A precision testing method for an inertial navigation orbit detection device

Abstract

The present application belongs to the technical field of precision detection, and particularly relates to a precision testing method for an inertial navigation track detection device. The present application introduces a high-precision navigation-grade IMU as a precision evaluation reference benchmark, and under the same measurement environment and conditions, makes the navigation-grade IMU and a low-cost consumer-grade IMU to be tested participate in algorithm fusion in parallel, takes the post-processing result of the combined navigation mode constructed by the high-precision navigation-grade IMU as a reference true value, systematically evaluates the internal and external compliance precision of the low-cost consumer-grade IMU after algorithm fusion, and thus reveals the size of the random error and the measurement accuracy of the combined navigation detection device based on the low-cost IMU, so as to determine whether the detection device meets the application requirements or to formulate targeted optimization measures to improve the detection precision.

Description

Technical Field

[0001] This invention belongs to the field of precision testing technology, specifically relating to a precision testing method for an inertial navigation track testing device. Background Technology

[0002] The increasing speed and intelligence of rail transit have placed higher demands on track condition monitoring. In the field of high-speed railway inspection, non-contact measurement based on inertial navigation technology is a commonly used method, which involves using inertial sensors to measure track irregularities. This technology essentially leverages the relative measurement capabilities of inertia to characterize the internal geometric parameters of the track by calculating the changes in the relative positions and attitudes of various measurement points along the track.

[0003] Chinese patent CN113085948B discloses a comprehensive track inspection system that measures parameters such as track gauge and elevation using an inertial measurement unit (IMU) and laser sensors. Chinese patent CN119428790A discloses a vehicle-mounted track inspection system that employs a multi-sensor fusion approach, integrating fiber optic inertial navigation and a 3D camera to achieve complementary signal and image data. Chinese patent CN110658543A discloses a non-contact measurement method for detecting geometric parameters of high-speed railway tracks, using a combined navigation calculation of an inertial navigation system and a high-precision differential satellite positioning system to obtain the spatial position and attitude information of the IMU. These solutions often use navigation-grade fiber optic gyroscopes or laser gyroscopes, combined with quartz accelerometers to form the IMU, and then integrate a satellite navigation and positioning system to jointly construct a high-precision track irregularity measurement system. While these solutions can achieve high-precision measurement, the high cost of navigation-grade equipment presents challenges for market promotion and industrialization, especially in applications such as trams and industrial railways. Faced with this engineering bottleneck, developing low-cost inertial navigation trajectory detection equipment has become an inevitable trend, such as multi-sensor fusion navigation trajectory detection equipment based on low-cost consumer-grade MEMS IMUs. However, while achieving low cost, it is equally important to meet the measurement accuracy requirements of practical applications. Therefore, determining the measurement accuracy of such low-cost multi-sensor fusion navigation detection systems has become a prerequisite for engineering applications, but existing technologies lack corresponding accuracy evaluation and testing methods. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the technical problem to be solved by this invention is to provide a method for testing the accuracy of inertial navigation track detection equipment, and to systematically evaluate the accuracy of a low-cost IMU after the fusion of combined navigation mode algorithms.

[0005] The inertial navigation track testing equipment to be tested relies on a testing vehicle as its carrier, with various testing sensors installed on the vehicle body. The testing sensors mainly include an inertial measurement unit (IMU), a satellite receiver, an odometer, and a laser scanner. The testing equipment is also equipped with a data acquisition computer. The IMU integrates a three-axis gyroscope and a three-axis accelerometer to measure the motion of the vehicle on the railway track and outputs the vehicle's attitude, speed, and position information in real time through an inertial navigation algorithm. A satellite receiver measures the vehicle's speed and position in the navigation coordinate system in real time. The odometer uses a pulse-type rotary encoder, fixed to the vehicle's axles. Based on the pulse signals output when the wheels rotate and the wheel diameter data, it acquires the vehicle's speed information in real time. Combined with the attitude information output by the inertial navigation system, it performs dead reckoning and also outputs the vehicle's attitude, speed, and position information in real time. The laser scanner contains two two-dimensional laser rangefinders, installed symmetrically at their axis centers. They scan the left and right rails respectively, accurately measuring the rail gauge and acquiring the lateral and vertical distances of the laser relative to the rails in real time. The data acquisition computer is responsible for synchronously acquiring data from all sensors, running the integrated navigation algorithm software for data processing, and deriving track irregularity detection indicators based on the track geometry calculation principle.

