CN104133379A - Simulation method for four-rotor aircraft - Google Patents

Abstract

The invention provides a simulation method for a four-rotor aircraft, and relates to a real-time simulation method of an unmanned aircraft. The simulation method for the four-rotor aircraft solves the problems that an existing semi-physical simulation system is huge in size, complex in connection, low in accuracy due to the fact that virtual state variables exist in simulation, and low in accuracy of a model obtained through modeling applying a controller design method based on the models and a numerical simulation system mechanism, wherein the controller design method based on the models and the numerical simulation system mechanism cannot be applied to modeling easily. According to the simulation method for the four-rotor aircraft, the actual aircraft is regarded as the controlled object and connected to a simulation circuit, and a reference is provided for the numerical simulation models obtained by mechanism modeling in MATLAB/Simulink, so that accuracy of aircraft mechanism modeling can be verified rapidly. A parameter recognition method is adopted so as to guarantee accuracy of objective models in numerical simulation, so that feasible controllers are verified in the numerical simulation and the controllers can be finally effectively applied to actual physical objects. The simulation method for the four-rotor aircraft is specifically applied to the field of four-rotor aircraft simulation.

Description

Quadrotor emulation mode

Technical field

The present invention relates to a kind of unmanned vehicle real-time emulation method.

Background technology

Quadrotor has the driving of owing, strong coupling and nonlinear feature in dynamics, and these have all increased the difficulty of Flight Controller Design.

In the design phase of controller, conventionally have two kinds of modes can go access control device performance: the first is flight in kind, and this verification method is with a high credibility, but have efficiency low, have a big risk, shortcoming that cost is high; Second method is numerical simulation; this method efficiency is high, cost is low, even if but conventionally can be because parameter estimation in numerical model be inaccurate, modeling is not dynamically and do not consider that the factors such as disturbance cause verifying in numerical simulation that feasible controller is also difficult to be applied on actual aircraft.

Hardware-in-the-loop simulation, is called again the emulation (Hardware in the Loop Simulation) of hardware in loop, refers to access part real-time simulation in kind in the emulation loop of simulating experimental system.Real-time is the prerequisite of carrying out hardware-in-the-loop simulation.Hardware-in-the-loop simulation is compared and is had the possibility that realizes economically higher validity with the emulation mode of other type, has the advantage of fast construction and deployment control algolithm simultaneously.Viewpoint from system, hardware-in-the-loop simulation allows access part material object in system, mean and part material object can be placed in system and investigate, thereby parts can be checked in the environment that meets entire system performance index, so hardware-in-the-loop simulation is to improve the reliability of system and the necessary means of development quality.

A kind of means that realize hardware-in-the-loop simulation are to adopt dSPACE real-time emulation system, it is a set of control system based on MATLAB/Simulink by the exploitation of German dSPACE company exploitation and test environment under real time environment, realized the seamless link of actual physics system and MATLAB/Simulink, but expensive.

The visual semi-matter simulating system based on xPC of another kind of supported data wants two computing machines (host, target machine) just can complete emulation, connects complicated cost higher.

A kind of Hardware In The Loop Simulation Method of quadrotor is fixed on aircraft on a universal joint being connected with ground, has limited its three degree of freedom in space translation, thereby has guaranteed the security of its flight; But this method has been used virtual location status, simulation nicety is not high, and and is not suitable for checking height control algolithm.

Summary of the invention

The present invention is in large scale in order to solve existing semi-matter simulating system, connect accuracy complicated, that emulation exists virtual state variable to cause low, be difficult to the not high problem of model accuracy that the controller design method of application based on model and Numeral Emulation System modelling by mechanism obtain, the invention provides a kind of quadrotor emulation mode.

Quadrotor Hardware In The Loop Simulation Method, it is realized based on following apparatus, and this device comprises simulation computer, airborne microcontroller, No. 1 wireless transceiver, No. 2 wireless transceivers and airborne sensor;

The data-signal input end of airborne microcontroller is connected with the data-signal output terminal of airborne sensor, the communication terminal of airborne microcontroller is connected with the signal input output end of No. 2 wireless transceivers, No. 2 wireless transceivers carry out data transmission by mode and No. 1 wireless transceiver of radio communication, the signal input output end of No. 1 wireless transceiver is connected with the signal input output end of simulation computer

Simulation computer is embedded in Matlab/Simulink emulation module, and described Matlab/Simulink emulation module comprises data input module, state estimation and filtration module, attitude and height control module, state of flight display module, data outputting module, hardware-in-the-loop simulation data recordin module and steering order generation module;

The detailed process of this Hardware In The Loop Simulation Method is:

First, by the data-signal of airborne microcontroller harvester set sensor output, and airborne microcontroller is delivered to the data-signal collecting the data input module of simulation computer by No. 2 wireless transceivers, No. 1 wireless transceiver successively, this data-signal comprises range finding reading, 3-axis acceleration reading, three axis angular rate readings and three axle magnetic induction density signals;

By data input module, receive the data-signal of input, and these data are delivered to state estimation and filtration module, state estimation and filtration module utilize the method for complementary filter to obtain two groups of data according to the data-signal receiving, first group of data in these two groups of data are attitude Eulerian angle estimated value and Height Estimation value, the estimated value of speed on the estimated value that second group of data are angular velocity and vertical direction

And first group of data delivered to state of flight display module, and state of flight display module is used for showing in real time the state of flight of quadrotor,

First group and second group of data are delivered to attitude and height control module, attitude and the steering order that highly control module produces according to steering order generation module, calculate and obtain attitude and height control signal, then through controlling the expectation rotating speed that distributes four motors that obtain quadrotor, the expectation rotating speed of these four motors is delivered to airborne microcontroller for controlling electron speed regulator and the motor of quadrotor successively by data outputting module after No. 1 wireless transceiver, No. 2 wireless transceivers

Hardware-in-the-loop simulation data recordin module is estimated the expectation rotating speed with two groups of data of filtration module output and four motors of quadrotor for store status.

Described airborne sensor comprises three axle electronic compasss, six axis movement sensors and ultrasonic range finder sensor,

Three axle electronic compasss are used for gathering three axle magnetic induction density signals,

Six axis movement sensors are used for gathering 3-axis acceleration reading and three axis angular rate readings,

Ultrasonic range finder sensor is used for gathering range finding reading.

