Virtual gyroscope reconstruction and precision improvement method based on distributed inertial navigation autocorrelation
Through the virtual gyroscope reconstruction method of distributed inertial autocorrelation, the accuracy of redundant multi-inertial navigation information is improved by using space-time normalization and Kalman filters, solving the problem of difficulty in improving accuracy in the existing technology, and achieving more High-precision navigation output and real-time calculation.
Patent Information
- Application Number
- CN202311421367.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-31
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2043-10-31
AI Technical Summary
The prior art is difficult to effectively improve the accuracy of redundant multi-inertial navigation information and cannot surpass the highest accuracy inertial navigation system.
Through a virtual gyroscope reconstruction method based on distributed inertial autocorrelation, a virtual gyroscope error model is constructed and information fusion is performed to generate a higher-precision virtual gyroscope output.
It achieves autonomous, simple and reliable navigation accuracy improvement without being restricted by the external environment, supports the engineering application goal of ‘1+1>2’, and ensures the real-time nature of navigation computing.
Smart Images

Figure CN117629244B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of navigation application technology, and in particular to a virtual gyroscope reconstruction and precision improvement method based on distributed inertial navigation autocorrelation. Background Art
[0002] Modern ships / submersibles often have multiple types of inertial navigation systems, such as backup main inertial navigation and payload sub-inertial navigation. After time and space normalization, the above-mentioned inertial navigation measures the movement of the same carrier, and the measurement information constitutes redundant measurements. At present, for redundant measurements, data fusion methods such as averaging or weighted averaging can generally be used, which can be started from the two perspectives of gyroscope and inertial navigation. The former performs data fusion on the measurement values of each axis of the gyroscope, and the latter performs data fusion on the instrument measurement values and navigation solution results (attitude, velocity and position). However, the averaging or weighted averaging method can only balance the noise and eliminate wild values, and cannot make the fusion result surpass the highest accuracy inertial navigation system.
[0003] The time-space normalized distributed inertial navigation is in the same physical space and measures the same physical parameters. Both the instrument measurement values and the navigation solution results are strongly correlated. Summary of the invention
[0004] Aiming at the problem of difficulty in improving the accuracy of redundant multi-inertial navigation information fusion and combining the characteristics of distributed inertial navigation autocorrelation, the present invention proposes a virtual gyroscope reconstruction and accuracy improvement method based on distributed inertial navigation autocorrelation, which can improve the accuracy of the gyroscope.
[0005] The virtual gyroscope reconstruction and accuracy improvement method based on distributed inertial navigation autocorrelation includes:
[0006] Step 1: Perform time-space normalization processing on redundant multi-INS and construct the correlation between INS physically;
[0007] Step 2: Based on the measurement output model of the gyroscope, construct a virtual gyroscope error model, including:
[0008] Step 2.1: Model the gyroscope's measurement output as:
[0009]
[0010] In the formula, Indicates the actual value of the gyroscope output; Represents the true value of the gyroscope output; ε b represents the random zero bias of the gyroscope, represented by a first-order Markov model; η g represents the gyroscope measurement white noise, τ is the gyroscope random zero bias time constant; η represents the random zero bias white noise;
[0011] Step 2.2: Select 18-dimensional state variables and construct state equations and measurement equations based on the gyroscope measurement output model:
[0012] The state variable X is expressed as:
[0013]
[0014] In the formula, represents the random zero bias vector of the inertial navigation gyroscope, Represents the output vector of the inertial navigation gyroscope i; i = 1, 2, 3;
[0015] The state equation is expressed as:
[0016]
[0017] In the formula, F represents the system matrix, W represents the system noise vector, and the expression is:
[0018]
[0019] W=[(η1) T (η2) T (η3) T (0) T (0) T (0) T ] T
[0020] In the formula, τ i represents the time correlation coefficient of the inertial navigation gyroscope, I 3×3 represents a 3×3 unit matrix, 0 3×3 represents a 3×3 zero matrix, 0 represents a 3rd-order zero vector, η i represents the random zero-bias white noise of the inertial navigation gyroscope;
[0021] Select the gyroscope outputs of three sets of inertial navigation as the measurement variable Z:
[0022]
[0023] In the formula, Indicates the actual measurement value of the inertial navigation gyroscope;
[0024] The measurement equation is expressed as:
[0025] Z=HX+V
[0026] In the formula, H represents the measurement matrix, V represents the measurement noise vector, and the specific expression is:
[0027]
[0028] V=[(η g1) T (η g2 ) T (η g3 ) T ] T
[0029] Where η gi represents the inertial navigation gyro measurement white noise;
[0030] Step 3: Based on the gyroscope array correlation analysis, solve the virtual gyroscope measurement noise array and system noise array, including:
[0031] Step 3.1: Based on the white noise assumption, construct the variance matrix R of the measurement noise vector V:
