
(2)X轴沿①号旋翼臂方向,通过机体重心指向飞行器前进方向;
(3)Y轴沿②、③号旋翼臂的对称轴方向,通过机体重心指向飞行器右侧运动方向;
(4)Z轴垂直于六旋翼平面向上,通过机体重心指向飞行器爬升方向;
(5)俯仰角θ为机体X轴与大地水平面之间的夹角,设定机身向下时为正;
(6)横滚角φ为机体Z轴与过机体X轴的大地竖直平面之间的夹角,设定机身向右时为正;
(7)偏航角ψ为机体X轴在大地水平面上的投影与大地坐标系中X轴之间的夹角,设定机头向左时为正。
根据不同的旋翼飞行器模型,建立针对该飞行器的物理模型等式以及动力学方程,可以通过如下飞行器动力学建模步骤完成动力学分析:
定义地面坐标系E和机体坐标系B,地面坐标系和机体坐标系可通过转换矩阵R建立联系:E=RB,R可表示为
Figure PCTCN2018112114-appb-000079
其中φ为横滚角,θ为俯仰角,ψ为偏航角;
忽略飞行器所受空气阻力作用,在机体坐标系下,飞行器系统受力分析(Newton-Euler形式)如下
Figure PCTCN2018112114-appb-000080
其中m为飞行器的总质量,I 3×3为单位矩阵,I为惯性矩阵,V为机体坐标系下的线速度,ω为机体坐标系下的角速度,F为合外力,τ为合力矩;
根据上述等式可得飞行器的动力学方程为
Figure PCTCN2018112114-appb-000081
其中l为臂长,g为重力加速度,x、y、z分别为飞行器在地面坐标系下的位置坐标,I x、I y、I z分别为飞行器在X、Y、Z轴上的转动惯量,u x=cosφsinθcosψ+sinφsinψ,u y=cosφsinθsinψ-sinφcosψ,u 1、u 2、u 3、u 4为输出控制量。
S2:根据多层零化神经动力学方法,设计无人飞行器的控制器;
由垂直高度z出发,设计关于输出控制量u 1的偏差函数,根据该偏差函数,设计多旋翼无人飞行器高度控制器,求解出u 1,再由水平位置x、y出发,设计关于u x、u y的偏差函数以及相应的多旋翼无人飞行器位置控制器,反解出目标姿态角度φ T和θ T,根据目标姿态角度,分别由横滚角φ、俯仰角θ以及偏航角ψ出发,设计关于输出控制量u 2~u 4的偏差函数,并设计相应的多层零化神经网络控制器,具体步骤如下:
针对垂直高度z,根据Z轴方向上的目标高度值z T以及实际高度值z,可以定义偏差函数为
e z1=z-z T (1)
并可以得到其导数为
Figure PCTCN2018112114-appb-000082
为了使实际值z能够收敛到目标值z T,根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000083
其中γ为一常数;
将等式(1)和(2)代入等式(3),整理得
Figure PCTCN2018112114-appb-000084
因为等式(4)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000085
并可以得到其导数为
Figure PCTCN2018112114-appb-000086
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000087
将等式(5)和(6)代入等式(7),整理得
Figure PCTCN2018112114-appb-000088
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000089
根据飞行器动力学方程,(9)可简化为
E z=a zu 1+b z (10)
其中
Figure PCTCN2018112114-appb-000090
并可以得到其导数为
Figure PCTCN2018112114-appb-000091
利用多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000092
将等式(10)和(11)代入等式(12)整理得
Figure PCTCN2018112114-appb-000093
针对水平位置x,根据X轴方向上的目标值x T以及实际值x,可以定义偏差函数
e x1=x-x T (14)
并可以得到其导数为
Figure PCTCN2018112114-appb-000094
为了使实际值x能够收敛到目标值x T,根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000095
将等式(14)和(15)代入等式(16)整理得
Figure PCTCN2018112114-appb-000096
因为等式(17)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000097
并可以得到其导数为
Figure PCTCN2018112114-appb-000098
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000099
将等式(18)和(19)代入等式(20)整理得
Figure PCTCN2018112114-appb-000100
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000101