The integrated circuit that the employing model of three described axle electronic compasss is HMC5883L is realized, and the integrated circuit that the model of six axis movement sensors is MPU6050 is realized, and ultrasonic range finder sensor adopts the integrated circuit that model is US-100 to realize.

Described airborne microcontroller adopts TMS320F28335 type DSP to realize.

Described attitude with height control module optimizing process be,

In Model Distinguish subsystem, parameter identification device receives the expectation rotating speed Ω of four motors of quadrotor ₁, Ω ₂, Ω ₃, Ω ₄, attitude Eulerian angle estimated value height Estimation value the estimated value of angular velocity estimated value with speed on vertical direction this parameter identification device, according to model structure to be identified, is tried to achieve the identification result of parameter vector, and then has determined aircraft movements and kinetic model, thereby completes the optimization with height control module to attitude according to numerical simulation subsystem,

Described model structure to be identified is described by following state space equation:

$\left\{\begin{array}{c}\stackrel{\·}{\mathrm{\φ}}=p\cos \mathrm{\θ}+r\sin \mathrm{\θ}\\ \stackrel{\·}{\mathrm{\θ}}=q-r\cos \mathrm{\θ}\tan \mathrm{\φ}+p\sin \mathrm{\θ}\tan \mathrm{\φ}\\ \stackrel{\·}{\mathrm{\ψ}}=r\cos \mathrm{\θ}\sec \mathrm{\φ}-p\sin \mathrm{\θ}\sec \mathrm{\φ}\\ \stackrel{\·}{p}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{yy}}-I_{\mathrm{zz}}}{I_{\mathrm{xx}}}\right)-\stackrel{\·}{\mathrm{\θ}}\frac{J_r}{I_{\mathrm{xx}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{xx}}}{U}_2\\ \stackrel{\·}{q}=\stackrel{\·}{\mathrm{\φ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{zz}}-I_{\mathrm{xx}}}{I_{\mathrm{yy}}}\right)+\stackrel{\·}{\mathrm{\φ}}\frac{J_r}{I_{\mathrm{yy}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{yy}}}{U}_3\\ \stackrel{\·}{r}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\φ}}\left(\frac{I_{\mathrm{xx}}-I_{\mathrm{yy}}}{I_{\mathrm{zz}}}\right)+\frac{1}{I_{\mathrm{zz}}}{U}_4\\ \stackrel{\·}{z}=w\\ \stackrel{\·}{w}=g-\left(\cos \mathrm{\φ}\cos \mathrm{\θ}\right)\frac{1}{m}{U}_1\end{array}\right.$ (formula one),

Wherein, $\left\{\begin{array}{c}{U}_1=b({{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ U_2=b(-{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_4^2)\\ U_3=b({{\mathrm{\Ω}}_1}^2-{\mathrm{\Ω}}_3^2)\\ U_4=d(-{{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2-{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ {\mathrm{\Ω}}_r={\mathrm{\Ω}}_2+{\mathrm{\Ω}}_4-{\mathrm{\Ω}}_1-{\mathrm{\Ω}}_3\end{array}\right.$ (formula two),

Described parameter identification device is according to model structure to be identified, and the detailed process that obtains the identification result of parameter vector is:

Expectation rotating speed Ω when four motors ₁, Ω ₂, Ω ₃, Ω ₄, during as the input of dummy vehicle, the numerical solution of trying to achieve part predicted state by formula one and formula two is

ξ(k)＝[φ(k),θ(k),ψ(k),p(k),q(k),r(k),z(k),w(k)] ^T，

Secondly, adopt cost function $V_N\left(r\right)=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}({\widehat{\mathrm{\ξ}}}^2\left(k\right)-{\mathrm{\ξ}}^2\left(k\right)),$ Ask for r ₀,

$r_0=\arg \underset{r}{\min }{V}_N\left(r\right),$

R ₀be the identification result of parameter vector, r is parameter vector to be identified, wherein, and r=[m, l, I _xx, I _yy, I _zz, J _r, b, d] ^t,

Parameter vector r to be identified comprises aircraft gross mass m, and relatively motor shaft is apart from l, body moment of inertia (I _xx, I _yy, I _zz), rotor and rotor be around motor shaft moment of inertia J _r, lift coefficient b, torque coefficient d, N is the sampling number of record; K is sampling time point, U ₁, U ₂, U ₃, U ₄and Ω _rbe intermediate variable, g is acceleration of gravity.

The numerical value emulation method of quadrotor, the detailed process of the method is,

First, parameter identification device receives from semi-matter simulating system and obtains data message, this parameter identification device is according to model structure to be identified, obtain the identification result of parameter vector, and by numerical simulation subsystem according to the identification result of the parameter vector obtaining, obtain aircraft movements and kinetic model, thereby complete the numerical simulation of quadrotor.

Described parameter identification device receives from semi-matter simulating system and obtains the expectation rotating speed Ω that data message comprises four motors of quadrotor ₁, Ω ₂, Ω ₃, Ω ₄, attitude Eulerian angle estimated value height Estimation value the estimated value of angular velocity estimated value with speed on vertical direction

Described model structure to be identified is described by following state space equation:

$\left\{\begin{array}{c}\stackrel{\·}{\mathrm{\φ}}=p\cos \mathrm{\θ}+r\sin \mathrm{\θ}\\ \stackrel{\·}{\mathrm{\θ}}=q-r\cos \mathrm{\θ}\tan \mathrm{\φ}+p\sin \mathrm{\θ}\tan \mathrm{\φ}\\ \stackrel{\·}{\mathrm{\ψ}}=r\cos \mathrm{\θ}\sec \mathrm{\φ}-p\sin \mathrm{\θ}\sec \mathrm{\φ}\\ \stackrel{\·}{p}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{yy}}-I_{\mathrm{zz}}}{I_{\mathrm{xx}}}\right)-\stackrel{\·}{\mathrm{\θ}}\frac{J_r}{I_{\mathrm{xx}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{xx}}}{U}_2\\ \stackrel{\·}{q}=\stackrel{\·}{\mathrm{\φ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{zz}}-I_{\mathrm{xx}}}{I_{\mathrm{yy}}}\right)+\stackrel{\·}{\mathrm{\φ}}\frac{J_r}{I_{\mathrm{yy}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{yy}}}{U}_3\\ \stackrel{\·}{r}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\φ}}\left(\frac{I_{\mathrm{xx}}-I_{\mathrm{yy}}}{I_{\mathrm{zz}}}\right)+\frac{1}{I_{\mathrm{zz}}}{U}_4\\ \stackrel{\·}{z}=w\\ \stackrel{\·}{w}=g-\left(\cos \mathrm{\φ}\cos \mathrm{\θ}\right)\frac{1}{m}{U}_1\end{array}\right.$ (formula one),

