Horizontal dynamic error suppression method for vehicle-mounted strapdown gravity vector repetition line measurement
By applying low-pass filtering and Kalman filtering methods in vehicle-mounted gravitational vector measurement, the problem of horizontal dynamic error in vehicle-mounted measurements is solved, and the measurement accuracy is significantly improved.
Patent Information
- Application Number
- CN202411757845.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-03
- Publication Date
- 2025-03-14
AI Technical Summary
There is horizontal dynamic error in vehicle-mounted gravity vector measurement, which makes the measurement accuracy unable to meet the high accuracy requirements.
Low-pass filtering is performed through frequency separation means, with a period of 200s, which suppresses the influence of centripetal acceleration generated by vehicle turning maneuver on horizontal relative force measurement; then Kalman filtering is used to fuse the measurement results of the forward and reverse measurement lines to achieve the cancellation of the measurement error of the horizontal accelerometer.
It effectively suppresses horizontal dynamic errors during the measurement process of vehicle-mounted gravity vector repeat lines, and significantly improves the accuracy of gravity vector measurement results in the horizontal direction.
Smart Images
Figure FT_1 
Figure FT_2 
Figure FT_3
Abstract
Description
Technical field:
[0001] The invention belongs to the technical field of gravity vector measurement. Background technology:
[0002] Vehicle-mounted strapdown gravity measurement is a low-cost, flexible and mobile gravity measurement method, which provides a good choice for the refinement of regional gravity fields. At present, vehicle-mounted strapdown gravity scalar measurement has developed relatively maturely, but vehicle-mounted strapdown gravity vector measurement still has certain technical bottlenecks and cannot meet the accuracy requirements for building a regional high-precision gravity field. One of the difficulties is the horizontal dynamic error caused by the vehicle-mounted measurement environment.
[0003] The dynamics of the carrier have a great impact on strapdown gravity vector measurement. In aerial gravity vector measurement, the measurement line is generally planned as a straight line, and the aircraft is required to maintain a uniform and stable attitude along the straight line to ensure measurement accuracy and quality; but in vehicle-mounted strapdown gravity vector measurement, the measurement line has a complex shape due to the limitations of the road shape. The vehicle cannot maintain a stable heading and speed during the measurement process, and the influence of centripetal acceleration will also be introduced during the maneuvering turn. These factors increase the complexity of the horizontal dynamic error of the vehicle-mounted strapdown gravity vector measurement and the difficulty of subsequent processing of the vehicle-mounted strapdown gravity vector data. Therefore, corresponding measures need to be taken to suppress the horizontal dynamic error of the vehicle-mounted strapdown gravity vector measurement and create conditions for subsequent vehicle-mounted strapdown gravity vector data processing. Summary of the invention:
[0004] The purpose of the present invention is to effectively suppress the horizontal dynamic error in the process of vehicle-mounted strapdown gravity vector repetitive line measurement by using frequency separation means and error cancellation means.
[0005] The present invention firstly performs low-pass filtering on the original signal of the horizontal direction of the vehicle-mounted strapdown gravity vector measurement, and the low-pass filtering period is 200s, so as to suppress the influence of the centripetal acceleration generated by the vehicle turning maneuver on the horizontal specific force measurement; Then, the horizontal measurement results of the forward and reverse survey lines are fused through Kalman filtering to cancel the measurement error of the horizontal accelerometer and suppress the influence of the heading angle change of the vehicle-mounted survey line on the horizontal gravity vector measurement result.