根据飞行器动力学方程,等式(22)可化简为
E x=a xu x+b x (23)
其中
Figure PCTCN2018112114-appb-000102
并可以得到其导数为
Figure PCTCN2018112114-appb-000103
利用多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000104
将等式(23)和(24)代入等式(25)整理得
Figure PCTCN2018112114-appb-000105
针对水平位置y,根据Y轴方向上的目标值y T以及实际值y,可以定义偏差函数
e y1=y-y T (27)
并可以得到其导数为
Figure PCTCN2018112114-appb-000106
为了使实际值y能够收敛到目标值y T,根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000107
将等式(27)和(28)代入等式(29)整理得
Figure PCTCN2018112114-appb-000108
因为等式(30)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000109
并可以得到其导数为
Figure PCTCN2018112114-appb-000110
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000111
将等式(31)和(32)代入等式(33)整理得
Figure PCTCN2018112114-appb-000112
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000113
根据飞行器动力学方程,等式(35)可化简为
E y=a yu y+b y (36)
其中
Figure PCTCN2018112114-appb-000114
并可以得到其导数为
Figure PCTCN2018112114-appb-000115
利用多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000116
将等式(36)和(37)代入等式(38)整理得
Figure PCTCN2018112114-appb-000117
由等式(26)和(39)可求解出u x、u y,又根据
Figure PCTCN2018112114-appb-000118
可反解出目标角度值φ T和θ T
Figure PCTCN2018112114-appb-000119
针对横滚角φ,根据(40)中求解出的目标角度φ T以及实际角度φ,可以定义偏差函数为
e φ1=φ-φ T (41)
并可以得到其导数为
Figure PCTCN2018112114-appb-000120
为了使实际值φ能够收敛到目标值φ T,根据多层零化神经动力学方法, 可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000121
将等式(41)和(42)代入等式(43)整理得
Figure PCTCN2018112114-appb-000122
因为等式(44)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000123
并可以得到其导数为
Figure PCTCN2018112114-appb-000124
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000125
将等式(45)和(46)代入等式(47)整理得
Figure PCTCN2018112114-appb-000126
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000127
根据飞行器动力学方程,等式(49)可简化为
E φ=a φu 2+b φ (50)
其中
Figure PCTCN2018112114-appb-000128
并可以得到其导数为
Figure PCTCN2018112114-appb-000129
根据多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000130
将等式(50)和(51)代入等式(52)整理得
Figure PCTCN2018112114-appb-000131
针对俯仰角θ,根据等式(40)中求解出的目标角度θ T以及实际角度θ,可以定义偏差函数为
e θ1=θ-θ T (54)
并可以得到其导数为
Figure PCTCN2018112114-appb-000132
为了使实际值θ能够收敛到目标值θ T,根据多层零化神经动力学方法, 可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000133
将等式(54)和(55)代入等式(56)整理得
Figure PCTCN2018112114-appb-000134
因为等式(57)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000135
并可以得到其导数为