[0032]
[0033] In the formula, represents the inertial navigation gyro measurement white noise η gi The corresponding variance matrix;
[0034] Step 3.2: The gyroscope array is installed on the same physical carrier and measures the same physical object. The gyroscopes are not independent of each other but have some correlation, that is, there is some correlation between the angular rate random walks. Based on the correlation analysis, the variance matrix Q of the system noise vector W is constructed:
[0035]
[0036] In the formula, represents the random zero bias white noise η of the inertial navigation gyroscope i i The corresponding variance matrix; ρ ij Indicates the correlation coefficient between the INS i and INS j gyroscopes; j = 1, 2, 3;
[0037] Step 4: Based on the virtual gyro error model in step 2 and the virtual gyro measurement noise matrix and system noise error matrix obtained in step 3, a Kalman filter is executed to perform information fusion to generate virtual gyro output information, including:
[0038] Using the output of three inertial navigation gyroscopes to construct the observation value Z k , real-time recursive calculation to obtain the state estimate at time k The specific solution process is as follows:
[0039]
[0040] In the formula, and P k,k-1 Respectively represent the k-1 moment state estimation and estimation error variance matrix of one-step prediction; K krepresents the filter gain matrix at time k; H k is the measurement matrix at time k; φ k,k-1 is the state transfer matrix from time k-1 to time k; is the state estimate at time k; Z k represents the measured variable at time k; H k represents the measurement matrix at time k; R k represents the variance matrix of the measurement noise vector V at time k; Q k-1 represents the variance matrix of the system noise vector W at time k-1; I represents the identity matrix.
[0041] Preferably, the step 1 comprises:
[0042] Step 1.1: Use the pulse-per-second signal of the satellite navigation equipment to synchronize the time of multiple inertial navigation systems;
[0043] Step 1.2: Install the redundant multi-INS on the same platform, and use the calibration device to calibrate the installation error between the redundant multi-INS, so that the installation error angle between the INS is 0;
[0044] Step 1.3: Based on the time-space normalization processing, the redundant multi-INS gyroscope measurements on the same axis have a strong correlation:
[0045]
[0046] In the formula, represents the x-axis gyroscope output value of inertial navigation i, represents the y-axis gyroscope output value of inertial navigation i, Indicates the z-axis gyroscope output value of inertial navigation i.
[0047] Preferably, in step 4, the Kalman filter is initialized as follows:
[0048]
[0049]
[0050] The present invention has the following beneficial effects:
[0051] 1. The present invention proposes a method for constructing and improving the accuracy of a virtual gyroscope based on distributed inertial navigation autocorrelation. The use of this method is not restricted by the external environment of the mission, does not require additional external auxiliary navigation sensors or information sources, and has the advantages of being completely autonomous, simple and reliable.
[0052] 2. The present invention can virtually reconstruct a virtual gyroscope with higher accuracy, thereby improving the accuracy of the navigation system from the device level and supporting the realization of the engineering application goal of "1+1>2".
[0053] 3. The virtual gyroscope reconstruction and precision improvement method adopted in the present invention has the advantages of small calculation amount and small data storage amount, and can effectively ensure the real-time performance of navigation calculation. BRIEF DESCRIPTION OF THE DRAWINGS
[0054] Figure 1 It is a schematic diagram of the virtual gyroscope reconstruction and precision improvement method based on distributed inertial navigation autocorrelation of the present invention. DETAILED DESCRIPTION
[0055] The technical solution of the present invention is further described below with reference to the accompanying drawings and three sets of distributed inertial navigation embodiments.
[0056] The present invention aims to improve the accuracy of redundant multi-INS information fusion by combining the characteristics of distributed INS autocorrelation, constructing a correlation model between gyroscope measurement values, and virtually reconstructing the redundant gyroscopes brought by redundant INS to achieve a breakthrough in gyroscope accuracy. Figure 1 As shown, the method for constructing a virtual gyroscope based on distributed inertial navigation autocorrelation according to the present invention comprises the following steps:
[0057] Step 1: Perform time-space normalization processing on redundant multi-INS to physically build a strong correlation between INS;
[0058] Step 1.1: Use the pulse-per-second signal of the satellite navigation device to synchronize the time of multiple inertial navigation systems, so that the parameter error caused by time asynchrony can be ignored;
[0059] Step 1.2: Install the redundant multi-INS on the same platform, and use the calibration device to calibrate the installation error between the redundant multi-INS, so that the installation error angle between the INS is approximately 0;
[0060] Step 1.3: Based on the time-space normalization processing, the redundant multi-INS gyroscope measurements on the same axis have a strong correlation:
[0061]
[0062] In the formula, Indicates the output value of the inertial navigation i (i=1,2,3) x-axis gyroscope, Indicates the output value of the y-axis gyroscope of the inertial navigation i (i=1,2,3), Indicates the output value of the inertial navigation i (i=1,2,3) z-axis gyroscope.