[0006] The core idea of ​​this invention is to introduce a high-precision navigation-grade IMU as a reference benchmark for accuracy evaluation. Under the same measurement environment and conditions, the navigation-grade IMU and the low-cost consumer-grade IMU under test participate in algorithm fusion in parallel. The post-processing result of the combined navigation mode constructed by the high-precision navigation-grade IMU is used as the reference truth value to systematically evaluate the internal and external consistency accuracy of the low-cost consumer-grade IMU after algorithm fusion. Specifically, the steps include the following:

[0007] Step S1, inspect vehicle body design and manufacturing

[0008] The inspection vehicle is a T-shaped hand-pushed trolley, mainly composed of five components: longitudinal beams, cross beams, push rods, auxiliary structures, and wheel structure.

[0009] As a bridge connecting the detection sensors and the railway track, the detection vehicle body is responsible for transmitting the uneven forces of the track to the sensors. The vehicle body's ability to transmit track undulations as a rigid body, as well as its performance in maintaining stability and ensuring continuous contact with the track during the measurement process, all depend on the quality of the vehicle body's mechanical design and manufacturing.

[0010] Step S2, Sensor Selection and Installation

[0011] Two levels of inertial measurement units (IMUs) are selected: a navigation-grade IMU and a consumer-grade IMU. These two IMUs and other sensors are integrated and installed on the detection vehicle obtained in step S1, with the sensitive axis aligned with the vehicle coordinate system and rigidly connected.

[0012] Step S3, Test circuit selection

[0013] A section of the line was selected as the test object. The total length of the test line was 400-500m, including straight and curved sections.

[0014] Step S4, Field Test

[0015] Preparation phase: Place the testing vehicle on the track selected in step S3, turn on the power of the data acquisition system, and let it stand for 5 minutes to preheat;

[0016] Data acquisition phase: Start the synchronous trigger serial port acquisition software to synchronously acquire triaxial IMU data, RTK satellite positioning data, laser measurement data and odometer data, and store them in real time;

[0017] Measurement phase: After initializing the navigation system, the test start and end points are marked on the rail surface. The operator uses a push rod to push the test vehicle along the track at a constant speed of 3-5 km / h to perform repetitive measurements. After each time the vehicle reaches the end point, it stops for 30 seconds and then returns to the start point.

[0018] Final stage: After the measurement is completed, the equipment is powered off, the measurement data of the entire process is exported and stored for offline processing, and finally the testing vehicle is moved off the track.

[0019] Step S5, Internal Consistency Accuracy Analysis

[0020] S5.1 Reference Setting: The heading angle obtained by the autonomous alignment of the navigation-grade IMU is used as the reference to set the initial heading value of the consumer-grade IMU;

[0021] S5.2 Data Calculation: Offline combined navigation calculation is performed by integrating the two levels of IMUs with satellite data, odometry data, and laser data respectively;

[0022] S5.3 Repeatability Accuracy Test: In a multi-round-trip measurement scenario, four measurement processes, including two forward measurements and two reverse measurements, are randomly selected. The repeatability accuracy is evaluated by analyzing the data differences between the four measurements.

[0023] S5.4 Consumer-Grade IMU Test Results Analysis

[0024] In track irregularity detection, good roll angle repeatability indicates high superelevation measurement accuracy; good pitch angle repeatability indicates high accuracy in gauge distortion and track height irregularity measurement; and good position repeatability indicates high accuracy in horizontal and vertical track irregularity measurement.

[0025] The data underwent position alignment processing, with relevant data aligned based on the northward position. Position observations included eastward and celestial position differences. Attitude angle parameters included roll angle, pitch angle, and yaw angle differences. For the deviations in attitude angle and position measurement values ​​observed in the four measurements, statistical values ​​of each attitude angle parameter and position observation for the two IMUs were listed, including mean, maximum, and root mean square values. The consistency of the measurement results was statistically analyzed and evaluated.