Figure PCTCN2018112114-appb-000136
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000137
将等式(58)和(59)代入等式(60)整理得
Figure PCTCN2018112114-appb-000138
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000139
根据飞行器动力学方程,等式(62)可简化为
E θ=a θu 3+b θ (63)
其中
Figure PCTCN2018112114-appb-000140
并可以得到其导数为
Figure PCTCN2018112114-appb-000141
根据多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000142
将等式(63)和(64)代入等式(65)整理得
Figure PCTCN2018112114-appb-000143
针对偏航角ψ,根据(40)中求解出的目标角度ψ T以及实际角度ψ,可以定义偏差函数为
e ψ1=ψ-ψ T (67)
并可以得到其导数为
Figure PCTCN2018112114-appb-000144
为了使实际值ψ能够收敛到目标值ψ T,根据多层零化神经动力学方法, 可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000145
将等式(67)和(68)代入等式(69)整理得
Figure PCTCN2018112114-appb-000146
因为等式(70)在初始情况下一般不成立且不包含输出控制量的相关信息,无法实现对控制量的求解,故需要进一步设计,于是定义
Figure PCTCN2018112114-appb-000147
并可以得到其导数为
Figure PCTCN2018112114-appb-000148
根据多层零化神经动力学方法,可以设计基于偏差函数的神经动力学方程为
Figure PCTCN2018112114-appb-000149
将等式(71)和(72)代入等式(73)整理得
Figure PCTCN2018112114-appb-000150
据此,可定义偏差函数
Figure PCTCN2018112114-appb-000151
根据飞行器动力学方程,等式(75)可简化为
E ψ=a ψu 4+b ψ (76)
其中
Figure PCTCN2018112114-appb-000152
并可以得到其导数为
Figure PCTCN2018112114-appb-000153
根据多层零化神经动力学方法,可以设计
Figure PCTCN2018112114-appb-000154
将等式(76)和(77)代入等式(78)整理得
Figure PCTCN2018112114-appb-000155
S3:利用获取的飞行器实时运行数据与目标姿态数据,通过设计的多层零化神经网络控制器求解飞行器电机的输出控制量;
根据多层零化神经网络等式(13)、(53)、(66)和(79),可得到无人飞行器控制器,该控制器能够通过网络结构化实现;该无人飞行器控制器能够控制无人飞行器稳定飞行,其中该控制器可以写成如下的形式:
Figure PCTCN2018112114-appb-000156
通过该控制器微分方程构建零化神经网络,通过零化神经网络解算无人飞行器控制量。
S4:将步骤4)的求解结果传递给飞行器电机调速器,根据控制器求解的得到的控制量与多旋翼无人机电机动力的关系,实现对电机动力的控制,控制无人飞行器的运动;
根据无人机动力分配方案,控制器求解得到的控制量与多旋翼无人机电机动力存在如下关系:
U=WF
其中U=[u 1 u 2 u 3 u 4] T为无人飞行器的控制量,F=[F 1 … F j] T为无人飞行器的电机动力,j为多旋翼无人机的电机个数,W为无人飞行器动力分配矩阵。
为了得到相应的电机所需得到的动力的多少,可以通过矩阵求逆或求伪逆的形式得到相应的电机动力F,也即
F=W -1U
若矩阵W为方阵且可逆,则W -1通过求逆运算得到,若W不为方阵,则通过相应的伪逆运算求解W -1;最终得到所需电机动力F并根据电机电压与动力的关系控制电机输入电压以控制电机转速,最终实现对电机动力的控制,完成无人机稳定飞行控制。由于不同的旋翼数目与结构会影响多旋翼无人机的控制方式,所以矩阵W根据不同结构与旋翼数目会有不同的形式。
以六旋翼无人飞行器为例,其动力分配存在如下关系:
Figure PCTCN2018112114-appb-000157
该关系可以进一步写成
Figure PCTCN2018112114-appb-000158
由于上述关系中W不是方阵,可以通过伪逆方式得到W -1,也即
Figure PCTCN2018112114-appb-000159
由此可以得到六旋翼无人飞行器的动力分配情况并得到相应的电机实际控制量用于控制电机运转。
以上所述仅为本发明优选的实施例,但本发明专利的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本发明专利所公开的范围内,根据本发明专利的发明构思或者技术方案加以等同替换或改变,都属于本发明专利的保护范围。