Wherein, $\left\{\begin{array}{c}{U}_1=b({{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ U_2=b(-{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_4^2)\\ U_3=b({{\mathrm{\Ω}}_1}^2-{\mathrm{\Ω}}_3^2)\\ U_4=d(-{{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2-{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ {\mathrm{\Ω}}_r={\mathrm{\Ω}}_2+{\mathrm{\Ω}}_4-{\mathrm{\Ω}}_1-{\mathrm{\Ω}}_3\end{array}\right.$ (formula two),

Described parameter identification device is according to model structure to be identified, and the detailed process that obtains the identification result of parameter vector is:

Expectation rotating speed Ω when four motors ₁, Ω ₂, Ω ₃, Ω ₄, during as the input of dummy vehicle, the numerical solution of trying to achieve part predicted state by formula one and formula two is ξ (k)=[φ (k), θ (k), ψ (k), p (k), q (k), r (k), z (k), w (k)] ^t,

Secondly, adopt cost function $V_N\left(r\right)=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}({\widehat{\mathrm{\ξ}}}^2\left(k\right)-{\mathrm{\ξ}}^2\left(k\right)),$ Ask for r ₀,

$r_0=\arg \underset{r}{\min }{V}_N\left(r\right),$

R ₀be the identification result of parameter vector, r is parameter vector to be identified, wherein, and r=[m, l, I _xx, I _yy, I _zz, J _r, b, d] ^t,

Parameter vector r to be identified comprises aircraft gross mass m, and relatively motor shaft is apart from l, body moment of inertia (I _xx, I _yy, I _zz), rotor and rotor be around motor shaft moment of inertia J _r, lift coefficient b, torque coefficient d, N is the sampling number of record; K is sampling time point, U ₁, U ₂, U ₃, U ₄and Ω _rbe intermediate variable, g is acceleration of gravity.

The present invention has set up a kind of quadrotor half physical varification platform, the access emulation loop using actual aircraft as controlled device, for the numerical simulation model of building up in MATLAB/Simulink provides a reference, aircraft numerical modeling accuracy can be verified fast, and then guarantee to verify that feasible controller finally can be applied on actual physical object effectively in numerical simulation.For dummy vehicle analysis and control parameter tuning provides a visual data display platform, improve the efficiency of designing and developing of state observer and controller simultaneously.

System comprises ingredient and function:

The simulation computer of MATLAB/Simulink is installed as the core of half physical varification system, it has moved mathematical model and the program of aircraft entity object and simulated environment, adopted stratification, modular modeling method, can be with the type of drive real time execution of data stream.

Real-time Windows Target (RTWT) product of MATLAB has been installed in simulation computer, it provides a Run-time engine to make the different module of Simulink realistic model under Windows system, to link and to connect in real time with hardware input/output board, the parameter regulatory function while simultaneously realizing the visual and model running of signal.

External schema allows Simulink engine as client (host), and the code that RTWT generates is as server (target), and the mode with shared drive between client and server communicates.Specifically, Simulink engine sends message request target and receives data and upload measurement data, and target is usingd the request of execution as responding.

Externally, under pattern, hardware-in-the-loop simulation model is the highest can be carried out with the frequency of 5KHz, and Simulink model has just become a graphic user interface, and not needing to recompilate just can adjustment model parameter.The realistic model operating in Simulink comprises:

(1) realistic model hardware interface comprises data input module and data outputting module, can complete respectively the binary coding of serial communication and the work of decoding, sets up data be connected with airborne microcontroller;

(2) numerical value object model, comprises kinematical equation, kinetics equation, the measurement model of sensor and the noise model of simulation of aircraft.Sensor comprises three axle electronic compasss, six axis movement sensors and ultrasonic range finder sensor.This part is corresponding with actual physical model, for the preliminary identification of filtering algorithm and control algolithm;

(3) state estimation and filtration module, according to the observation of magnetic vector and gravitational vector and the integration of three axle orientation read signals over the ground, utilize the method for complementary filter to obtain the estimation to attitude information.According to measured value, the estimated value of attitude and the measured value of acceleration of height, use equally the estimation of the method acquisition altitude rate of complementary filter.

(4) attitude and height control module, comprise that four take respectively roll angle, the angle of pitch, crab angle and be highly the PID controller of controlled variable, through controlling distribution, obtain the expectation rotating speed of four motors as output.

Formation control loop in (2) (3) (4) can be realized numerical simulation above, and above (1) (3) (4) form control loop and also coordinate actual physical model can complete hardware-in-the-loop simulation,

The airborne control panel of quadrotor has carried digital signal processor (DSP) TMS320F28335 for gathering flight attitude and positional information is passed to simulation computer and carried out the steering order that simulation computer sends.

The I of DSP ²c interface is connected with three axle electronic compass HMC5883L with six axis movement sensor MPU6050 (comprising three axis accelerometer and three-axis gyroscope) respectively.By UART, be connected in order to measuring height with ultrasound measurement module US-100.Simulation computer swap data by spi bus with operation Simulink: send the sensing data collecting according to the request of receiving to simulation computer, thereby generate pwm signal and issue electron speed regulator and control motor speed according to the control signal of receiving.In the incipient stage of emulation, thereby aircraft center of gravity should be fixed on to the security of warranty test on a universal joint.

Radio-Frequency Wireless Communication has partly comprised a NRF24L01 wireless transceiver connecting with DSP, with the USB wireless transceiver of a compatible NRF24L01 who is connected with simulation computer, has realized the wireless data exchange between simulation computer and DSP.

Quadrotor Hardware In The Loop Simulation Method proposed by the invention, can for the various Flight Control Algorithms that have been proved to be successful, in conjunction with practical object, carry out hardware-in-the-loop simulation intuitively in numerical simulation, state and the controlled quentity controlled variable change curve that can present truly four rotor unmanned aircrafts, again can online modification wave filter and controller parameter, the test result that has reference value can be provided for the design of embedded controller, can greatly reduce the R&D cycle, save flight experiment cost simultaneously.