[0006] The detailed steps of the present invention are: S1: low-pass filter the continuous raw signal of the horizontal direction of the vehicle-mounted strapdown gravity vector measurement, and the period of the low-pass filter is selected as 200s; S2: The continuous gravity vector measurement data in the horizontal direction after low-pass filtering is divided into measurement lines. The measurement line from the starting point to the end point is called the forward measurement line, and the measurement line from the end point back to the starting point is called the reverse measurement line. A pair of adjacent forward measurement lines and reverse measurement lines that go from the starting point to the end point and then back to the starting point is called a group of measurement lines. S3: The measurement data on the first group of survey lines are numbered according to the mileage, where the starting point is 0 by default and the end point is the total mileage of the survey line by default; S4: aligning the measurement data at the same position of the forward survey line and the reverse survey line in the first group of survey lines by interpolation; S5: using the Kalman filter method to fuse the measurement data at the same position of the first group of measurement lines, and obtain the effective measurement results of the first group of measurement lines; S6: S3 to S5 are repeated in a loop, and the remaining forward survey lines and reverse survey lines are processed in turn, and finally the horizontal dynamic error of the vehicle-mounted strapdown gravity vector measurement is suppressed.
[0007] The present invention can effectively suppress the horizontal dynamic error in the process of repeated line measurement of vehicle-mounted strapdown gravity vector and significantly improve the accuracy of gravity vector measurement results in the horizontal direction. Description of the drawings:
[0008] Figure 1 is a flow chart of the method of the present invention; Figure 2 It is a schematic diagram of the centripetal acceleration of the vehicle during turning; Figure 3 It is a simulation trajectory diagram of vehicle turning; Figure 4 It is the horizontal force diagram directly caused by the centripetal acceleration when the vehicle turns; Figure 5 It is the east and north force comparison diagram directly caused by the centripetal acceleration when the vehicle turns; Figure 6 This is the east- and north-direction specific forces (200s low-pass filter) caused by the centripetal acceleration when the vehicle turns; Figure 7 is the measurement error of the accelerometer on the vehicle forward and reverse measurement lines ▽ n Graph of changes in Figure 8 This is the survey line trajectory diagram of the horizontal accelerometer measurement error cancellation test; Figure 9a This is the result of the horizontal east-direction gravity vector measurement directly solved by the traditional method; the six lines on the right side of the figure are arranged from top to bottom as follows: Line 6, Line 2, Line 1, Line 3, Line 4, Line 5; Figure 9b This is a diagram of the horizontal north gravity vector measurement result directly solved by the traditional method; in the figure, the six lines on the left end are arranged from top to bottom: survey line 5, survey line 3, survey line 1, survey line 4, survey line 2, and survey line 6; Fig.10a This is the horizontal eastward gravity vector measurement result diagram after the fusion processing of the forward and reverse survey lines; Fig.10bThis is the measurement result of the horizontal north gravity vector after the fusion of the forward and reverse survey lines. Specific implementation method:
[0009] The present invention firstly performs low-pass filtering on the original signal of the horizontal direction of the vehicle-mounted strapdown gravity vector measurement by means of frequency separation, and the low-pass filtering period is 200s, so as to suppress the influence of the centripetal acceleration generated by the vehicle turning maneuver on the horizontal specific force measurement; then, by means of error cancellation, the Kalman filter is used to fuse the horizontal direction measurement results of the forward survey line and the reverse survey line, so as to achieve the cancellation of the horizontal accelerometer measurement error and suppress the influence of the vehicle-mounted survey line heading angle change on the horizontal gravity vector measurement result. The method provided by the present invention can effectively suppress the horizontal dynamic error in the process of repeated line measurement of the vehicle-mounted strapdown gravity vector, create good conditions for the subsequent vehicle-mounted strapdown gravity vector data processing, and significantly improve the accuracy of the gravity vector measurement result in the horizontal direction.
[0011] The main implementation mode of the present invention is: Firstly, the original signal of the horizontal direction of the vehicle-mounted strapdown gravity vector measurement is low-pass filtered with a period of 200s to suppress the influence of the centripetal acceleration generated by the vehicle turning maneuver on the horizontal specific force measurement.
[0012] Then, the horizontal measurement results of the forward and reverse survey lines are fused through Kalman filtering to cancel the measurement error of the horizontal accelerometer and suppress the influence of the heading angle change of the vehicle-mounted survey line on the horizontal gravity vector measurement result.
[0013] By adopting the method provided by the present invention, the horizontal dynamic error in the process of repeated line measurement of vehicle-mounted strapdown gravity vector can be effectively suppressed, and the accuracy of the gravity vector measurement result in the horizontal direction can be significantly improved.