Claims (9)

  1. 一种无人飞行器稳定飞行控制方法,其特征在于:所述方法包括:
    1)获取飞行器自身的飞行实时运行数据,对飞行器的运动学问题进行解析处理,建立飞行器动力学模型;
    2)根据步骤1)获取的飞行实时运行数据与目标姿态数据,构建偏差函数;利用多层零化神经动力学方法,构建基于偏差函数的神经动力学方程;所有参数对应的基于偏差函数的神经动力学方程,共同构成无人飞行器的控制器,控制器的微分方程解算的输出量为飞行器电机的输出控制量;
    3)根据步骤2)求解得到的输出控制量与多旋翼无人机电机动力的关系,控制电机动力,完成无人飞行器运动的控制,具体步骤为:
    根据无人机动力分配方案,控制器求解得到的控制量与多旋翼无人机电机动力存在如下关系:
    U=WF
    其中U=[u 1 u 2 u 3 u 4] T为无人飞行器的输出控制量,F=[F 1 … F j] T为无人飞行器的电机动力,j为多旋翼无人机的电机个数,W为无人飞行器动力分配矩阵,矩阵W根据不同结构与旋翼数目会有不同的形式,需要根据其结构和旋翼数确定;
    通过矩阵求逆或求伪逆的形式得到相应的电机动力F,也即:
    F=W -1U
    若矩阵W为方阵且可逆,则W -1通过求逆运算得到,若W不为方阵,则通过相应的伪逆运算求解W -1;最终得到所需电机动力F并根据电机电压与动力的关系控制电机输入电压以控制电机转速,最终实现对电机动力的控制,完成无人机稳定飞行控制。
  2. 根据权利要求1所述的一种无人飞行器稳定飞行控制方法,其特征在于:所述通过所搭载的处理器对飞行器的运动学问题进行解析处理,具体包括:
    定义地面坐标系E和机体坐标系B,地面坐标系和机体坐标系通过转换矩阵R建立联系:E=RB,R表示为
    Figure PCTCN2018112114-appb-100001
    其中φ为横滚角,θ为俯仰角,ψ为偏航角;
    忽略飞行器所受空气阻力作用,在机体坐标系下,飞行器系统受力分 析如下
    Figure PCTCN2018112114-appb-100002
    其中m为飞行器的总质量,I 3×3为单位矩阵,I为惯性矩阵,V为机体坐标系下的线速度,ω为机体坐标系下的角速度,F为合外力,τ为合力矩。
  3. 根据权利要求2所述的一种无人飞行器稳定飞行控制方法,其特征在于:所述建立飞行器动力学模型,具体包括:
    根据定义的地面坐标系E和机体坐标系B、两者通过转换矩阵R建立的联系:E=RB以及在机体坐标系下,飞行器系统的受力分析,得到多旋翼飞行器的动力学方程为
    Figure PCTCN2018112114-appb-100003
    其中l为臂长,g为重力加速度,x、y、z分别为飞行器在地面坐标系下的位置坐标,
    Figure PCTCN2018112114-appb-100004
    分别表示x(t)、y(t)、z(t)的二阶导数,φ、θ、ψ分别表示横滚角、俯仰角以及偏航角,
    Figure PCTCN2018112114-appb-100005
    分别表示对应参数的二阶导数,
    Figure PCTCN2018112114-appb-100006
    分别表示对应参数的一阶导数,I x、I y、I z分别为飞行器在X、Y、Z轴上的转动惯量,u x=cosφsinθcosψ+sinφsinψ,u y=cosφsinθsinψ-sinφcosψ,u 1、u 2、u 3、u 4为输出控制量。
  4. 根据权利要求1所述的一种无人飞行器稳定飞行控制方法,其特征在于:所述根据多层零化神经动力学方法,设计无人飞行器的控制器,具体包括:
    (2-1)过多层零化神经动力学方法,由垂直高度z出发,设计关于输出控 制量u 1的偏差函数,根据该偏差函数,设计无人机高度控制器;
    (2-2)通过多层零化神经动力学方法,由水平位置x、y出发,设计关于u x、u y的偏差函数,根据该偏差函数,设计无人机位置控制器,再反解出目标姿态角度φ T和θ T
    (2-3)通过多层零化神经动力学方法,分别由横滚角φ、俯仰角θ以及偏航角ψ出发,设计输出控制量u 2~u 4的偏差函数,根据该偏差函数,设计姿态控制器。
  5. 根据权利要求4所述的一种无人飞行器稳定飞行控制方法,其特征在于:所述通过多层零化神经动力学方法,由垂直高度z出发,设计输出控制量u 1的偏差函数,根据该偏差函数,设计无人机高度控制器,具体包括:
    针对垂直高度z,根据Z轴方向上的目标高度值z T以及实际高度值z,定义偏差函数为
    Figure PCTCN2018112114-appb-100007
    根据飞行器动力学方程,(9)简化为
    E z=a zu 1+b z      (10)
    其中
    Figure PCTCN2018112114-appb-100008
    γ为一常数,并得到其导数为
    Figure PCTCN2018112114-appb-100009
    利用多层零化神经动力学方法,设计
    Figure PCTCN2018112114-appb-100010
    将等式(10)和(11)代入等式(12)整理得
    Figure PCTCN2018112114-appb-100011
  6. 根据权利要求4所述的一种无人飞行器稳定飞行控制方法,其特征在于:所述设计关于u x、u y的偏差函数,以及无人机位置控制器,具体包括:
    针对水平位置x,根据X轴方向上的目标值x T以及实际值x,定义偏差函数
    Figure PCTCN2018112114-appb-100012
    根据飞行器动力学方程,等式(22)化简为
    E x=a xu x+b x     (23)
    其中
    Figure PCTCN2018112114-appb-100013
    并得到其导数为
