A method for system-level calibration of rotating strapdown inertial navigation system errors

Abstract

The invention discloses a method for calibrating the system error of a rotating strapdown inertial navigation system. On the basis that the rotating inertial navigation system has a rotating mechanism, a 30-position rotation calibration strategy is set, and the position flipping of the inertial navigation system in the rotation strategy can fully stimulate various errors of the inertial measurement unit; secondly, a Kalman filter model is established, and the speed error and position error are used as measurement quantities to estimate the inertial navigation error; finally, through simulation experiments, it is verified that the method for calibrating the system error of the 30-position rotation strapdown inertial navigation system designed by the invention can calibrate the constant error, scale coefficient error and installation error of the gyroscope and accelerometer at one time with high precision.

Description

Technical Field

[0001] The invention relates to the technical field of strapdown inertial navigation, and in particular to a system-level calibration method for a rotating strapdown inertial navigation system error. Background Art

[0002] The main component of the inertial navigation system is the inertial measurement unit. In the production and manufacturing, there are inevitably related device errors, such as constant error, scale factor error, installation error, etc. In order to suppress the influence of these errors on the inertial navigation system, in addition to the rotation modulation technology, the error calibration technology is often used to estimate and compensate the error to achieve the purpose of high-precision navigation. Calibration is mainly divided into discrete calibration and system-level calibration. The discrete calibration technology installs the inertial navigation system on the rotating mechanism, calibrates the gyroscope and accelerometer through different position arrangements, and compensates the inertial navigation system with the calibrated errors. However, the discrete calibration has the disadvantages of cumbersome disassembly and assembly, high precision requirements for the rotating mechanism, and inability to calibrate in the field. The rotating strapdown inertial navigation system can use its own rotating mechanism for system-level calibration, which can avoid the complexity of disassembly and assembly from the carrier and provide a realistic basis for long-duration precise navigation.

[0003] System-level calibration stimulates the error of the inertial measurement unit by designing a reasonable position arrangement, establishes the input and output mathematical model of the inertial measurement unit, takes the navigation error (speed error, position error) as the observed quantity, and estimates the error parameters through the least square method or Kalman filter. SAGEM designed an 18-position calibration scheme (Camberlein L, Mazzanti F. Calibration technique for laser gyro strapdown inertial navigation systems [J]. Ortung Und Navigation, 1985: 5.0-5.13.) to calibrate the laser gyro inertial navigation system. After calibration, the system-level accuracy requirements can be achieved. In addition, this scheme has rich engineering practice experience and is a traditional error calibration method. The literature (Xie Bo, Qin Yongyuan, Wan Yanhui. Multi-position calibration method of laser gyro strapdown inertial navigation system [J]. Journal of Chinese Inertial Technology, 2011, 19(02): 157-162+169.) proposed a 19-position calibration method. Through multiple initial alignment, position flipping and static navigation process to stimulate the error parameters of the inertial measurement unit, the least square method can be used to calibrate 21 error parameters in a short time. The rotation calibration scheme is a key factor in the system-level calibration technology. Summary of the invention

[0004] The purpose of the present invention is to solve the error calibration problem of an inertial navigation system, and to provide a new error calibration method for a rotating strapdown inertial navigation system, which can simultaneously calibrate the constant error, scale coefficient error and installation error of a gyroscope and an accelerometer, and compensate the inertial navigation system, thereby achieving the purpose of improving the system accuracy.

[0005] To achieve the above object, the present invention provides a method for calibrating the error of a rotating strapdown inertial navigation system, comprising the following steps:

[0006] Step S11: define the coordinate system and define the constant error parameters, scale factor error parameters and installation error parameters of the gyroscope and accelerometer in the inertial measurement unit

[0007] Determine the inertial coordinate system (set as system i) where the inertial measurement unit of the rotating inertial navigation system is located, and the origin o of the inertial coordinate system i Located at the center of the Earth, the distance from the origin to the North Pole is o i -z i The axis starts from the origin and points to the mean vernal equinox at o i -x i Axis, o i -y i Axis and o i -z i The axes form a right-handed rectangular coordinate system;

[0008] Determine the navigation coordinate system (set as n system) in which the inertial measurement unit of the rotating inertial navigation system is located. This coordinate system is the local geographic coordinate system, and its origin is o n Located at the center of mass of the carrier, its o n -x n , o n -y n , o n -z n The axes point from the origin to the east, north, and celestial directions respectively;

[0009] Determine the carrier coordinate system (set as system b) where the inertial measurement unit of the rotating inertial navigation system is located, and the center of mass of the carrier is the origin of the carrier coordinate system o b , the coordinate system o b -x b , o b -y b and b -z b The axes are directed from the center of mass of the carrier to the right side, the front side and the upper side of the carrier respectively;

[0010] Determine the inertial measurement unit coordinate system (set as p system). The inertial measurement unit consists of three sets of orthogonally installed gyroscopes and accelerometers. The origin of the inertial measurement unit coordinate system is o p Located at the center of mass of the inertial measurement unit, its op -x p , o p -y p , o p -z p The axis points from the origin to be parallel to the directions of three sets of orthogonally mounted gyroscopes and accelerometers;

[0011] Determine the installation coordinate system (set as m system). When making an inertial measurement unit, it is impossible to ensure that the three sets of gyroscopes and accelerometers are installed in an ideal orthogonal manner. Suppose the coordinate system of the inertial measurement unit during actual installation is the installation coordinate system, and the origin of the installation coordinate system is o m Same as p series;

[0012] The constant errors of the gyroscope and accelerometer are denoted as ε and The specific expressions are as follows:

[0013]

[0014] Among them, ε x , ε y , ε z Gyroscope o p -x p , o p -y p , o p -z p Constant error of the axis, are the constant error of the accelerometer in o p -x p , o p -y p , o p -z p The weight of the axis.

[0015] The scale coefficient errors of the gyroscope and accelerometer are expressed as δK g and δK a , specifically expressed as follows:

[0016]

[0017] Among them, δK gx ,δK gy ,δK gz are the gyroscope scale coefficient errors in o p -x p , o p -y p , o p -z p Axis component, δK ax ,δK ay ,δK azThe error of the accelerometer scale factor is p -x p , o p -y p , o p -z p The weight of the axis.

[0018] The installation errors of the gyroscope and accelerometer are expressed as δA g and δA a , specifically expressed as follows:

[0019]

[0020] Among them, δA gyx for o m -y m Axis gyroscope with o p -x p -y p Installation error angle of the plane, δA gzx for o m -z m Axis gyroscope with o p -x p -z p Installation error angle of the plane, δA gzy for o m -z m Axis gyroscope with o p -y p -z p Installation error angle of the plane, δA gxy for o m -x m Axis gyroscope with o p -x p -y p Installation error angle of the plane, δA gxz for o m -x m Axis gyroscope with o p -x p -z p Installation error angle of the plane, δA gyz for o m -y m Axis gyroscope with o p -y p -z p Installation error angle of the plane, δA ayx for o m -y m Axis accelerometer with o p -x p -y p Installation error angle of the plane, δA azx for om -z m Axis accelerometer with o p -x p -z p Installation error angle of the plane, δA azy for o m -z m Axis accelerometer with o p -y p -z p Installation error angle of the plane, δA axy for o m -x m Axis accelerometer with o p -x p -y p Installation error angle of the plane, δA axz for o m -x m Axis accelerometer with o p -x p -z p Installation error angle of the plane, δA ayz for o m -y m Axis accelerometer with o p -y p -z p Installation error angle of the plane.

[0021] Step S21: Construct gyroscope and accelerometer error models

[0022] In order to reduce the number of calibration parameters, improve the calculation rate and shorten the calibration time, it is assumed that the gyroscope orthogonal system o p -x p Axis and installation coordinate system o m -x m The axes coincide, and o p -y p Axis in o m -x m -y m plane. Then the gyroscope error model is as follows:

[0023]

[0024] in, are the gyroscope angle increment errors in o b -x b , o b -y b , o b -z b The component in the axial direction, are the gyroscope angle increments in o b-x b , o b -y b , o b -z b Component in the direction of the axis.

[0025] Accelerometer error model:

[0026]

[0027] in, are the accelerometer specific force increment errors in o b -x b , o b -y b , o b -z b The component in the axial direction, are the accelerometer specific force increment in o b -x b , o b -y b , o b -z b Component in the direction of the axis.