The present invention is in simulation process, and its controlled device adopts entity four rotor unmanned aircrafts, but not the kinetic model of pure values form, attitude sensor can read real Flight Condition Data in real time, and simulated effect is pressed close to straight truth condition.

The present invention uses the disturbance response data of half physical varification system, the method for System Discrimination is applied in the correction work of numerical model, for the modelling by mechanism method of quadrotor provides effective verification tool.

Two data terminals of analogue system (simulation computer and platform in kind) are used Radio-Frequency Wireless Communication, closely in situation, can guarantee the exchanges data of fast and stable, have improved the dirigibility of emulation.

Because needed control cycle is in 200Hz, the real-time kernel that can be implanted in Windows that MATLAB provides can be competent at the data bandwidth of wanting required for the present invention completely, simulation computer involved in the present invention has completed the task that practical flight device model is controlled in real time when human-computer interaction function is provided, has saved the worry of considering special purpose computer cost and compatibility issue.

And the present invention improves the existing semi-matter simulating system based on xPC, the simulated program of operation is transferred in the real-time kernel that the MATLAB in host provides and is moved in above-mentioned target machine, both guaranteed the real-time of system operation, reduced again the scale of system hardware, reduce the complexity that module connects, improved reliability of operation.Because the present invention has adopted the hardware-in-the-loop simulation of full degree of freedom, aircraft is not limited on universal joint, so the data that emulation obtains have higher authenticity.

The present invention accesses emulation loop using actual aircraft as controlled device, for building up the numerical simulation model that modelling by mechanism obtains in MATLAB/Simulink, provide a reference, significant for the validity of confirming the numerical simulation model that modelling by mechanism obtains

For the control problem of quadrotor, the controller design method of data-driven (such as PID) does not rely on the parameter of concrete model, so the process of parameter identification does not have too large meaning.In the past few decades, the various novel controller design methods based on model continue to bring out.In order to utilize these methods, first must obtain plant model accurately.The realistic model obtaining with modelling by mechanism is compared, and parameter identification device proposed by the invention can obtain object model more accurately, and this application for the controller design method based on model provides necessary condition.

Accompanying drawing explanation

Fig. 1 is the principle schematic of quadrotor Hardware In The Loop Simulation Method of the present invention;

Fig. 2 is the principle schematic of the numerical value emulation method of the quadrotor described in embodiment six.

Embodiment

Embodiment one: present embodiment is described referring to Fig. 1, quadrotor Hardware In The Loop Simulation Method described in present embodiment, it is realized based on following apparatus, and this device comprises simulation computer 1,3, No. 2 wireless transceivers 4 of 2, No. 1 wireless transceiver of airborne microcontroller and airborne sensor 5;

The data-signal input end of airborne microcontroller 2 is connected with the data-signal output terminal of airborne sensor 5, the communication terminal of airborne microcontroller 2 is connected with the signal input output end of No. 2 wireless transceivers 4, No. 2 wireless transceivers 4 carry out data transmission by mode and No. 1 wireless transceiver 3 of radio communication, the signal input output end of No. 1 wireless transceiver 3 is connected with the signal input output end of simulation computer 1

Simulation computer 1 is embedded in Matlab/Simulink emulation module, and described Matlab/Simulink emulation module comprises data input module 1-1, state estimation and filtration module 1-2, attitude and height control module 1-3, state of flight display module 1-4, data outputting module 1-5, hardware-in-the-loop simulation data recordin module 1-6 and steering order generation module 1-7;

The detailed process of this Hardware In The Loop Simulation Method is:

First, by the data-signal of airborne microcontroller 2 harvester set sensor 5 outputs, and airborne microcontroller 2 is delivered to the data-signal collecting the data input module 1-1 of simulation computer 1 by 4, No. 1 wireless transceiver 3 of No. 2 wireless transceivers successively, this data-signal comprises range finding reading, 3-axis acceleration reading, three axis angular rate readings and three axle magnetic induction density signals;

By data input module 1-1, receive the data-signal of input, and these data are delivered to state estimation and filtration module 1-2, state estimation and filtration module 1-2 utilize the method for complementary filter to obtain two groups of data according to the data-signal receiving, first group of data in these two groups of data are attitude Eulerian angle estimated value and Height Estimation value, the estimated value of speed on the estimated value that second group of data are angular velocity and vertical direction

And first group of data delivered to state of flight display module 1-4, and state of flight display module 1-4 is used for showing in real time the state of flight of quadrotor,

First group and second group of data are delivered to attitude and height control module 1-3, attitude and the steering order that highly control module 1-3 produces according to steering order generation module 1-7, calculate and obtain attitude and height control signal, then through controlling the expectation rotating speed that distributes four motors that obtain quadrotor, the expectation rotating speed of these four motors is delivered to airborne microcontroller 2 for controlling electron speed regulator and the motor of quadrotor successively by data outputting module 1-5 after 3, No. 2 wireless transceivers 4 of No. 1 wireless transceiver

Hardware-in-the-loop simulation data recordin module 1-6 estimates the expectation rotating speed with two groups of data of filtration module 1-2 output and four motors of quadrotor for store status.

Embodiment two: the difference of the quadrotor Hardware In The Loop Simulation Method described in present embodiment and embodiment one is, described airborne sensor 5 comprises three axle electronic compass 5-1, six axis movement sensor 5-2 and ultrasonic range finder sensor 5-3

Three axle electronic compass 5-1 are used for gathering three axle magnetic induction density signals,

Six axis movement sensor 5-2 are used for gathering 3-axis acceleration reading and three axle orientation readings,

Ultrasonic range finder sensor 5-3 is used for gathering range finding reading.

Embodiment three: the difference of the quadrotor Hardware In The Loop Simulation Method described in present embodiment and embodiment two is, the integrated circuit that the employing model of three described axle electronic compass 5-1 is HMC5883L is realized, the model of six axis movement sensor 5-2 is that the integrated circuit of MPU6050 is realized, and ultrasonic range finder sensor 5-3 adopts the integrated circuit that model is US-100 to realize.