[0063] Step 2: Construct a virtual gyro error model based on the gyro measurement output model;
[0064] Step 2.1: Model the gyroscope's measurement output as:
[0065]
[0066] In the formula, Indicates the actual value of the gyroscope output; Represents the true value of the gyroscope output; ε b represents the random zero bias of the gyroscope, represented by a first-order Markov model; η g represents the gyroscope measurement white noise, τ is the gyroscope random zero bias time constant; η represents random zero bias white noise.
[0067] Step 2.2: Select 18-dimensional state variables and construct state equations and measurement equations based on the gyroscope measurement output model in step 1.1;
[0068] The state variable X is expressed as:
[0069]
[0070] In the formula, represents the random zero bias vector of the inertial navigation gyroscope i (i=1,2,3), Represents the output vector of inertial navigation gyroscope i (i=1,2,3).
[0071] The state equation is expressed as:
[0072]
[0073] In the formula, F represents the system matrix, W represents the system noise vector, and the specific expression is:
[0074]
[0075] W=[(η1) T (η2) T (η3) T (0) T (0) T (0) T ] T (6)
[0076] In the formula, τ i represents the time correlation coefficient of inertial navigation i (i=1,2,3) gyroscope, I 3×3 represents a 3×3 unit matrix, 0 3×3 represents a 3×3 zero matrix, 0 represents a 3rd-order zero vector, η i represents the random zero-bias white noise of inertial navigation gyroscope i (i=1,2,3).
[0077] Select the gyroscope outputs of three sets of inertial navigation as the measurement variable Z:
[0078]
[0079] In the formula, Indicates the actual measurement value of inertial navigation gyroscope i (i=1,2,3);
[0080] The measurement equation is expressed as:
[0081] Z=HX+V (8)
[0082] In the formula, H represents the measurement matrix, V represents the measurement noise vector, and the specific expression is:
[0083]
[0084] V=[(η g1 ) T (η g2 ) T (η g3 ) T ] T (10)
[0085] Where η gi Represents the white noise measured by inertial navigation gyroscope i (i=1,2,3).
[0086] Step 3: Based on the correlation analysis of the gyroscope array, solve the virtual gyroscope measurement noise array and the system noise array;
[0087] Step 3.1: Based on the white noise assumption, construct the variance matrix R of the measurement noise vector V:
[0088]
[0089] In the formula, represents the white noise η measured by the inertial navigation gyroscope i (i=1,2,3) gi The corresponding variance matrix;
[0090] Step 3.2: The gyroscope array is installed on the same physical carrier and measures the same physical object. The gyroscopes are not independent of each other but have some correlation, that is, there is some correlation between the angular rate random walks. Based on the correlation analysis, the variance matrix Q of the system noise vector W is constructed:
[0091]
[0092] In the formula, represents the random zero bias white noise η of the inertial navigation gyroscope i (i=1,2,3) i The corresponding variance matrix; ρ ij (i=1,2,3, j=1,2,3) represents the correlation coefficient between the inertial navigation i and inertial navigation j gyroscopes.
[0093] Step 4: Based on the virtual gyro error model in step 2 and the virtual gyro measurement noise matrix and system noise error matrix obtained by solving step 3, Kalman filter is executed to perform information fusion to generate virtual gyro output information.
[0094] Step 4.1: Kalman filter initialization;
[0095]
[0096]
[0097] Step 4.2: Construct the observation value Z using the output of the three inertial navigation gyroscopes k , real-time recursive calculation to obtain the state estimate at time k The specific solution process is as follows:
[0098]
[0099] In the formula, and P k,k-1 Respectively represent the k-1 moment state estimation and estimation error variance matrix of one-step prediction; K k represents the filter gain matrix at time k; H k is the measurement matrix at time k; φ k,k-1 is the state transfer matrix from time k-1 to time k; is the state estimate at time k; Z k represents the measured variable at time k; H k represents the measurement matrix at time k; R k represents the variance matrix of the measurement noise vector V at time k; Q k-1 represents the variance matrix of the system noise vector W at time k-1; I represents the identity matrix.