[0028] Step S31: The error parameters defined in step 11 are obtained by flipping the gyroscope and accelerometer error models in step 21.

[0029] Initially, the three axes of the inertial measurement unit are o p -x p , o p -y p , o p -z p Pointing to the sky, east and north respectively, the central axis of the rotating mechanism and the o of the inertial measurement unit p -y p The axes coincide, the inner axis of the rotating mechanism and the o of the inertial measurement unit p -z p The axes coincide with each other, and the outer axis of the rotating mechanism coincides with the o of the inertial measurement unit p -x pAxes coincide. The 30-position calibration path is designed as follows: 1. Rotate 90 degrees forward around the middle frame axis of the rotating mechanism, with an angular velocity of ω degrees / second, and stop for t seconds after the rotation is completed; 2. Rotate 90 degrees forward around the middle frame axis of the rotating mechanism, with an angular velocity of ω degrees / second, and stop for t seconds after the rotation is completed; 3. Rotate 90 degrees forward around the middle frame axis of the rotating mechanism, with an angular velocity of ω degrees / second, and stop for t seconds after the rotation is completed; 4. Rotate 180 degrees in the opposite direction around the middle frame axis of the rotating mechanism, with an angular velocity of ω degrees / second, and stop for t seconds after the rotation is completed; 5. Rotate 90 degrees in the opposite direction around the middle frame axis of the rotating mechanism, with an angular velocity of ω degrees / second, and stop for t seconds after the rotation is completed; 6. Rotate 90 degrees forward around the inner frame axis of the rotating mechanism, with an angular velocity of ω degrees / second, and stop for t seconds after the rotation is completed; 7. Rotate 90 degrees forward around the middle frame axis of the rotating mechanism, with an angular velocity of ω degrees / second, and stop for t seconds after the rotation is completed; 8. Rotate 90 degrees forward around the middle frame axis of the rotating mechanism, with an angular velocity of ω degrees / second, and stop for t seconds. 15. Rotate 90 degrees in the opposite direction around the inner frame axis of the rotating mechanism at an angular velocity of ω degrees / second, and stop for t seconds after the rotation is completed; 16-30 positions have the same rotation axis order as 1-15 positions, but the rotation direction is opposite.

[0030] Step S41: Establish a Kalman filter model based on the error parameters in step 31

[0031] After a series of flips in step 31, the gyroscope and accelerometer can be stimulated to generate errors. Then, the errors are introduced into the inertial navigation attitude error, velocity error and position error equations to obtain the system state error model.

[0032] The 30-dimensional Kalman state equation is established, whose state vector includes attitude error, velocity error, position error, gyroscope and accelerometer constant error, scale coefficient error and installation error:

[0033]

[0034] Among them, the 30-dimensional state vector X(t) is

[0035]

[0036] is the differential of the state vector X, F is the state transfer matrix, and the expression is as follows:

[0037]

[0038] The specific expressions of each block matrix are as follows:

[0039]

[0040] Among them, ω ie is the Earth's rotation angular rate, R M , R N They are the curvature radii of the earth's meridian and meridian respectively, L is the latitude of the carrier, and h is the altitude of the carrier.

[0041]

[0042]

[0043] Among them, v E 、v N are the north and north speeds respectively.

[0044]

[0045] in, They are the components of the accelerometer output in the east, north and celestial directions in the navigation coordinate system respectively.

[0046]

[0047] Among them, v U The celestial speed.

[0048]

[0049]

[0050]

[0051] Attitude transformation matrix from the carrier coordinate system to the navigation coordinate system as follows:

[0052]

[0053]

[0054]

[0055] In formula (3), G is the system noise driving matrix, which is expressed as follows:

[0056]

[0057] W(t) is the system noise matrix, which is composed of the gyroscope noise W g (t) and accelerometer noise W a (t), which are all zero-mean Gaussian white noise vector sequences, and their expressions are as follows:

[0058]

[0059] The following Kalman filter observation equation is established with velocity error and position error as observation quantities:

[0060] Z(t)=H(t)X(t)+V(t) (19)

[0061] Among them, Z(t) is the observed quantity, which consists of velocity error and position error, and the expression is as follows:

[0062] Z(t)=[δv E δv N δv U δL δλ δh] T (20)

[0063] H is the observation matrix, which is expressed as follows:

[0064]

[0065] V(t) is the observation noise matrix, which represents the system noise that can obtain the observed quantity, including velocity observation noise and position observation noise, both of which are zero-mean Gaussian white noise vector sequences, and the above W(t) and V(t) are independent of each other, and the expression is as follows:

[0066]

[0067] Step S51: Discretize the Kalman filter model to obtain the constant error, scale coefficient error and installation error estimation of the gyroscope and accelerometer

[0068] Given the system state space model:

[0069]

[0070] Where, X k is the discretized state vector, Z k is the discretized measurement vector, F k / k-1 , H k and Γ k / k-1 are the state transfer matrix, measurement matrix and system noise driving matrix respectively, W k-1 and V kare the system noise vector and the measurement noise vector respectively.

[0071] The discrete state transfer matrix F from time k-1 to time k k / k-1 The specific expression is as follows:

[0072]

[0073] Where I is the identity matrix, T s is the filtering period.

[0074] The complete set of Kalman filtering algorithms is as follows:

[0075] One-step prediction of state

[0076]

[0077] State one-step prediction mean square error matrix

[0078]

[0079] Filter gain

[0080]

[0081] State Estimation

[0082]

[0083] State Estimation Mean Square Error

[0084] P k =(IK k H k ) k / k-1 (28)

[0085] In formula (25), Q is the mean square error matrix of the white noise output by the gyroscope and accelerometer, and R is the observation noise variance matrix.

[0086] In summary, combined with a series of flipping movements in step 31, the inertial measurement unit output data is subjected to inertial navigation solution, and Kalman filtering is performed at the same time. Under the given system state vector initial value X0, state estimation mean square error initial value P0, gyroscope and accelerometer output white noise mean square error matrix Q0 and observation noise variance matrix initial value R0, and the selection of observation quantity Z, state transfer matrix F and observation matrix H is determined, the state quantity X at any time can be recursively estimated by equations (24) to (28), so that the constant error, scale coefficient error and installation error estimation of the gyroscope and accelerometer can be obtained.

[0087] Step S61: Correct the system using the constant error, scale factor error and installation error estimate of the gyroscope and accelerometer obtained in step 51. The error correction model is as follows:

[0088] Gyroscope and accelerometer constant error feedback:

[0089]

[0090]

[0091] Among them, ε and is the system constant error after feedback, and is the system constant error before feedback, and Constant error estimate.

[0092] Gyroscope scale coefficient error and installation error feedback:

[0093]

[0094] Where K g and A g is the scale factor and installation angle matrix after gyroscope feedback, and is the gyro feedback pre-scale coefficient and installation angle matrix, δK g and δA g are the estimated values ​​of gyroscope scale coefficient error and installation error.

[0095] Gyroscope and accelerometer installation error feedback:

[0096]

[0097] Where K a and A a is the scale factor and installation angle matrix after accelerometer feedback, and is the accelerometer feedback pre-scale coefficient and installation angle matrix, δK a and δA a are estimated values ​​of accelerometer scale factor error and installation error.

[0098] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention utilizes the advantages of the rotating mechanism of the rotating inertial navigation system itself and the designed 30-position rotation calibration scheme, which can fully stimulate the errors of inertial navigation related devices in an outdoor environment, and then establishes a 30-dimensional Kalman filter model, using velocity error and position error as observation quantities to estimate the constant error, scale coefficient error and installation error of the gyroscope and accelerometer with high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0099] Figure 1 It is a schematic diagram of the rotating mechanism;

[0100] Figure 2 It is a rotation calibration strategy diagram;

[0101] Figure 3 Comparison of simulation and experimental error parameter estimation curves

[0102] Figure 4 Comparison of navigation positioning error after estimated inertial device error compensation DETAILED DESCRIPTION

[0103] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0104] The feasibility of the present invention can be verified by simulation experiments, such as Figure 1 The inertial measurement unit is installed on the rotating mechanism, and the calibration rotation scheme under the b system is as follows Figure 2 As shown, the simulation experiment adopts a comparative experiment, taking the traditional 18-position scheme designed by SAGEM as a reference, and comparing the calibration effect of the scheme of the present invention:

[0105] Experimental conditions: initial latitude 30.6 degrees, initial longitude 114.2 degrees, initial altitude 25 meters; initial attitude (0°, -90°, -90°); static simulation is adopted, so the initial velocity is 0; initial attitude error (0.5′, -0.5′, 0.5′); initial velocity error (0.1m / s, 0.1m / s, 0.1m / s,); initial position error (1m, 1m, 3m); simulation time is 4 hours.

[0106] Kalman filter model initial parameter settings:

[0107] The initial value of the system state vector X0 = [0 30×1 ];

[0108] The initial value of the state estimation mean square error P0:

[0109] P0=diag{(0.5′) 2 ,(-0.5′) 2 ,(0.5′) 2 ,(0.1m / s) 2 ,(0.1m / s) 2 ,(0.1m / s) 2 ,(1m) 2 ,(1m) 2 ,(3m) 2 ,(0.05° / h) 2 ,(0.05° / h) 2,(0.05° / h) 2 ,(80ug) 2 ,(90ug) 2 ,(100ug) 2 ,(40ppm) 2 ,(10″) 2 ,(10″) 2 ,(40ppm) 2 ,(10″) 2 ,(40ppm) 2 ,(20ppm) 2 ,(10″) 2 ,(10″) 2 ,(10″) 2 ,(20ppm) 2 ,(10″) 2 ,(10″) 2 ,(10″) 2 ,(20ppm) 2}

[0110] The mean square error matrix Q0 of white noise is:

[0111] Q0=diag{(0.001° / h) 2 ,(0.001° / h) 2 ,(0.001° / h) 2 ,(1ug) 2 ,(1ug) 2 ,(1ug) 2}

[0112] The initial value of the observation noise variance matrix R0:

[0113] R0=diag{(0.1m / s) 2 ,(0.1m / s) 2 ,(0.1m / s) 2 ,(1m) 2 ,(1m) 2 ,(1m) 2}

[0114] The parameter setting values ​​and estimated values ​​of the constant error, installation error, and scale coefficient error of the gyroscope and accelerometer in the simulation experiment are shown in Table 1.

[0115] Table 1 Calibration error parameter setting values ​​and estimated value results

[0116]

[0117] As attached Figure 33a gyroscope x-axis constant error, 3b gyroscope y-axis constant error, 3c gyroscope z-axis constant error, 3d accelerometer x-axis constant error, 3e accelerometer y-axis constant error, 3f accelerometer z-axis constant error, 3g gyroscope x-axis scale coefficient error, 3h gyroscope y-axis scale coefficient error, 3i gyroscope z-axis scale coefficient error, 3j accelerometer x-axis scale coefficient error, 3k accelerometer y-axis scale coefficient error, 3l accelerometer z-axis scale coefficient error, 3m gyroscope installation error yx, 3n gyroscope installation error zx, 3o gyroscope installation error zy, 3p accelerometer installation error yx, 3q accelerometer installation error zx, 3r accelerometer installation error xy, 3s accelerometer installation error zy, 3t accelerometer installation error xz, 3u accelerometer installation error yz show the estimation curves of 21 related error parameters of gyroscopes and accelerometers. It can be seen from the figure that the error parameters of the designed 30-position scheme have converged under the calibration rotation strategy and Kalman filtering, just like the traditional 18-position scheme, and most of the error parameters have better convergence effect.

[0118] Combined with the results in Table 1, the residual errors of 16 out of the 21 error parameters estimated by the designed 30-position calibration scheme are smaller than those of the 18-position scheme. After the inertial device errors estimated by the two calibration schemes are compensated for by the IMU, a 48-hour static base navigation is performed. The navigation positioning error curves of the two schemes are shown in the attached figure. Figure 4 ,As can be seen from the figure, the maximum positioning error of the designed 30-position ,calibration scheme is 3639 meters smaller than the traditional 18-position scheme, which ,can verify that the calibration scheme proposed in the present ,invention has higher calibration accuracy.