[0014] The steps of the present invention are as follows: S1: Low-pass filter the continuous raw signal of the horizontal direction of the vehicle-mounted strapdown gravity vector measurement, and the period of the low-pass filter is selected as 200s.
[0015] S2: The continuous gravity vector measurement data in the horizontal direction after low-pass filtering is divided into survey lines. The survey line from the starting point to the end point is called a forward survey line, and the survey line from the end point back to the starting point is called a reverse survey line. An adjacent pair of forward survey lines and reverse survey lines that go from the starting point to the end point and then back to the starting point is called a group of survey lines.
[0016] S3: The measurement data on the first group of survey lines are numbered according to mileage, where the starting point is 0 by default and the end point is the total mileage of the survey line by default.
[0017] S4: Align the measurement data at the same position of the forward survey line and the reverse survey line in the first group of survey lines by interpolation.
[0018] S5: The Kalman filter method is used to fuse the measurement data at the same position of the first group of measurement lines to obtain the effective measurement results of the first group of measurement lines.
[0019] S6: S3 to S5 are repeated in a loop, and the remaining forward survey lines and reverse survey lines are processed in turn, and finally the horizontal dynamic error of the vehicle-mounted strapdown gravity vector measurement is suppressed.
[0020] The method of the present invention is described in more detail below. The present invention comprises the following steps: S1: Low-pass filter the continuous raw signal of the horizontal direction of the vehicle-mounted strapdown gravity vector measurement, and the period of the low-pass filter is selected as 200s.
[0021] In the S1 stage, the influence of the centripetal acceleration generated by the vehicle turning maneuver on the horizontal specific force measurement is suppressed by frequency separation. The relationship between angular velocity and frequency can be converted by the following formula: f=2ωπ (1) Where f is the frequency and ω is the angular velocity.
[0022] According to the angular velocity that a vehicle can reach when turning on the road, the frequencies corresponding to the angular velocities of 5° / s, 10° / s and 25° / s are 0.0139 Hz, 0.0278 Hz and 0.0694 Hz, respectively, while the cutoff frequency of the 200s low-pass filter is 0.005 Hz. Therefore, the horizontal specific force caused by the vehicle turning maneuver during the on-board measurement can be effectively weakened.
[0023] S2: The continuous gravity vector measurement data in the horizontal direction after low-pass filtering is divided into survey lines. The survey line from the starting point to the end point is called a forward survey line, and the survey line from the end point back to the starting point is called a reverse survey line. An adjacent pair of forward survey lines and reverse survey lines that go from the starting point to the end point and then back to the starting point is called a group of survey lines.
[0024] S3: The measurement data on the first group of survey lines are numbered according to mileage, where the starting point is 0 by default and the end point is the total mileage of the survey line by default.
[0025] S4: Align the measurement data at the same position of the forward survey line and the reverse survey line in the first group of survey lines by interpolation.
[0026] S5: The Kalman filter method is used to fuse the measurement data at the same position of the first group of measurement lines to obtain the effective measurement results of the first group of measurement lines.
[0027] In the S2-S5 stage, the influence of the heading angle change of the vehicle-mounted survey line on the horizontal gravity vector measurement results is suppressed by means of error cancellation. Taking advantage of the fact that the heading angle at the same position of the vehicle-mounted forward and reverse survey lines differs by 180°, the measurement error of the horizontal accelerometer at the same position of the vehicle-mounted forward and reverse survey lines has a mirror symmetry phenomenon. The horizontal accelerometer measurement error can be cancelled by fusing the measurement data at the same position of the vehicle-mounted forward and reverse survey lines using Kalman filtering. The advantage of this method is that there is no need to specially calibrate the random constant zero bias of the accelerometer. In the absence of on-site calibration, the influence of complex heading angle changes on the horizontal direction measurement results is suppressed, the problem of the difficulty in effectively estimating and compensating the random constant zero bias of the horizontal accelerometer is solved, and the interference of the complex shape survey line error on the vehicle-mounted strapdown gravity vector solution is eliminated. The method is innovative and has engineering application value.