    Figure PCTCN2018112114-appb-100014
    利用多层零化神经动力学方法,设计
    Figure PCTCN2018112114-appb-100015
    将等式(23)和(24)代入等式(25)整理得
    Figure PCTCN2018112114-appb-100016
    针对水平位置y,根据Y轴方向上的目标值y T以及实际值y,定义偏差函数
    Figure PCTCN2018112114-appb-100017
    根据飞行器动力学方程,等式(35)化简为
    E y=a yu y+b y      (36)
    其中
    Figure PCTCN2018112114-appb-100018
    并得到其导数为
    Figure PCTCN2018112114-appb-100019
    利用多层零化神经动力学方法,设计
    Figure PCTCN2018112114-appb-100020
    将等式(36)和(37)代入等式(38)整理得
    Figure PCTCN2018112114-appb-100021
  7. 根据权利要求6所述的一种无人飞行器稳定飞行控制方法,其特征在于:所述反解目标姿态角度φ T和θ T,其计算公式为:
    由位置控制器等式(26)和(39)求解出的u x、u y
    Figure PCTCN2018112114-appb-100022
    所以反解出的目标角度值φ T和θ T
    Figure PCTCN2018112114-appb-100023
  8. 根据权利要求4所述的一种无人飞行器稳定飞行控制方法,其特征在于:所述通过多层零化神经动力学方法,分别由横滚角φ、俯仰角θ以 及偏航角ψ出发,设计关于输出控制量u 2~u 4的偏差函数,根据该偏差函数,设计姿态控制器,具体包括:
    针对横滚角φ,根据(40)中求解出的目标角度φ T以及实际角度φ,定义偏差函数为
    Figure PCTCN2018112114-appb-100024
    根据飞行器动力学方程,等式(49)简化为
    E φ=a φu 2+b φ      (50)
    其中
    Figure PCTCN2018112114-appb-100025
    并得到其导数为
    Figure PCTCN2018112114-appb-100026
    根据多层零化神经动力学方法,设计
    Figure PCTCN2018112114-appb-100027
    将等式(50)和(51)代入等式(52)整理得
    Figure PCTCN2018112114-appb-100028
    针对俯仰角θ,根据等式(40)中求解出的目标角度θ T以及实际角度θ,定义偏差函数为
    Figure PCTCN2018112114-appb-100029
    根据飞行器动力学方程,等式(62)简化为
    E θ=a θu 3+b θ      (63)
    其中
    Figure PCTCN2018112114-appb-100030
    并得到其导数为
    Figure PCTCN2018112114-appb-100031
    根据多层零化神经动力学方法,设计
    Figure PCTCN2018112114-appb-100032
    将等式(63)和(64)代入等式(65)整理得
    Figure PCTCN2018112114-appb-100033
    针对偏航角ψ,根据人为设定的角度ψ T以及实际角度ψ,定义偏差函数为
    Figure PCTCN2018112114-appb-100034
    根据飞行器动力学方程,等式(75)简化为
    E ψ=a ψu 4+b ψ      (76)
    其中
    Figure PCTCN2018112114-appb-100035
    并得到其导数为
    Figure PCTCN2018112114-appb-100036
    根据多层零化神经动力学方法,设计
    Figure PCTCN2018112114-appb-100037
    将等式(76)和(77)代入等式(78)整理得
    Figure PCTCN2018112114-appb-100038
  9. 根据权利要求4所述的一种无人飞行器稳定飞行控制方法,其特征在于:所述根据所设计的高度控制器、位置控制器以及姿态控制器,共同构成的多旋翼无人飞行器的稳定飞行器,具体包括:
    Figure PCTCN2018112114-appb-100039
    Figure PCTCN2018112114-appb-100040
    Figure PCTCN2018112114-appb-100041
    Figure PCTCN2018112114-appb-100042
    根据等式(13)、(53)、(66)和(79),得到无人飞行器控制器,该控制器能够通过网络结构化实现;该无人飞行器控制器能够控制无人飞行器稳定飞行;其中该控制器写成如下的形式:
    Figure PCTCN2018112114-appb-100043
    通过该控制器微分方程构建多层零化神经网络,通过多层零化神经网络解算无人飞行器控制量。
PCT/CN2018/112114 2017-11-01 2018-10-26 一种无人飞行器稳定飞行控制方法 WO2019085834A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/652,457 US11721219B2 (en) 2017-11-01 2018-10-26 Method for controlling steady flight of unmanned aircraft