Claims

1. A method for calibrating the error of a rotating strapdown inertial navigation system at the system level, characterized by: The calibration method is as follows: Step S11: define the coordinate system and define the constant error parameters, scale factor error parameters and installation error parameters of the gyroscope and accelerometer in the inertial measurement unit Step S21: Construct gyroscope and accelerometer error models Step S31: The error parameters defined in step 11 are obtained by flipping the gyroscope and accelerometer error models in step 21. Step S41: Establish a Kalman filter model based on the error parameters in step 31 Step S51: Discretize the Kalman filter model to obtain the constant error, scale coefficient error and installation error estimation of the gyroscope and accelerometer Step S61: Correct the system using the constant error, scale factor error and installation error estimation values ​​of the gyroscope and accelerometer obtained in step 51; The specific process of establishing the Kalman filter model according to the error parameters in step S41 is as follows: After the flipping in step 31, the errors generated by the gyroscope and the accelerometer are excited; then the errors are introduced into the inertial navigation attitude error, velocity error and position error equations to obtain the system state error model; The 30-dimensional Kalman state equation is established, whose state vector includes attitude error, velocity error, position error, gyroscope and accelerometer constant error, scale coefficient error and installation error: Among them, the 30-dimensional state vector X(t) is is the differential of the state vector X, F is the state transfer matrix, and the expression is as follows:

2. The method for calibrating the rotational strapdown inertial navigation system error at the system level according to claim 1, characterized in that: In the step S11, Determine the inertial coordinate system i in which the inertial measurement unit of the rotating inertial navigation system is located, and the origin o of the inertial coordinate system i Located at the center of the Earth, the distance from the origin to the North Pole is o i -z i The axis starts from the origin and points to the mean vernal equinox at o i -x i Axis, o i -y i Axis and o i -z i The axes form a right-handed rectangular coordinate system; Determine the navigation coordinate system n of the rotating inertial navigation system inertial measurement unit, which is the local geographic coordinate system with its origin o n Located at the center of mass of the carrier, its o n -x n , o n -y n , o n -z n The axes point from the origin to the east, north, and celestial directions respectively; Determine the carrier coordinate system b in which the inertial measurement unit of the rotating inertial navigation system is located, and the center of mass of the carrier is the origin o of the carrier coordinate system b , the coordinate system o b -x b , o b -y b and b -z b The axes are directed from the center of mass of the carrier to the right side, the front side and the upper side of the carrier respectively; Determine the inertial measurement unit coordinate system p. The inertial measurement unit consists of three sets of orthogonally mounted gyroscopes and accelerometers. The origin of the inertial measurement unit coordinate system is o p Located at the center of mass of the inertial measurement unit, its o p -x p , o p -y p , o p -z p The axis points from the origin to be parallel to the directions of three sets of orthogonally mounted gyroscopes and accelerometers; Determine the installation coordinate system m system, assume that the inertial measurement unit coordinate system during actual installation is the installation coordinate system, and the origin of the installation coordinate system is o m and the origin o of the p system p same.

3. The method for calibrating the rotational strapdown inertial navigation system error at the system level according to claim 2, characterized in that: In the step S11, The constant errors of the gyroscope and accelerometer are denoted as ε and The specific expressions are as follows: Among them, ε x , ε y , ε z Gyroscope o p -x p , o p -y p , o p -z p Constant error of the axis, are the constant error of the accelerometer in o p -x p , o p -y p , o p -z p The weight of the axis; The scale coefficient errors of the gyroscope and accelerometer are expressed as δK g and δK a , specifically expressed as follows: Among them, δK gx ,δK gy ,δK gz are the gyroscope scale coefficient errors in o p -x p , o p -y p , o p -z p Axis component, δK ax ,δK ay ,δK az The error of the accelerometer scale factor is p -x p , o p -y p , o p -z p The weight of the axis; The installation errors of the gyroscope and accelerometer are expressed as δA g and δA a , specifically expressed as follows: Among them, δA gyx for o m -y m Axis gyroscope with o p -x p -y p Installation error angle of the plane, δA gzx for o m -z m Axis gyroscope with o p -x p -z p Installation error angle of the plane, δA gzy for o m -z m Axis gyroscope with o p -y p -z p Installation error angle of the plane, δA gxy for o m -x m Axis gyroscope with o p -x p -y p Installation error angle of the plane, δA gxz for o m -x m Axis gyroscope with o p -x p -z p Installation error angle of the plane, δA gyz for o m -y m Axis gyroscope with o p -y p -z p Installation error angle of the plane, δA ayx for o m -y m Axis accelerometer with o p -x p -y p Installation error angle of the plane, δA azx for o m -z m Axis accelerometer with o p -x p -z p Installation error angle of the plane, δA azy for o m -z m Axis accelerometer with o p -y p -z p Installation error angle of the plane, δA axy for o m -x m Axis accelerometer with o p -x p -y p Installation error angle of the plane, δA axz for o m -x m Axis accelerometer with o p -x p -z p Installation error angle of the plane, δA ayz for o m -y m Axis accelerometer with o p -y p -z p Installation error angle of the plane.