[0028] S6: S3 to S5 are repeated in a loop, and the remaining forward survey lines and reverse survey lines are processed in turn, and finally the horizontal dynamic error of the vehicle-mounted strapdown gravity vector measurement is suppressed.
[0029] The technical solution of the present invention will be clearly and completely described below through embodiments: The horizontal dynamic error of the vehicle-mounted strapdown gravity vector repetitive line measurement is mainly affected by the centripetal acceleration generated by the vehicle turning maneuver and the change of the heading angle of the vehicle-mounted measurement line.
[0030] For vehicle-mounted gravity vector measurement, the measuring vehicle will inevitably face turning when driving on the road. Even if the measuring vehicle is driving at a constant speed, the centripetal acceleration will be generated due to the curved motion during the turning process, such as Figure 2 As shown, the horizontal arrow represents the centripetal acceleration a n A specific force measurement will be generated in the horizontal direction, so it is necessary to study the effect of the centripetal acceleration on the horizontal specific force during vehicle turning.
[0031] The formula for calculating centripetal acceleration is: a n =v·ω (2) Among them, a n represents the centripetal acceleration, v represents the linear velocity, which is the forward velocity of the car in this scenario, and ω represents the turning angular velocity.
[0032] The following is a set of simulation experiments to explore the amplitude of the horizontal specific force generated by the car turning process, such as Figure 3As shown in the figure, the simulation trajectory is composed of three sections, including two straight lines and a 50° arc, which simulates the turning process of a car on the road; the car travels along the trajectory at a constant forward speed of 20m / s, which is a commonly used speed in vehicle-mounted measurements; this simulation experiment sets three different turning angular velocities, namely 5° / s, 10° / s and 25° / s, and turns 50° with different turning radii at the same speed to simulate various slow and sharp curves faced by cars when driving on the road.
[0033] The simulation results of the horizontal specific force directly caused by the centripetal acceleration generated by the above three turning modes are as follows: Figure 4 As shown in the figure, the simulation results of 5° / s, 10° / s and 25° / s are plotted respectively. In order to better demonstrate the effect, this paper only retains the simulation results of 300s before and after the turn. It can be seen that the centripetal acceleration during the turn will cause a very large specific force measurement value in the horizontal direction. Under the same vehicle speed, the greater the turning angular velocity, the greater the centripetal acceleration and the greater the horizontal specific force.
[0034] Table 1 lists the theoretical values of the centripetal acceleration calculated using formula (2), and also statistics the maximum absolute value of the horizontal specific force simulation results. It can be seen that the simulation results are basically consistent with the theoretical calculation results, proving that the simulation results are reliable.
[0035] Table 1 Statistical table of centripetal acceleration theoretical values and horizontal specific force simulation results
[0036] Figure 5 From top to bottom, the simulation results of the eastward specific force and the northward specific force directly caused by the centripetal acceleration when the vehicle turns are plotted respectively. They are the components of the horizontal specific force simulation results in the eastward and northward directions. It can also be concluded from the figure that when the vehicle speed is the same, the greater the angular velocity, the greater the eastward and northward horizontal specific forces.
[0037] In order to suppress the influence of the centripetal acceleration generated by the vehicle turning maneuver on the horizontal specific force measurement, according to the scheme proposed by the present invention, a low-pass filter with a filtering period of 200s is selected here. Figure 5 The simulation results are low-pass filtered.