Embodiment four: the difference of the quadrotor Hardware In The Loop Simulation Method described in present embodiment and embodiment one, two or three is, described airborne microcontroller 2 employing TMS320F28335 type DSP realizations.

Embodiment five: the difference of the quadrotor Hardware In The Loop Simulation Method described in present embodiment and embodiment one, two or three is, described attitude with the optimizing process of height control module 1-3 is,

In Model Distinguish subsystem, parameter identification device receives the expectation rotating speed Ω of four motors of quadrotor ₁, Ω ₂, Ω ₃, Ω ₄, attitude Eulerian angle estimated value height Estimation value the estimated value of angular velocity estimated value with speed on vertical direction this parameter identification device, according to model structure to be identified, is tried to achieve the identification result of parameter vector, and then has determined aircraft movements and kinetic model, thereby completes the optimization with height control module 1-3 to attitude according to numerical simulation subsystem,

Described model structure to be identified is described by following state space equation:

$\left\{\begin{array}{c}\stackrel{\·}{\mathrm{\φ}}=p\cos \mathrm{\θ}+r\sin \mathrm{\θ}\\ \stackrel{\·}{\mathrm{\θ}}=q-r\cos \mathrm{\θ}\tan \mathrm{\φ}+p\sin \mathrm{\θ}\tan \mathrm{\φ}\\ \stackrel{\·}{\mathrm{\ψ}}=r\cos \mathrm{\θ}\sec \mathrm{\φ}-p\sin \mathrm{\θ}\sec \mathrm{\φ}\\ \stackrel{\·}{p}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{yy}}-I_{\mathrm{zz}}}{I_{\mathrm{xx}}}\right)-\stackrel{\·}{\mathrm{\θ}}\frac{J_r}{I_{\mathrm{xx}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{xx}}}{U}_2\\ \stackrel{\·}{q}=\stackrel{\·}{\mathrm{\φ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{zz}}-I_{\mathrm{xx}}}{I_{\mathrm{yy}}}\right)+\stackrel{\·}{\mathrm{\φ}}\frac{J_r}{I_{\mathrm{yy}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{yy}}}{U}_3\\ \stackrel{\·}{r}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\φ}}\left(\frac{I_{\mathrm{xx}}-I_{\mathrm{yy}}}{I_{\mathrm{zz}}}\right)+\frac{1}{I_{\mathrm{zz}}}{U}_4\\ \stackrel{\·}{z}=w\\ \stackrel{\·}{w}=g-\left(\cos \mathrm{\φ}\cos \mathrm{\θ}\right)\frac{1}{m}{U}_1\end{array}\right.$ (formula one),

Wherein, $\left\{\begin{array}{c}{U}_1=b({{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ U_2=b(-{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_4^2)\\ U_3=b({{\mathrm{\Ω}}_1}^2-{\mathrm{\Ω}}_3^2)\\ U_4=d(-{{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2-{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ {\mathrm{\Ω}}_r={\mathrm{\Ω}}_2+{\mathrm{\Ω}}_4-{\mathrm{\Ω}}_1-{\mathrm{\Ω}}_3\end{array}\right.$ (formula two),

Described parameter identification device is according to model structure to be identified, and the detailed process that obtains the identification result of parameter vector is:

Expectation rotating speed Ω when four motors ₁, Ω ₂, Ω ₃, Ω ₄, during as the input of dummy vehicle, the numerical solution of trying to achieve part predicted state by formula one and formula two is ξ (k)=[φ (k), θ (k), ψ (k), p (k), q (k), r (k), z (k), w (k)] ^t,

Secondly, adopt cost function $V_N\left(r\right)=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}({\widehat{\mathrm{\ξ}}}^2\left(k\right)-{\mathrm{\ξ}}^2\left(k\right)),$ Ask for r ₀,

$r_0=\arg \underset{r}{\min }{V}_N\left(r\right),$

R ₀be the identification result of parameter vector, r is parameter vector to be identified, wherein, and r=[m, l, I _xx, I _yy, I _zz, J _r, b, d] ^t,

Parameter vector r to be identified comprises aircraft gross mass m, and relatively motor shaft is apart from l, body moment of inertia (I _xx, I _yy, I _zz), rotor and rotor be around motor shaft moment of inertia J _r, lift coefficient b, torque coefficient d, N is the sampling number of record; K is sampling time point, U ₁, U ₂, U ₃, U ₄and Ω _rbe intermediate variable, g is acceleration of gravity.

Embodiment six: present embodiment is described referring to Fig. 2, the numerical value emulation method of the quadrotor described in present embodiment, first, parameter identification device receives from semi-matter simulating system and obtains data message, this parameter identification device is according to model structure to be identified, obtains the identification result of parameter vector, and by numerical simulation subsystem according to the identification result of the parameter vector of acquisition, obtain aircraft movements and kinetic model, thereby complete the numerical simulation of quadrotor.

Described parameter identification device receives from semi-matter simulating system and obtains the expectation rotating speed Ω that data message comprises four motors of quadrotor ₁, Ω ₂, Ω ₃, Ω ₄, attitude Eulerian angle estimated value height Estimation value the estimated value of angular velocity estimated value with speed on vertical direction

Described model structure to be identified is described by following state space equation:

$\left\{\begin{array}{c}\stackrel{\·}{\mathrm{\φ}}=p\cos \mathrm{\θ}+r\sin \mathrm{\θ}\\ \stackrel{\·}{\mathrm{\θ}}=q-r\cos \mathrm{\θ}\tan \mathrm{\φ}+p\sin \mathrm{\θ}\tan \mathrm{\φ}\\ \stackrel{\·}{\mathrm{\ψ}}=r\cos \mathrm{\θ}\sec \mathrm{\φ}-p\sin \mathrm{\θ}\sec \mathrm{\φ}\\ \stackrel{\·}{p}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{yy}}-I_{\mathrm{zz}}}{I_{\mathrm{xx}}}\right)-\stackrel{\·}{\mathrm{\θ}}\frac{J_r}{I_{\mathrm{xx}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{xx}}}{U}_2\\ \stackrel{\·}{q}=\stackrel{\·}{\mathrm{\φ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{zz}}-I_{\mathrm{xx}}}{I_{\mathrm{yy}}}\right)+\stackrel{\·}{\mathrm{\φ}}\frac{J_r}{I_{\mathrm{yy}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{yy}}}{U}_3\\ \stackrel{\·}{r}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\φ}}\left(\frac{I_{\mathrm{xx}}-I_{\mathrm{yy}}}{I_{\mathrm{zz}}}\right)+\frac{1}{I_{\mathrm{zz}}}{U}_4\\ \stackrel{\·}{z}=w\\ \stackrel{\·}{w}=g-\left(\cos \mathrm{\φ}\cos \mathrm{\θ}\right)\frac{1}{m}{U}_1\end{array}\right.$ (formula one),

Wherein, $\left\{\begin{array}{c}{U}_1=b({{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ U_2=b(-{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_4^2)\\ U_3=b({{\mathrm{\Ω}}_1}^2-{\mathrm{\Ω}}_3^2)\\ U_4=d(-{{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2-{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ {\mathrm{\Ω}}_r={\mathrm{\Ω}}_2+{\mathrm{\Ω}}_4-{\mathrm{\Ω}}_1-{\mathrm{\Ω}}_3\end{array}\right.$ (formula two),

Described parameter identification device is according to model structure to be identified, and the detailed process that obtains the identification result of parameter vector is:

Expectation rotating speed Ω when four motors ₁, Ω ₂, Ω ₃, Ω ₄, during as the input of dummy vehicle, the numerical solution of trying to achieve part predicted state by formula one and formula two is ξ (k)=[φ (k), θ (k), ψ (k), p (k), q (k), r (k), z (k), w (k)] ^t,

Secondly, adopt cost function $V_N\left(r\right)=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}({\widehat{\mathrm{\ξ}}}^2\left(k\right)-{\mathrm{\ξ}}^2\left(k\right)),$ Ask for r ₀,

$r_0=\arg \underset{r}{\min }{V}_N\left(r\right),$

R ₀be the identification result of parameter vector, r is parameter vector to be identified, wherein, and r=[m, l, I _xx, I _yy, I _zz, J _r, b, d] ^t,

(φ, θ, ψ) is attitude Eulerian angle, and z is height, and (p, q, r) is angular velocity, and w is speed on vertical direction,

Parameter vector r to be identified comprises aircraft gross mass m, and relatively motor shaft is apart from l, body moment of inertia (I _xx, I _yy, I _zz), rotor and rotor be around motor shaft moment of inertia J _r, lift coefficient b, torque coefficient d, N is the sampling number of record; K is sampling time point, U ₁, U ₂, U ₃, U ₄and Ω _rbe intermediate variable, g is acceleration of gravity.

Simulated environment configuration and operation steps in the present invention:

Before starting to carry out hardware-in-the-loop simulation, first need in MATLAB 2013b, complete following configuration:

1. real-time kernel is installed

RTWT needs a real-time kernel with windows interface, and real-time kernel can, by the real-time executable file of the non-dispensing of the highest execution priority, can move uninterruptedly it under the sample frequency of setting.Use rtwintgt – install order that real-time kernel can be installed.

2. set external schema code and generate parameter

After setting up Simulink model, just can input work Simulink Coder carries out code generation and has set up the simulation parameter of real-time application.In model configuration parameter dialog box, select code to generate panel, in target selection, partly click navigation button, in aims of systems listed files, select for setting up the aims of systems file rtwin.tlc of Real-time Windows Target application program, dialog box can be automatically set as template spanned file rtwin.tmf, generates to order and be set as make_rtw like this.At hardware, realize in panel, equipment supplier selects Generic, and device type is selected 32 x86 compatibilities, chooses " testing hardware is identical with products-hardware " check box.

3. set oscillograph parameter

Just can input oscillograph parameter after in Simulink model for signal trace adding oscillograph module.Open oscillograph parameter dialog box, in time range text box input time scope the upper bound.Right click coordinate axis is selected coordinate axis attribute, can set the indication range of y axle.

4. external schema control panel arranges

Under code menu, select external schema control panel.Click data filing button, chooses and enables filing, Simulink oscillograph and can storing data on hard disk to work space module under pattern externally.

5. online modification parameter: the first is to double-click the module that need to revise parameter, clicks OK after revising parameter; The second is to be variable assignments again in order line, by shortcut Ctrl+D, just modification can be downloaded in executable file and goes.

With Simulink, substituting C language is dsp program:

Utilize Embedded Coder and DSP System Toolbox tool box in Simulink, can realize the cross compile matching with this type figure signal processor, Simulink is built to the .out file that program generating digital signal processor can directly be carried out, break away from numerous and diverse C Programming with Pascal Language, greatly improved development efficiency.

Utilize the Embedded Coder of Simulink automatically to generate the executable code of C2000, first need to configure translation and compiling environment, be about to Simulink environment and be connected with CCS environment, concrete steps are:

1, first following program is installed:

MATLAB 2013B, CCS4.1.2, compiler (Compiler) CGT 5.2.3, linker (linker), real time operating system BIOS 5.41.02.14, DSP Real-Time Component XDC Tools 3.16.02.32 and flashburn tools Flash Tools 2.10.

2, MATLAB real-time kernel configuration:

In command boxe, input >>rtwintgt – setup

3, connect Simulink environment and CCS environment:

In command boxe, input >>checkEnvSetup (' ccsv4', ' f28335', ' setup'), and selection tool place file successively, connection can be completed.

4, Makefile compiling configuration:

In command boxe, input >>xmakefilesetup, and select ticcs_c2000_ccsv4_clone code build environment.

After configuring translation and compiling environment, to build block diagram and carry out cross compile, concrete steps are:

1, enter Simulink environment and utilize DSP System Tools and Embedded Coder tool box, build model;

2, simulation parameter is configured, solver (solver) is chosen as to discrete value fixed step size solver.

3, enter code and generate (Code Generation) panel, selecting ert.tlc is that aims of systems file, C are target language, target hardware is selected TI Delfino F2833x (boot from flash), the instrument chain of generative process is selected CCSv4 (C2000), is configured to generate fast to shorten the compiling link time.

4, after successful generating code, in CCS environment .out file is loaded in target.