[0026] Step S6, External Consistency Accuracy Analysis

[0027] External conformity accuracy characterizes the absolute measurement capability of track detection equipment, and is reflected in the statistical characteristics of the error between the measured value and the true value. Using the average value of multiple combined navigation measurements obtained from a navigation-grade IMU as the reference true value, an absolute error statistical analysis is performed on the combined navigation measurement results of the attitude and position of a consumer-grade IMU.

[0028] Preferably, in step S1, the main frame of the inspection vehicle is made of 6031 aluminum alloy, and the connecting parts are fixed by positioning pins to ensure that the deformation and repeated assembly errors of the inspection vehicle during dynamic measurement are strictly controlled within a limited range; the coplanarity error of the three running wheels is calibrated on a marble platform by adjusting the screw torque; ceramic wheels and axles are used, and the running wheel mechanism is screened by measuring the full circumference deformation of the wheels with a dial indicator; the auxiliary structure is a spring mechanical device that provides lateral extrusion force to the two longitudinal wheel sides, forcing the single wheel side to fit tightly against the rail and accurately sensing changes in the track trajectory; a push rod supports and fixes the inspection computer and is also used to manually push the inspection vehicle.

[0029] By adopting the above technical solution, the beneficial effect of this invention is that the accuracy testing method of the inertial navigation orbit detection device proposed in this invention can systematically evaluate the internal and external conformity accuracy of a low-cost consumer-grade IMU after the fusion of integrated navigation mode algorithms. This reveals the magnitude of the impact of random errors on the integrated navigation detection device based on the low-cost IMU and the measurement accuracy, so as to determine whether the detection device meets the application requirements or to formulate targeted optimization measures to improve the detection accuracy. Attached Figure Description

[0030] Figure 1 To detect the initial value of the heading angle of the vehicle on the track.

[0031] Figure 2 The relative deviations of the orbital attitude measured by IMU_N in three orbital operations.

[0032] Figure 3 The relative deviations of the orbital positions measured by IMU_N for three orbital maneuvers.

[0033] Figure 4The relative deviations of the orbital positions measured by the IMU_M in four orbital maneuvers.

[0034] Figure 5 The relative deviations of the orbital attitude measured by the IMU_M during four orbital operations.

[0035] Figure 6 The absolute deviation of the orbital attitude measured by IMU_N for three orbital operations.

[0036] Figure 7 The absolute deviation of the orbital position measured by IMU_N for three orbital runs.

[0037] Figure 8 The absolute deviation of the orbital attitude measured by the IMU_M in three orbital operations.

[0038] Figure 9 The absolute deviation of the orbital position measured by the IMU_M for three orbital runs. Detailed Implementation

[0039] The technical solution of the present invention will now be described more clearly and completely with reference to the accompanying drawings.

[0040] This embodiment introduces a high-precision navigation-grade IMU as a reference benchmark for accuracy evaluation. Under the same measurement environment and conditions, the navigation-grade IMU and the low-cost consumer-grade IMU under test participate in the algorithm fusion in parallel. The post-processing result of the combined navigation mode constructed by the high-precision navigation-grade IMU is used as the reference truth value to systematically evaluate the internal and external conformance accuracy of the low-cost consumer-grade IMU after algorithm fusion.

[0041] First, a standard T-shaped hand-push testing cart was designed and manufactured as the testing platform. This cart mainly consists of five components: longitudinal beams, transverse beams, push rods, auxiliary structures, and wheel structures. The cart body is made of 6031 aluminum alloy, a high-strength, one-piece machined material. Connecting parts are fixed with locating pins to ensure that deformation and repeated assembly errors during dynamic measurement are strictly controlled within limits. The coplanarity error of the three running wheels is calibrated on a marble platform by precisely adjusting the screw torque. Precision-machined ceramic wheels and axles are used, and the wheel deformation is measured with a dial indicator to screen qualified running wheel mechanisms. The auxiliary structure is a spring-loaded mechanical device that provides lateral compressive force to the longitudinal two wheel sides, forcing each wheel side to tightly conform to the rail, accurately sensing changes in the track trajectory. Handles are designed at both ends of the cart for easy transport, and the push rod angle is adjustable, saving effort and allowing for the installation of a testing computer. The traveling wheels of the inspection vehicle are insulated from the rail being tested by cylindrical ceramic wheels; the positioning wheels are insulated from the inside of the rail being tested by ceramic wheels; the insulation between the traveling wheel module and the gauge detection module and the longitudinal and transverse beams is achieved by adding insulating pads to the mounting surface; the insulation between the protruding parts at the bottom of the front and rear of the longitudinal beams and the protruding parts at the bottom of the transverse beams is achieved by adding insulating sleeves.