[0038] Figure 6The 200s low-pass filtering results of the east and north relative forces caused by the centripetal acceleration when the vehicle turns are shown. It can be seen from the figure that after low-pass filtering, the amplitudes of the east and north relative forces are greatly reduced, and there is no obvious difference in the relative forces under different turning angular velocities, which is due to the filtering effect of the low-pass filter on high-frequency noise. According to formula (1), the relationship between angular velocity and frequency can be calculated. It can be seen that the frequencies corresponding to angular velocities of 5° / s, 10° / s and 25° / s are 0.0139Hz, 0.0278Hz and 0.0694Hz respectively, and the cutoff frequency of the 200s low-pass filter is 0.005Hz. The frequencies of the three angular velocities that the vehicle may reach when turning are all within the band stop range of the 200s low-pass filter. Therefore, the horizontal relative force caused by the vehicle turning maneuver during the on-board measurement process can be effectively weakened:
[0039] Table 2 summarizes the maximum absolute values of the eastward and northward relative force simulation results before and after low-pass filtering. It can be seen that after low-pass filtering, the maximum absolute values of the eastward and northward relative force simulation results are reduced from 7.80 m / s to 2 and 4.16m / s 2 Respectively reduced to 0.12m / s 2 and 0.06m / s 2 .
[0040] Table 2 Statistics of the maximum absolute values of the simulation results of the eastward and northward specific forces before and after using the low-pass filter
[0041] Therefore, the 200s low-pass filter can suppress the influence of the centripetal acceleration generated by the vehicle turning maneuver on the horizontal specific force measurement.
[0042] Vehicle-mounted gravity measurement is generally carried out on a road survey line in a round trip, so there will be positive and negative survey lines in opposite directions. The heading angle change of the vehicle-mounted survey line will lead to a complex form of accelerometer measurement error projection in the horizontal direction. This error is difficult to effectively estimate. However, the heading angles of the vehicle-mounted positive and negative survey lines differ by 180°, so the horizontal accelerometer measurement errors of the positive and negative survey lines have a mirror symmetry phenomenon, such as Figure 7 As shown, the accelerometer measurement error in the horizontal direction is E and N It is approximately mirror-symmetric about the zero coordinate axis.
[0043] The reason for this phenomenon is that after the vehicle turns around, the heading angle at the same position on the forward and reverse measuring lines differs by about 180°. Since the z-axis of the gravimeter basically maintains a soaring posture during the measurement process, the accelerometer in the horizontal direction has a random constant zero bias of ▽ E and NThe redistribution of is closely related to the heading angle. A heading angle difference of about 180° will cause the sign of the projection to change, thus producing an effect of approximate mirror symmetry about the zero coordinate axis.
[0044] Based on this feature, the present invention suppresses the influence of the change of the heading angle of the vehicle-mounted survey line on the horizontal gravity vector measurement result by means of error cancellation. The main idea of this method is to align the gravity vector solution results of the forward and reverse survey lines according to the position and then perform Kalman filter fusion, and use the error characteristics of approximately mirror-symmetrical changes to cancel the horizontal accelerometer measurement errors at the same position of the forward and reverse survey lines to achieve the purpose of error suppression.
[0045] The measured vehicle data is used to verify the horizontal accelerometer measurement error cancellation method based on the fusion of forward and reverse measurement lines.
[0046] A vehicle-mounted strapdown gravity vector measurement test was carried out in Changsha city. Figure 8 The test line setting of the test is shown. The length of a single test line from the starting point to the end point in the figure is about 15km. The measuring vehicle made three round trip measurements along this section of road, and a total of 6 valid test line data were obtained. The vehicle test was carried out during a period of low traffic volume. The vehicle speed was set at 60km / h through the cruise control function of the measuring vehicle. Driving operations that affect the measurement stability, such as acceleration, braking or lane changing, were avoided as much as possible unless there were special circumstances.
[0047] (1) Direct solution using traditional methods The gravity vector measurement results directly calculated by traditional methods, such as Figure 9a and Figure 9b As shown in the figure, it can be seen that the horizontal gravity vector measurement results contain the measurement error of the mirror change. It should be noted that in vehicle-mounted measurement, considering that the survey line may not be evenly distributed and monotonically changing in the latitude and longitude directions, the measurement results are usually plotted with the mileage of the measuring point relative to the endpoint of the survey line as the horizontal coordinate.
[0048] The accuracy evaluation of the horizontal gravity vector measurement results directly solved by the above traditional method is shown in Table 3.