Claims (6)

1. quadrotor Hardware In The Loop Simulation Method, it is characterized in that, it is realized based on following apparatus, and this device comprises simulation computer (1), airborne microcontroller (2), No. 1 wireless transceiver (3), No. 2 wireless transceivers (4) and airborne sensor (5);

The data-signal input end of airborne microcontroller (2) is connected with the data-signal output terminal of airborne sensor (5), the communication terminal of airborne microcontroller (2) is connected with the signal input output end of No. 2 wireless transceivers (4), No. 2 wireless transceivers (4) carry out data transmission by mode and No. 1 wireless transceiver (3) of radio communication, the signal input output end of No. 1 wireless transceiver (3) is connected with the signal input output end of simulation computer (1)

Simulation computer (1) is embedded in Matlab/Simulink emulation module, and described Matlab/Simulink emulation module comprises data input module (1-1), state estimation and filtration module (1-2), attitude and height control module (1-3), state of flight display module (1-4), data outputting module (1-5), hardware-in-the-loop simulation data recordin module (1-6) and steering order generation module (1-7);

The detailed process of this Hardware In The Loop Simulation Method is:

First, by the data-signal of airborne microcontroller (2) harvester set sensor (5) output, and airborne microcontroller (2) is delivered to the data-signal collecting the data input module (1-1) of simulation computer (1) by No. 2 wireless transceivers (4), No. 1 wireless transceiver (3) successively, this data-signal comprises range finding reading, 3-axis acceleration reading, three axis angular rate readings and three axle magnetic induction density signals;

By data input module (1-1), receive the data-signal of input, and these data are delivered to state estimation and filtration module (1-2), state estimation and filtration module (1-2) utilize the method for complementary filter to obtain two groups of data according to the data-signal receiving, first group of data in these two groups of data are attitude Eulerian angle estimated value and Height Estimation value, the estimated value of speed on the estimated value that second group of data are angular velocity and vertical direction

And first group of data delivered to state of flight display module (1-4), and state of flight display module (1-4) is for showing in real time the state of flight of quadrotor,

First group and second group of data are delivered to attitude and height control module (1-3), attitude and the steering order that highly control module (1-3) produces according to steering order generation module (1-7), calculate and obtain attitude and height control signal, then through controlling the expectation rotating speed that distributes four motors that obtain quadrotor, the expectation rotating speed of these four motors is delivered to airborne microcontroller (2) for controlling electron speed regulator and the motor of quadrotor successively by data outputting module (1-5) after No. 1 wireless transceiver (3), No. 2 wireless transceivers (4)

Hardware-in-the-loop simulation data recordin module (1-6) is estimated the expectation rotating speed with two groups of data of filtration module (1-2) output and four motors of quadrotor for store status.

2. quadrotor Hardware In The Loop Simulation Method according to claim 1, it is characterized in that, described airborne sensor (5) comprises three axle electronic compasss (5-1), six axis movement sensors (5-2) and ultrasonic range finder sensor (5-3)

Three axle electronic compasss (5-1) are for gathering three axle magnetic induction density signals,

Six axis movement sensors (5-2) are for gathering 3-axis acceleration reading and three axis angular rate readings,

Ultrasonic range finder sensor (5) is for gathering range finding reading.

3. quadrotor Hardware In The Loop Simulation Method according to claim 2, it is characterized in that, the integrated circuit that the employing model of three described axle electronic compasss (5-1) is HMC5883L is realized, the integrated circuit that the model of six axis movement sensors (5-2) is MPU6050 is realized, and ultrasonic range finder sensor (5) adopts the integrated circuit that model is US-100 to realize.

4. according to the quadrotor Hardware In The Loop Simulation Method described in claim 1,2 or 3, it is characterized in that, described airborne microcontroller (2) adopts TMS320F28335 type DSP to realize.

5. according to the quadrotor Hardware In The Loop Simulation Method described in claim 1,2 or 3, it is characterized in that, described attitude with height control module (1-3) optimizing process be,

In Model Distinguish subsystem, parameter identification device receives the expectation rotating speed Ω of four motors of quadrotor ₁, Ω ₂, Ω ₃, Ω ₄, attitude Eulerian angle estimated value height Estimation value the estimated value of angular velocity estimated value with speed on vertical direction this parameter identification device, according to model structure to be identified, is tried to achieve the identification result of parameter vector, and then has determined aircraft movements and kinetic model, thereby completes the optimization with height control module (1-3) to attitude according to numerical simulation subsystem,

Described model structure to be identified is described by following state space equation:

$\left\{\begin{array}{c}\stackrel{\·}{\mathrm{\φ}}=p\cos \mathrm{\θ}+r\sin \mathrm{\θ}\\ \stackrel{\·}{\mathrm{\θ}}=q-r\cos \mathrm{\θ}\tan \mathrm{\φ}+p\sin \mathrm{\θ}\tan \mathrm{\φ}\\ \stackrel{\·}{\mathrm{\ψ}}=r\cos \mathrm{\θ}\sec \mathrm{\φ}-p\sin \mathrm{\θ}\sec \mathrm{\φ}\\ \stackrel{\·}{p}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{yy}}-I_{\mathrm{zz}}}{I_{\mathrm{xx}}}\right)-\stackrel{\·}{\mathrm{\θ}}\frac{J_r}{I_{\mathrm{xx}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{xx}}}{U}_2\\ \stackrel{\·}{q}=\stackrel{\·}{\mathrm{\φ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{zz}}-I_{\mathrm{xx}}}{I_{\mathrm{yy}}}\right)+\stackrel{\·}{\mathrm{\φ}}\frac{J_r}{I_{\mathrm{yy}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{yy}}}{U}_3\\ \stackrel{\·}{r}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\φ}}\left(\frac{I_{\mathrm{xx}}-I_{\mathrm{yy}}}{I_{\mathrm{zz}}}\right)+\frac{1}{I_{\mathrm{zz}}}{U}_4\\ \stackrel{\·}{z}=w\\ \stackrel{\·}{w}=g-\left(\cos \mathrm{\φ}\cos \mathrm{\θ}\right)\frac{1}{m}{U}_1\end{array}\right.$ (formula one),

Wherein, $\left\{\begin{array}{c}{U}_1=b({{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ U_2=b(-{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_4^2)\\ U_3=b({{\mathrm{\Ω}}_1}^2-{\mathrm{\Ω}}_3^2)\\ U_4=d(-{{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2-{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ {\mathrm{\Ω}}_r={\mathrm{\Ω}}_2+{\mathrm{\Ω}}_4-{\mathrm{\Ω}}_1-{\mathrm{\Ω}}_3\end{array}\right.$ (formula two),