[0042] Two IMUs of different classes were selected: a navigation-grade fiber optic IMU (denoted as IMU_N) POS_TDK and a consumer-grade IMU (denoted as IMU_M) PA_GS02. The two IMUs (IMU_N and IMU_M) were rigidly mounted on the test vehicle, with their sensitive axes aligned with the vehicle's coordinate system and rigidly connected.

[0043] A section of the Haoji Railway in Xiangyang, China, was selected as the test line. This newly built line has good smoothness. The test line is 488m long, runs north-south (heading angle of approximately 262°), and includes both straight and curved sections.

[0044] 1. Testing Process

[0045] (1) Place the testing vehicle on the track, turn on the power of the data acquisition system, and let it stand for 5 minutes to preheat;

[0046] (2) Start the synchronous trigger serial port acquisition software to synchronously acquire triaxial IMU data, RTK satellite positioning data, laser measurement data and odometer data, and store them in real time;

[0047] (3) After the navigation system is initialized, the test start point and end point are marked on the rail surface. The operator pushes the test vehicle along the track at a constant speed of 3~5 km / h with the help of a push rod to perform repetitive tests. After reaching the end point each time, the vehicle stops for 30 seconds and then returns to the start point.

[0048] (4) After the test is completed, the equipment is powered off, the test data of the whole process is exported and stored for offline processing, and finally the track measurement and inspection vehicle is moved away from the track.

[0049] 2. Internal compliance accuracy analysis

[0050] In track irregularity detection, roll angle repeatability characterizes ultra-high measurement accuracy, while pitch angle repeatability characterizes the measurement accuracy of gauge distortion and track elevation irregularities. Furthermore, good position repeatability can effectively assess both horizontal and vertical track irregularities.

[0051] Before performing integrated navigation, the initial attitude angles of the IMU must be accurately determined, with the heading angle being particularly critical. Since consumer-grade IMUs lack autonomous alignment capabilities, this embodiment uses the heading angle obtained from the autonomous alignment of the navigation-grade IMU as a reference to set the initial heading value of the consumer-grade IMU. For example... Figure 1 As shown, the heading angle determined by the autonomous alignment is 262.2°, which is consistent with the actual route direction.

[0052] The two IMUs are used to perform offline integrated navigation calculations with RTK satellite data, odometry encoder data, and laser data, respectively, and the internal consistency accuracy of attitude and position of each integrated navigation result is calculated.

[0053] Internal compliance accuracy was determined using repeatability accuracy analysis. Four test cycles (two forward and two reverse) were randomly selected from a multi-round-trip test scenario. Repeatability accuracy was assessed by analyzing the differences between the four tests. The four tests were labeled T1, T2, T3, and T4, and the repeatability differences between any two tests were defined as T12, T13, T14, T23, T24, and T34, respectively. For example, the symbol T12 represents the attitude or position difference between the first and second tests. The data underwent position alignment, with relevant data aligned to the northward position. Position observations included eastward and celestial position differences; attitude angle parameters included roll, pitch, and yaw angle differences.

[0054] The integrated navigation results of the navigation-level inertial measurement unit (IMU_N) are shown in Figure 2 and... Figure 3 As shown, the results of three orbital tests on the navigation-grade IMU are as follows: For attitude repeatability, the differences in pitch and roll angles between any two tests are all less than 0.02°, and the differences in heading angles are all less than 0.05°. The repeatability accuracy of horizontal attitude (pitch and roll) is significantly better than that of azimuth attitude (heading). For position repeatability, the eastward position deviation is less than 5 mm, and the azimuth position deviation is less than 8 mm. The position repeatability accuracy shows a characteristic of eastward being better than azimuth.