4. The method for calibrating the rotational strapdown inertial navigation system error at the system level according to claim 3, characterized in that: The specific process of constructing the gyroscope and accelerometer error model in step S21 is as follows: Assume that the gyro orthogonal system o p -x p Axis and installation coordinate system o m -x m The axes coincide, and o p -y p Axis in o m -x m -y m In the plane, the gyroscope error model is as follows: in, are the gyroscope angle increment errors in o b -x b , o b -y b , o b -z b The component in the axial direction, are the gyroscope angle increments in o b -x b , o b -y b , o b -z b Component in the axial direction; Accelerometer error model: in, are the accelerometer specific force increment errors in o b -x b , o b -y b , o b -z b The component in the axial direction, are the accelerometer specific force increment in o b -x b , o b -y b , o b -z b Component in the direction of the axis.

5. The method for calibrating the rotational strapdown inertial navigation system error at system level according to claim 4, characterized in that: In step S31, the error parameters defined in step 12 are obtained by flipping the gyroscope and accelerometer error models in step 21 and exciting them. The specific process is as follows: Initially, the three axes of the inertial measurement unit are o p -x p , o p -y p , o p -z p Pointing to the sky, east and north respectively, the central axis of the rotating mechanism and the o of the inertial measurement unit p -y p The axes coincide, the inner axis of the rotating mechanism and the o of the inertial measurement unit p -z p The axes coincide with each other, and the outer axis of the rotating mechanism coincides with the o of the inertial measurement unit p -x p Axis coincidence; set the 30 position calibration path as follows:

1. Rotate 90 degrees forward around the middle frame axis of the rotating mechanism, the rotation angular velocity is ω degrees / second, and stop for t seconds after the rotation is completed; 2. Rotate 90 degrees forward around the middle frame axis of the rotating mechanism, the rotation angular velocity is ω degrees / second, and stop for t seconds after the rotation is completed; 3. Rotate 90 degrees forward around the middle frame axis of the rotating mechanism, the rotation angular velocity is ω degrees / second, and stop for t seconds after the rotation is completed; 4. Rotate 180 degrees in the opposite direction around the middle frame axis of the rotating mechanism, the rotation angular velocity is ω degrees / second, stop for t seconds after the rotation is completed; 5. Rotate 90 degrees in the opposite direction around the middle frame axis of the rotating mechanism, the angular velocity of rotation is ω degrees / second, and stop for t seconds after the rotation is completed; 6. Rotate 90 degrees forward around the inner frame axis of the rotating mechanism, the angular velocity of rotation is ω degrees / second, and stop for t seconds after the rotation is completed; 7. Rotate 90 degrees forward around the middle frame axis of the rotating mechanism, the angular velocity of rotation is ω degrees / second, and stop for t seconds after the rotation is completed; 8. Rotate 90 degrees forward around the middle frame axis of the rotating mechanism, the angular velocity of rotation is ω degrees / second , stop for t seconds after the rotation is completed; 9. Rotate 90 degrees forward around the middle frame axis of the rotating mechanism, the angular velocity of rotation is ω degrees / second, and stop for t seconds after the rotation is completed; 10. Rotate 180 degrees in the opposite direction around the middle frame axis of the rotating mechanism, the angular velocity of rotation is ω degrees / second, and stop for t seconds after the rotation is completed; 11. Rotate 90 degrees in the opposite direction around the middle frame axis of the rotating mechanism, the angular velocity of rotation is ω degrees / second, and stop for t seconds after the rotation is completed; 12. Rotate 90 degrees forward around the inner frame axis of the rotating mechanism, the angular velocity of rotation is ω degrees / second, and stop for t seconds after the rotation is completed; 13. Rotate 90 degrees forward around the inner frame axis of the rotating mechanism, the angular velocity of rotation is ω degrees / second, and stop for t seconds after the rotation is completed; 14. Rotate 180 degrees in the opposite direction around the inner frame axis of the rotating mechanism, the angular velocity of rotation is ω degrees / second, and stop for t seconds after the rotation is completed; 15. Rotate 90 degrees in the opposite direction around the inner frame axis of the rotating mechanism, the angular velocity of rotation is ω degrees / second, and stop for t seconds after the rotation is completed; The order of the rotation axes of positions 16-30 is the same as that of positions 1-15, but the rotation direction is opposite.