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201711059466.9A CN107957730B (zh) 2017-11-01 2017-11-01 一种无人飞行器稳定飞行控制方法
CN201711059466.9 2017-11-01

Publications (1)

Publication Number Publication Date
WO2019085834A1 true WO2019085834A1 (zh) 2019-05-09

Family

ID=61963496

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/112114 WO2019085834A1 (zh) 2017-11-01 2018-10-26 一种无人飞行器稳定飞行控制方法

Country Status (3)

Country Link
US (1) US11721219B2 (zh)
CN (1) CN107957730B (zh)
WO (1) WO2019085834A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110254741A (zh) * 2019-05-17 2019-09-20 李泽波 一种飞行控制系统的设计方法
CN110398971A (zh) * 2019-08-07 2019-11-01 大连海事大学 一种直流电机推进无人船的速度控制方法
CN110989649A (zh) * 2019-12-26 2020-04-10 中国航空工业集团公司沈阳飞机设计研究所 面向高机动固定翼无人机的飞行动作控制装置及训练方法
CN113520413A (zh) * 2021-08-25 2021-10-22 长春工业大学 一种基于表面肌电信号的下肢多关节角度估计方法
CN113791638A (zh) * 2021-08-29 2021-12-14 西北工业大学 一种多无人飞行器协同绳系吊装运输系统的稳定控制方法
CN114115322A (zh) * 2021-12-15 2022-03-01 西北工业大学 一种绳系飞行器系统的跟踪控制方法