Described parameter identification device is according to model structure to be identified, and the detailed process that obtains the identification result of parameter vector is:

Expectation rotating speed Ω when four motors ₁, Ω ₂, Ω ₃, Ω ₄, during as the input of dummy vehicle, the numerical solution of trying to achieve part predicted state by formula one and formula two is ξ (k)=[φ (k), θ (k), ψ (k), p (k), q (k), r (k), z (k), w (k)] ^t,

Secondly, adopt cost function $V_N\left(r\right)=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}({\widehat{\mathrm{\ξ}}}^2\left(k\right)-{\mathrm{\ξ}}^2\left(k\right)),$ Ask for r ₀,

$r_0=\arg \underset{r}{\min }{V}_N\left(r\right),$

R ₀be the identification result of parameter vector, r is parameter vector to be identified, wherein, and r=[m, l, I _xx, I _yy, I _zz, J _r, b, d] ^t,

(φ, θ, ψ) is attitude Eulerian angle, and z is height, and (p, q, r) is angular velocity, and w is speed on vertical direction,

Parameter vector r to be identified comprises aircraft gross mass m, and relatively motor shaft is apart from l, body moment of inertia (I _xx, I _yy, I _zz), rotor and rotor be around motor shaft moment of inertia J _r, lift coefficient b, torque coefficient d, N is the sampling number of record; K is sampling time point, U ₁, U ₂, U ₃, U ₄and Ω _rbe intermediate variable, g is acceleration of gravity.

6. the numerical value emulation method of quadrotor, is characterized in that, the detailed process of the method is,

First, parameter identification device receives from semi-matter simulating system and obtains data message, this parameter identification device is according to model structure to be identified, obtain the identification result of parameter vector, and by numerical simulation subsystem according to the identification result of the parameter vector obtaining, obtain aircraft movements and kinetic model, thereby complete the numerical simulation of quadrotor.

Described parameter identification device receives from semi-matter simulating system and obtains the expectation rotating speed Ω that data message comprises four motors of quadrotor ₁, Ω ₂, Ω ₃, Ω ₄, attitude Eulerian angle estimated value height Estimation value the estimated value of angular velocity estimated value with speed on vertical direction

Described model structure to be identified is described by following state space equation:

$\left\{\begin{array}{c}\stackrel{\·}{\mathrm{\φ}}=p\cos \mathrm{\θ}+r\sin \mathrm{\θ}\\ \stackrel{\·}{\mathrm{\θ}}=q-r\cos \mathrm{\θ}\tan \mathrm{\φ}+p\sin \mathrm{\θ}\tan \mathrm{\φ}\\ \stackrel{\·}{\mathrm{\ψ}}=r\cos \mathrm{\θ}\sec \mathrm{\φ}-p\sin \mathrm{\θ}\sec \mathrm{\φ}\\ \stackrel{\·}{p}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{yy}}-I_{\mathrm{zz}}}{I_{\mathrm{xx}}}\right)-\stackrel{\·}{\mathrm{\θ}}\frac{J_r}{I_{\mathrm{xx}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{xx}}}{U}_2\\ \stackrel{\·}{q}=\stackrel{\·}{\mathrm{\φ}}\stackrel{\·}{\mathrm{\ψ}}\left(\frac{I_{\mathrm{zz}}-I_{\mathrm{xx}}}{I_{\mathrm{yy}}}\right)+\stackrel{\·}{\mathrm{\φ}}\frac{J_r}{I_{\mathrm{yy}}}{\mathrm{\Ω}}_r+\frac{l}{I_{\mathrm{yy}}}{U}_3\\ \stackrel{\·}{r}=\stackrel{\·}{\mathrm{\θ}}\stackrel{\·}{\mathrm{\φ}}\left(\frac{I_{\mathrm{xx}}-I_{\mathrm{yy}}}{I_{\mathrm{zz}}}\right)+\frac{1}{I_{\mathrm{zz}}}{U}_4\\ \stackrel{\·}{z}=w\\ \stackrel{\·}{w}=g-\left(\cos \mathrm{\φ}\cos \mathrm{\θ}\right)\frac{1}{m}{U}_1\end{array}\right.$ (formula one),

Wherein, $\left\{\begin{array}{c}{U}_1=b({{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ U_2=b(-{\mathrm{\Ω}}_2^2+{\mathrm{\Ω}}_4^2)\\ U_3=b({{\mathrm{\Ω}}_1}^2-{\mathrm{\Ω}}_3^2)\\ U_4=d(-{{\mathrm{\Ω}}_1}^2+{\mathrm{\Ω}}_2^2-{\mathrm{\Ω}}_3^2+{\mathrm{\Ω}}_4^2)\\ {\mathrm{\Ω}}_r={\mathrm{\Ω}}_2+{\mathrm{\Ω}}_4-{\mathrm{\Ω}}_1-{\mathrm{\Ω}}_3\end{array}\right.$ (formula two),

Described parameter identification device is according to model structure to be identified, and the detailed process that obtains the identification result of parameter vector is:

Expectation rotating speed Ω when four motors ₁, Ω ₂, Ω ₃, Ω ₄, during as the input of dummy vehicle, the numerical solution of trying to achieve part predicted state by formula one and formula two is ξ (k)=[φ (k), θ (k), ψ (k), p (k), q (k), r (k), z (k), w (k)] ^t,

Secondly, adopt cost function $V_N\left(r\right)=\underset{k=1}{\overset{N}{\mathrm{\Σ}}}({\widehat{\mathrm{\ξ}}}^2\left(k\right)-{\mathrm{\ξ}}^2\left(k\right)),$ Ask for r ₀,

$r_0=\arg \underset{r}{\min }{V}_N\left(r\right),$

R ₀be the identification result of parameter vector, r is parameter vector to be identified, wherein, and r=[m, l, I _xx, I _yy, I _zz, J _r, b, d] ^t,

Parameter vector r to be identified comprises aircraft gross mass m, and relatively motor shaft is apart from l, body moment of inertia (I _xx, I _yy, I _zz), rotor and rotor be around motor shaft moment of inertia J _r, lift coefficient b, torque coefficient d, N is the sampling number of record; K is sampling time point, U ₁, U ₂, U ₃, U ₄and Ω _rbe intermediate variable, g is acceleration of gravity.