[0055] The integrated navigation results of the consumer-grade inertial measurement unit (IMU_M) are as follows: Figure 4 and Figure 5 As shown, the results obtained from four orbital tests of the consumer-grade IMU exhibit systematic differences. Regarding attitude repeatability, the pitch and roll angle differences between any two tests were better than 0.5°, and the yaw angle differences were better than 2°, indicating that the horizontal attitude accuracy was significantly higher than the yaw attitude accuracy. Regarding position repeatability, the eastward position deviation was better than 1 m, and the azimuth position deviation was better than 0.5 m.

[0056] Table 1 - Repeatability statistics of relative attitude deviations in four IMU_M measurements

[0057]

[0058] Table 1 shows the statistical results of the consumer-grade inertial measurement unit (IMU_M) in four attitude repeatability tests. The mean (MEAN), maximum (MAX), and root mean square (RMS) roll angles are -0.025°, 0.380°, and 0.115°, respectively; the corresponding statistics for pitch angles are -0.043°, 0.505°, and 0.147°, respectively; and the corresponding statistics for yaw angles are -0.506°, 0.769°, and 0.473°, respectively.

[0059] Table 2 - Repeatability Statistics of Four Measurement Locations for IMU_M

[0060]

[0061] Table 2 presents the statistical results of four position repeatability tests for the consumer-grade inertial measurement unit (IMU_M), with the relevant data aligned to the north-facing position. The mean (MEAN), maximum (MAX), and root mean square (RMS) values ​​for the east-facing position are 0.180 m, 0.558 m, and 0.367 m, respectively. The mean (MEAN), maximum (MAX), and RMS values ​​for the celestial position are 0.144 m, 0.383 m, and 0.126 m, respectively. The inherent deviation of the celestial position of the track is relatively small, and the positioning accuracy in this direction is significantly better than that in the horizontal direction.

[0062] 3. External compliance accuracy analysis

[0063] In the field of track irregularity measurement, external conformity accuracy characterizes the absolute measurement capability of the system, manifested as the statistical characteristics of the error between the measured value and the true value. This embodiment uses the average value of three combined navigation tests obtained from a navigation-grade IMU as the reference true value to perform absolute error statistical analysis on the attitude and position measurement results of a consumer-grade IMU.

[0064] like Figure 6 and Figure 7 As shown, the position deviation calculated by the navigation-grade IMU integrated navigation is on the order of millimeters, with the absolute deviation of the eastward position being less than 2 mm and the absolute deviation of the celestial position being less than 3 mm. The absolute errors of the attitude angles are generally less than 0.05°, with the average absolute deviations of the roll and pitch angles being approximately 0.01° and the average absolute deviation of the heading angle being approximately 0.02°.

[0065] Table 3 - Repeatability Statistics of Absolute Deviation of IMU_N Measurement Locations for Three Measurements

[0066]

[0067] Based on the statistical average of the absolute position deviations of IMU_N in Table 3, the eastward deviations (MEAN, MAX, and RMS) of the combined navigation position deviations of the navigation-level IMUs are 0.310 mm, 2.138 m, and 0.702 m, respectively. The celestial deviations (MEAN, MAX, and RMS) are 0.232 mm, 1.90 m, and 0.862 m, respectively.

[0068] Table 4 - Repeatability statistics of absolute attitude deviation from three IMU_N measurements

[0069]

[0070] According to the statistical average values ​​of the absolute attitude deviations of IMU_N in Table 4, the roll angle deviations (MEAN, MAX, and RMS) of the navigation-level IMU's integrated navigation attitude deviations are 0.002°, 0.079°, and 0.004°, respectively; the pitch angle deviations (MEAN, MAX, and RMS) are 0.004°, 0.090°, and 0.006°, respectively; and the heading angle deviations (MEAN, MAX, and RMS) are 0.012°, 0.085°, and 0.010°, respectively.

[0071] like Figure 8 and Figure 9 As shown, the absolute attitude deviation of the consumer-grade IMU integrated navigation system is as follows: horizontal deviation is less than 0.5°, and heading deviation is less than 1°. The absolute position deviation is as follows: eastward deviation is less than 0.4 m, and azimuth deviation is less than 0.2 m.