6. The method for calibrating the rotational strapdown inertial navigation system error at system level according to claim 5, characterized in that: The specific expressions of each block matrix in formula (5) are as follows: Among them, ω ie is the Earth's rotation angular rate, R M , R N are the curvature radii of the earth's meridian and meridian respectively, L is the latitude of the carrier, and h is the height of the carrier; Among them, v E 、v N are the north and north speeds, respectively; in, are the components of the accelerometer output in the east, north and celestial directions in the navigation coordinate system respectively; Among them, v U is the celestial velocity; Attitude transformation matrix from the carrier coordinate system to the navigation coordinate system as follows: In formula (3), G is the system noise driving matrix, which is expressed as follows: W(t) is the system noise matrix, which is composed of the gyroscope noise W g (t) and accelerometer noise W a (t), which are all zero-mean Gaussian white noise vector sequences, and their expressions are as follows: The following Kalman filter observation equation is established with velocity error and position error as observation quantities: Z(t)=H(t)X(t)+V(t) (19) Among them, Z(t) is the observed quantity, which consists of velocity error and position error, and the expression is as follows: Z(t)=[δv E dv N dv U [Lδλδh] T (20) H is the observation matrix, which is expressed as follows: V(t) is the observation noise matrix, which represents the system noise that can obtain the observed quantity, including velocity observation noise and position observation noise, both of which are zero-mean Gaussian white noise vector sequences, and the above W(t) and V(t) are independent of each other, and the expression is as follows:

7. The method for calibrating the rotational strapdown inertial navigation system error at system level according to claim 6, characterized in that: The specific process of discretizing the constant error, scale coefficient error and installation error estimation value of the gyroscope and accelerometer using the Kalman filter model in step S51 is as follows: Given the system state space model: Where, X k is the discretized state vector, Z k is the discretized measurement vector, F k / k-1 , H k and Γ k / k-1 are the state transfer matrix, measurement matrix and system noise driving matrix respectively, W k-1 and V k are the system noise vector and the measurement noise vector respectively; The discrete state transfer matrix F from time k-1 to time k k / k-1 The specific expression is as follows: Where I is the identity matrix, T s is the filtering period; The complete set of Kalman filtering algorithms is as follows: One-step state prediction State one-step prediction mean square error matrix Filter gain State Estimation State Estimation Mean Square Error P k =(I-K k H k )P k / k-1 (28) In formula (25), Q is the mean square error matrix of the white noise output by the gyroscope and accelerometer, and R is the observation noise variance matrix.

8. The method for calibrating the rotation strapdown inertial navigation system error at system level according to claim 7, characterized in that: The specific process of correcting the system using the constant error, scale coefficient error and installation error estimation value of the gyroscope and accelerometer obtained in step 51 in step S61 is as follows: The error correction model is shown below: Gyroscope and accelerometer constant error feedback: Among them, ε and is the system constant error after feedback, and is the system constant error before feedback, and constant error estimate; Gyroscope scale coefficient error and installation error feedback: Where K g and A g is the scale factor and installation angle matrix after gyroscope feedback, and is the gyro feedback pre-scale coefficient and installation angle matrix, δK g and δA g is the estimated value of gyroscope scale coefficient error and installation error; Gyroscope and accelerometer installation error feedback: Where K a and A a is the scale factor and installation angle matrix after accelerometer feedback, and is the accelerometer feedback pre-scale coefficient and installation angle matrix, δK a and δA a are estimated values ​​of accelerometer scale factor error and installation error.

Images