Families Citing this family (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107957730B (zh) * 2017-11-01 2020-02-18 华南理工大学 一种无人飞行器稳定飞行控制方法
CN108717265B (zh) * 2018-05-30 2021-05-18 重庆邮电大学 一种基于控制变量参数化的无人飞行器巡航跟踪控制系统及控制方法
CN108958289B (zh) * 2018-07-28 2021-08-13 天津大学 基于相对速度障碍的集群无人机避碰方法
CN110806688B (zh) * 2019-11-20 2022-12-09 中山大学 一种时变非线性质量弹簧阻尼系统的运动控制方法
CN111857171B (zh) * 2020-07-30 2021-09-21 华南理工大学 一种基于质量自适应多层神经网络的无人机控制方法
CN112706165A (zh) * 2020-12-22 2021-04-27 中山大学 一种面向轮式移动机械臂的跟踪控制方法及系统
CN116848485A (zh) * 2021-03-15 2023-10-03 深圳市大疆创新科技有限公司 无人飞行器控制方法、无人飞行器和存储介质
CN113359440B (zh) * 2021-05-24 2023-06-20 华南理工大学 基于多层积分神经动力学的多旋翼无人机控制器设计方法
CN113485437B (zh) * 2021-07-29 2022-03-04 合肥工业大学 一种抗干扰无人机控制方法及系统
CN114594675B (zh) * 2022-03-07 2024-06-04 东华理工大学长江学院 一种改进型pid的四旋翼飞行器控制系统及方法
KR102750447B1 (ko) * 2022-04-21 2025-01-09 한국과학기술원 무인 항공기용 뉴럴 네트워크 시스템, 이를 포함하는 무인 항공기 제어 시스템 및 이를 이용한 무인 항공기 제어 방법
CN114721273B (zh) * 2022-04-22 2024-04-12 湖南师范大学 一种固定时间收敛零化神经网络的多智能体编队控制方法
CN117666368B (zh) * 2024-02-02 2024-04-16 国网湖北省电力有限公司 基于物联网的无人机多机协同的作业方法及系统
CN118464024B (zh) * 2024-07-12 2025-01-07 广东海洋大学 基于零值化神经网络的auv轨迹追踪方法
CN118550313B (zh) * 2024-07-26 2024-12-06 湖南大学 一种绳索牵引飞行机器人抗扰动控制方法
CN119024860A (zh) * 2024-08-15 2024-11-26 苏州云锋航空科技有限公司 一种倾转机翼eVTOL的控制分配方法

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6092919A (en) * 1995-08-01 2000-07-25 Guided Systems Technologies, Inc. System and method for adaptive control of uncertain nonlinear processes
CN104932512A (zh) * 2015-06-24 2015-09-23 北京科技大学 一种基于mimo非线性不确定反步法的四旋翼位姿控制方法
CN106155076A (zh) * 2016-08-23 2016-11-23 华南理工大学 一种多旋翼无人飞行器的稳定飞行控制方法
CN106444809A (zh) * 2016-10-12 2017-02-22 湖南绿野航空科技有限公司 一种无人机飞行控制器
CN106502262A (zh) * 2015-09-08 2017-03-15 中国农业机械化科学研究院 一种农用无人机飞行平台及其控制系统和控制方法
CN107264804A (zh) * 2017-05-12 2017-10-20 华南农业大学 一种基于gps的无人飞行器变量喷雾控制装置与方法
CN107957730A (zh) * 2017-11-01 2018-04-24 华南理工大学 一种无人飞行器稳定飞行控制方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9958874B2 (en) * 2014-05-30 2018-05-01 SZ DJI Technology Co., Ltd Aircraft attitude control methods
CN105676641B (zh) * 2016-01-25 2018-10-16 南京航空航天大学 基于反步和滑模控制的非线性鲁棒控制器的设计方法
CN106647781B (zh) * 2016-10-26 2019-09-06 广西师范大学 基于重复控制补偿神经模糊pid四旋翼飞行器的控制方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6092919A (en) * 1995-08-01 2000-07-25 Guided Systems Technologies, Inc. System and method for adaptive control of uncertain nonlinear processes
CN104932512A (zh) * 2015-06-24 2015-09-23 北京科技大学 一种基于mimo非线性不确定反步法的四旋翼位姿控制方法
CN106502262A (zh) * 2015-09-08 2017-03-15 中国农业机械化科学研究院 一种农用无人机飞行平台及其控制系统和控制方法
CN106155076A (zh) * 2016-08-23 2016-11-23 华南理工大学 一种多旋翼无人飞行器的稳定飞行控制方法
CN106444809A (zh) * 2016-10-12 2017-02-22 湖南绿野航空科技有限公司 一种无人机飞行控制器
CN107264804A (zh) * 2017-05-12 2017-10-20 华南农业大学 一种基于gps的无人飞行器变量喷雾控制装置与方法
CN107957730A (zh) * 2017-11-01 2018-04-24 华南理工大学 一种无人飞行器稳定飞行控制方法