[0072] Table 5 - Repeatability Statistics of Absolute Position Deviation of IMU_M Measurements (Three Times)

[0073]

[0074] According to the statistical average values ​​of the absolute deviations of the IMU_M position in Table 5, the eastward deviations MAX, RMS, and STD are 0.239 m, 0.201 m, and 0.137 m, respectively, while the celestial deviations MAX, RMS, and STD are 0.243 m, 0.119 m, and 0.059 m, respectively.

[0075] Table 6 - Repeatability statistics of absolute attitude deviation in three IMU_M measurements

[0076]

[0077] According to the statistical average values ​​of the absolute attitude deviations measured by the IMU_M three times in Table 6, the MAX, RMS and STD of the roll angle deviation are 0.310°, 0.094° and 0.075°, respectively, and the MAX, RMS and STD of the pitch angle deviation are 0.239°, 0.075° and 0.072°, respectively.

Claims

1. A method of testing the accuracy of an inertial navigation track detection device, characterized in that, The steps include the following: Step S1: Inspect the vehicle body design and manufacturing; The inspection vehicle is a T-shaped hand-pushed trolley, including longitudinal beams, cross beams, push rods, auxiliary structures, and wheel structure. Step S2, Sensor selection and installation; Two levels of inertial measurement units (IMUs) are used: a navigation-level IMU and a consumer-level IMU. The navigation-level IMU and the consumer-level IMU are installed on the test vehicle obtained in step S1. Step S3, test line selection; A section of the line was selected as the test object. The total length of the test line was 400-500m, including straight and curved sections. Step S4, field test; Preparation phase: Place the testing vehicle on the track selected in step S3, turn on the power of the data acquisition system, and let it stand for 5 minutes to preheat; Data acquisition phase: Start the synchronous trigger serial port acquisition software to synchronously acquire triaxial IMU data, RTK satellite positioning data, laser measurement data and odometer data, and store them in real time; Measurement phase: After initializing the navigation system, the test start and end points are marked on the rail surface. The operator uses a push rod to push the test vehicle along the track at a constant speed of 3-5 km / h to perform repetitive measurements. After each time the vehicle reaches the end point, it stops for 30 seconds and then returns to the start point. Final stage: After the measurement is completed, the equipment is powered off, the measurement data of the entire process is exported and stored for offline processing, and finally the testing vehicle is moved off the track. Step S5, Internal compliance accuracy analysis; S5.1 Reference Setting: The heading angle obtained by the autonomous alignment of the navigation-grade IMU is used as the reference to set the initial heading value of the consumer-grade IMU; S5.2 Data Calculation: The two levels of IMUs are combined with satellite data, odometry data and laser data for offline navigation calculation; S5.3 Repeatability Accuracy Test: In a round-trip measurement scenario, four measurement processes, including two forward measurements and two reverse measurements, are randomly selected. The repeatability accuracy is evaluated by analyzing the data differences between the four measurements. Analysis of S5.4 consumer-grade IMU test results; The data underwent position alignment processing, with relevant data aligned based on the northward position. Position observations included eastward and celestial position differences. Attitude angle parameters included roll angle, pitch angle, and yaw angle differences. For the deviations in attitude angle and position measurement values ​​observed in the four measurements, statistical values ​​of each attitude angle parameter and position observation for the two IMUs were listed, including mean, maximum, and root mean square values. The consistency of the measurement results was statistically analyzed and evaluated. Step S6, External compliance accuracy analysis; The average value of the combined navigation measurement results obtained by the navigation-grade IMU through more than four combined navigation measurements is used as the reference true value to perform absolute error statistical analysis on the combined navigation measurement results of the attitude and position of the consumer-grade IMU.

2. The accuracy testing method according to claim 1, characterized in that, In step S1, the main frame of the detection vehicle is made of 6031 aluminum alloy, and the connecting parts are fixed by positioning pins; ceramic wheels and axles are used; the auxiliary structure is a spring mechanical device to provide lateral extrusion force to the longitudinal two wheel sides.

Images