[0049] Table 3 Statistics of the internal accuracy of horizontal gravity vector measurement results directly solved by traditional methods (mGal)
[0050] (2) Using the forward and reverse line measurement results of the present invention to fuse the results The above results are further processed by using the horizontal accelerometer measurement error cancellation method based on forward and reverse measurement line fusion proposed in the present invention. Fig.10a and Fig.10bis the processed gravity vector measurement result. In actual measurement, the survey vehicle conducts round-trip measurement on the road. The odd-numbered survey lines and even-numbered survey lines represent the forward survey lines and the reverse survey lines, respectively. Therefore, in the following description, survey line 12 will be used to refer to the fusion result of the forward survey line 1 and the reverse survey line 2. Fig.10a and Fig.10b It can be seen that after processing the six forward and reverse measurement lines, three forward and reverse fusion measurement lines are obtained. The accuracy of the processed gravity vector measurement results is shown in Table 4.
[0051] Table 4 Internal accuracy statistics of horizontal gravity vector measurement results after forward and reverse line fusion processing (mGal)
[0052] Compare Figure 9a , 9b and Fig.10a , 10b From the horizontal gravity vector measurement results before and after the forward and reverse survey line fusion processing, it can be found that the measurement error of the mirror change in the horizontal gravity vector measurement results after processing is effectively suppressed, and the internal consistency between the horizontal gravity vector measurement results is improved, showing an obvious change trend; by comparing the internal consistency accuracy of the gravity vector measurement results before and after the forward and reverse survey line fusion processing in Table 3 and Table 4, it can be found that the total internal consistency accuracy in the horizontal direction is improved, and the eastward accuracy is increased from 3.04mGal to 0.90mGal, and the northward accuracy is increased from 4.83mGal to 2.41mGal.
[0053] The results of real vehicle-mounted tests show that the horizontal accelerometer measurement error cancellation method based on the fusion of forward and reverse measurement lines proposed in the present invention is effective. By fusing the measurement results of the horizontal direction of the forward measurement line and the reverse measurement line through Kalman filtering, the influence of the heading angle change of the vehicle-mounted measurement line on the horizontal gravity vector measurement result can be suppressed.
Claims
1. A method for suppressing horizontal dynamic errors in vehicle-mounted strapdown gravity vector repetitive line measurement, characterized in that: Firstly, the original signal of the horizontal direction of the vehicle-mounted strapdown gravity vector measurement is low-pass filtered with a period of 200s to suppress the influence of the centripetal acceleration generated by the vehicle turning maneuver on the horizontal specific force measurement. Then, the horizontal measurement results of the forward and reverse survey lines are fused through Kalman filtering to cancel the measurement error of the horizontal accelerometer and suppress the influence of the heading angle change of the vehicle-mounted survey line on the horizontal gravity vector measurement result.
2. The horizontal dynamic error suppression method of vehicle-mounted strapdown gravity vector repetitive line measurement according to claim 1 is characterized in that: S1: low-pass filter the continuous raw signal of the horizontal direction of the vehicle-mounted strapdown gravity vector measurement, and the period of the low-pass filter is selected as 200s; S2: The continuous gravity vector measurement data in the horizontal direction after low-pass filtering is divided into measurement lines. The measurement line from the starting point to the end point is called the forward measurement line, and the measurement line from the end point back to the starting point is called the reverse measurement line. A pair of adjacent forward measurement lines and reverse measurement lines that go from the starting point to the end point and then back to the starting point is called a group of measurement lines. S3: The measurement data on the first group of survey lines are numbered according to the mileage, where the starting point is 0 by default and the end point is the total mileage of the survey line by default; S4: aligning the measurement data at the same position of the forward survey line and the reverse survey line in the first group of survey lines by interpolation method; S5: using the Kalman filter method to fuse the measurement data at the same position of the first group of measurement lines, and obtain the effective measurement results of the first group of measurement lines; S6: S3 to S5 are repeated in a loop, and the remaining forward survey lines and reverse survey lines are processed in turn, and finally the horizontal dynamic error of the vehicle-mounted strapdown gravity vector measurement is suppressed.