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110254741A (zh) * 2019-05-17 2019-09-20 李泽波 一种飞行控制系统的设计方法
CN110398971A (zh) * 2019-08-07 2019-11-01 大连海事大学 一种直流电机推进无人船的速度控制方法
CN110398971B (zh) * 2019-08-07 2022-04-01 大连海事大学 一种直流电机推进无人船的速度控制方法
CN110989649A (zh) * 2019-12-26 2020-04-10 中国航空工业集团公司沈阳飞机设计研究所 面向高机动固定翼无人机的飞行动作控制装置及训练方法
CN110989649B (zh) * 2019-12-26 2023-07-25 中国航空工业集团公司沈阳飞机设计研究所 面向高机动固定翼无人机的飞行动作控制装置及训练方法
CN113520413A (zh) * 2021-08-25 2021-10-22 长春工业大学 一种基于表面肌电信号的下肢多关节角度估计方法
CN113791638A (zh) * 2021-08-29 2021-12-14 西北工业大学 一种多无人飞行器协同绳系吊装运输系统的稳定控制方法
CN113791638B (zh) * 2021-08-29 2023-06-30 西北工业大学 一种多无人飞行器协同绳系吊装运输系统的稳定控制方法
CN114115322A (zh) * 2021-12-15 2022-03-01 西北工业大学 一种绳系飞行器系统的跟踪控制方法
CN114115322B (zh) * 2021-12-15 2023-06-30 西北工业大学 一种绳系飞行器系统的跟踪控制方法

Also Published As

Publication number Publication date
CN107957730B (zh) 2020-02-18
US11721219B2 (en) 2023-08-08
CN107957730A (zh) 2018-04-24
US20220036739A1 (en) 2022-02-03

Similar Documents

Publication Publication Date Title
CN107957730B (zh) 一种无人飞行器稳定飞行控制方法
CN107368091B (zh) 一种基于有限时间神经动力学的多旋翼无人飞行器的稳定飞行控制方法
CN106155076B (zh) 一种多旋翼无人飞行器的稳定飞行控制方法
Lozano Unmanned aerial vehicles: Embedded control
Flores et al. Quad-tilting rotor convertible mav: Modeling and real-time hover flight control
CN205891228U (zh) 一种飞行机器人
Bapst et al. Design and implementation of an unmanned tail-sitter
CN109991991B (zh) 一种无人直升机鲁棒容错跟踪方法
CN106249745A (zh) 四轴无人机的控制方法
Rodić et al. Modeling and simulation of quad-rotor dynamics and spatial navigation
CN105549617B (zh) 一种仿鸟群行为和虚拟结构的无人机编队控制方法
Beyeler et al. 3D vision-based navigation for indoor microflyers
Rodić et al. The modeling and simulation of an autonomous quad-rotor microcopter in a virtual outdoor scenario
Jithu et al. Quadrotor modelling and control
CN108121354A (zh) 基于指令滤波反步法的四旋翼无人机稳定跟踪控制方法
CN110231828A (zh) 基于nftsm的四旋翼无人机视觉伺服控制方法
Chen Research on AI application in the field of quadcopter UAVs
CN109308074A (zh) 一种无人机重心偏移的补偿方法及系统
CN111857171B (zh) 一种基于质量自适应多层神经网络的无人机控制方法
CN118778673A (zh) 一种鲁棒的固定翼无人机飞控系统及方法
Barve et al. Modelling, simulation and altitude-range-analysis of quad-copter uav
CN102706360A (zh) 一种利用光流传感器和速率陀螺对飞行器状态估计的方法
Itaketo et al. Dynamic modeling and performance analysis of an autonomous quadrotor using linear and nonlinear control techniques
Tahar et al. Control of an under-actuated x4-flyer using integral backstepping controller
CN119611832B (zh) 一种基于舵面控制的全驱动共轴旋翼无人机及其控制方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18871836

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 14/08/2020)

122 Ep: pct application non-entry in european phase

Ref document number: 18871836

Country of ref document: EP

Kind code of ref document: A1