[0100] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A virtual gyroscope reconstruction and precision improvement method based on distributed inertial navigation autocorrelation, characterized in that: include: Step 1: Perform time-space normalization processing on redundant multi-INS and construct the correlation between INS physically; Step 2: Based on the measurement output model of the gyroscope, construct a virtual gyroscope error model, including: Step 2.1: Model the gyroscope's measurement output as: In the formula, Indicates the actual value of the gyroscope output; Represents the true value of the gyroscope output; ε b represents the random zero bias of the gyroscope, represented by a first-order Markov model; η g represents the gyroscope measurement white noise, τ is the gyroscope random zero bias time constant; η represents the random zero bias white noise; Step 2.2: Select 18-dimensional state variables and construct state equations and measurement equations based on the gyroscope measurement output model: The state variable X is expressed as: In the formula, represents the random zero bias vector of the inertial navigation gyroscope, Represents the output vector of the inertial navigation gyroscope i; i = 1, 2, 3; The state equation is expressed as: In the formula, F represents the system matrix, W represents the system noise vector, and the expression is: W=[(η1) T (n2) T (n3) T (0) T (0) T (0) T ] T In the formula, τ i represents the time correlation coefficient of the inertial navigation gyroscope, I 3×3 represents a 3×3 unit matrix, 0 3×3 represents a 3×3 zero matrix, 0 represents a 3rd-order zero vector, η i represents the random zero-bias white noise of the inertial navigation gyroscope; Select the gyroscope outputs of three sets of inertial navigation as the measurement variable Z: In the formula, Indicates the actual measurement value of the inertial navigation gyroscope; The measurement equation is expressed as: Z=HX+V In the formula, H represents the measurement matrix, V represents the measurement noise vector, and the specific expression is: V=[(η g1 ) T (or g2 ) T (or g3 ) T ] T Where η gi represents the inertial navigation gyro measurement white noise; Step 3: Based on the gyroscope array correlation analysis, solve the virtual gyroscope measurement noise array and system noise array, including: Step 3.1: Based on the white noise assumption, construct the variance matrix R of the measurement noise vector V: In the formula, represents the inertial navigation gyro measurement white noise η gi The corresponding variance matrix; Step 3.2: The gyroscope array is installed on the same physical carrier and measures the same physical object. The gyroscopes are not independent of each other but have some correlation, that is, there is some correlation between the angular rate random walks. Based on the correlation analysis, the variance matrix Q of the system noise vector W is constructed: In the formula, represents the random zero bias white noise η of the inertial navigation gyroscope i i The corresponding variance matrix; ρ ij Indicates the correlation coefficient between the INS i and INS j gyroscopes; j = 1, 2, 3; Step 4: Based on the virtual gyro error model in step 2 and the virtual gyro measurement noise matrix and system noise error matrix obtained in step 3, a Kalman filter is executed to perform information fusion to generate virtual gyro output information, including: Using the output of three inertial navigation gyroscopes to construct the observation value Z k , real-time recursive calculation to obtain the state estimate at time k The specific solution process is as follows: In the formula, and P k,k-1 Respectively represent the k-1 moment state estimation and estimation error variance matrix of one-step prediction; K k represents the filter gain matrix at time k; H k is the measurement matrix at time k; φ k,k-1 is the state transfer matrix from time k-1 to time k; is the state estimate at time k; Z k represents the measured variable at time k; H k represents the measurement matrix at time k; R k represents the variance matrix of the measurement noise vector V at time k; Q k-1 represents the variance matrix of the system noise vector W at time k-1; I represents the identity matrix.
2. The method for virtual gyroscope reconstruction and precision improvement based on distributed inertial navigation autocorrelation as claimed in claim 1, characterized in that: The step 1 comprises: Step 1.1: Use the pulse-per-second signal of the satellite navigation equipment to synchronize the time of multiple inertial navigation systems; Step 1.2: Install the redundant multi-INS on the same platform, and use the calibration device to calibrate the installation error between the redundant multi-INS, so that the installation error angle between the INS is 0; Step 1.3: Based on the time-space normalization processing, the redundant multi-INS gyroscope measurements on the same axis have a strong correlation: In the formula, represents the x-axis gyroscope output value of inertial navigation i, represents the y-axis gyroscope output value of inertial navigation i, Indicates the z-axis gyroscope output value of inertial navigation i.
3. The method for virtual gyroscope reconstruction and precision improvement based on distributed inertial navigation autocorrelation as claimed in claim 1, characterized in that: In step 4, the Kalman filter is initialized as follows:
