A method for generating three-dimensional trajectory of high-speed moving objects based on high-frequency two-dimensional data

Abstract

The present invention discloses a method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data, and the steps are as follows: building a system architecture, including a UWB positioning module and a data processing module; installing a UWB positioning tag on the moving object, and setting a UWB positioning base station; selecting the points where the object has deep scratches on the track and the points that can be accurately positioned, removing inappropriate points, and using the rest as feature points; the UWB positioning module sends the continuous sampling data of the high-speed moving object to the data processing module for calculation to obtain the two-dimensional continuous trajectory coordinates of the moving object; obtaining multiple groups of translation and rotation equations through multiple groups of corresponding feature points; and finally realizing the generation of a three-dimensional continuous trajectory of a high-speed moving object based on high-frequency two-dimensional data acquisition using multiple groups of translation and rotation equations and interpolation fitting. The present invention realizes the high-precision continuous trajectory data acquisition of high-speed moving objects in complex scenes, and improves the accuracy and continuity of data during data processing or trajectory prediction.

Description

Technical Field

[0001] The present invention relates to a technology for generating a three-dimensional trajectory of a moving object, in particular to a method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data. Background Art

[0002] For objects moving at high speed on complex fixed tracks such as indoor and outdoor bobsleigh and luge, coaches and players need to obtain the running parameters of the car in a timely manner, including accurate position, speed, angle and other information, in order to better train and adapt to the track. In this complex scenario where the speed, direction of movement and acceleration are constantly changing, extremely high positioning frequency is required to achieve high-precision centimeter-level continuous trajectory positioning.

[0003] In order to accurately and comprehensively locate the continuous trajectory of high-speed moving objects, the following technologies are currently considered:

[0004] 1. RFID radio frequency identification technology, barcode technology, and detection hole technology have slow recognition speed and long response time;

[0005] 2.UWB (Ultra Wide Band) positioning technology has a faster speed, but its measurement accuracy is negatively correlated with the frequency and is not suitable for high-speed mobile scenarios;

[0006] 3. Single laser ranging is very fast, but it can only obtain the distance of the reflection point, and cannot obtain the coordinate information;

[0007] 4. High-speed cameras can obtain positioning information, but they are expensive, consume a lot of power, and have complex data processing. In dark environments, icy environments, and environments with complex lighting, the feedback from high-speed cameras will be greatly disturbed. In addition, although high-speed cameras can obtain the three-dimensional coordinate information of objects by rotating, the image is blurred during rotation, and the rotation speed is very slow, so it is impossible to accurately measure high-speed moving objects. For positioning scenarios, the field of view angle is demanding, which cannot be met by existing high-speed cameras.

[0008] 5. RTK positioning technology, through the synchronous observation of GNSS base station and mobile station, uses carrier phase observation value to achieve high-precision positioning; its measurement accuracy is high, and the resolution can reach 2cm, but the measurement output frequency is low;

[0009] 6. Odometer technology uses data from motion sensors, cameras, etc. to calculate the motion path to achieve positioning, but it is costly, has poor real-time performance, and has low commercialization. Its measurement accuracy is low, the error is large, and the measurement output frequency is low.

[0010] Ultra Wide Band (UWB) technology is a wireless carrier communication technology that transmits data by sending nanosecond non-sinusoidal narrow pulses to measure the distance between devices. The ranging accuracy is extremely high and can reach the centimeter level. Compared with traditional narrowband systems, ultra-wideband systems have the advantages of short transmission and reception time, good anti-multipath effect, high system security, and low overall power consumption. Therefore, UWB technology can be used for fast, high-precision positioning, tracking and navigation of stationary or moving people and objects indoors and outdoors.

[0011] However, UWB technology also has corresponding defects. First, the measurement accuracy of UWB technology is negatively correlated with frequency. Under the condition of ensuring centimeter-level high measurement accuracy, the sampling frequency of the currently used UWB positioning system can only reach 200Hz at most. As a result, the data collected when the object moves at high speed on the track is only a set of discrete sampling sets, and the complete trajectory data cannot be obtained. Secondly, UWB three-dimensional positioning technology has relatively high requirements for hardware cost, hardware installation location, and ranging algorithm, and is relatively difficult. Summary of the invention

[0012] In view of the shortcomings of the existing UWB technology, such as the measurement accuracy and frequency cannot obtain complete trajectory data, and the high requirements and difficulty of the ranging algorithm, the technical problem to be solved by the present invention is to provide a method for generating three-dimensional trajectories of high-speed moving objects based on high-frequency two-dimensional data, which can realize high-precision continuous trajectory data collection of high-speed moving objects, and is conducive to improving the accuracy and continuity of data during data processing or trajectory prediction.

[0013] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0014] The present invention provides a method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data, comprising the following steps:

[0015] 1) Build the system architecture, including the UWB positioning module and the data processing module. The UWB positioning module includes the positioning tag and the UWB positioning base station. The UWB positioning tag and the UWB positioning base station communicate through the UWB technology, and the UWB positioning module and the data processing module use the network cable for data transmission.

[0016] 2) Install UWB positioning tags on moving objects and set up UWB positioning base stations in key areas. Use three-dimensional coordinate measuring instruments to obtain the full track trajectory data and establish a three-dimensional SLAM model of the track.

[0017] 3) Let the object slide on the track for multiple times, select the points with deep scratches on the track and the points that can be accurately located, remove the inappropriate points after filtering, and use the rest as the final feature points;

[0018] 4) The UWB positioning module sends the continuous sampling data of the high-speed moving object to the data processing module. The data processing module infers the relative position of the object to be positioned relative to each UWB positioning base station and obtains the two-dimensional continuous trajectory coordinates of the high-speed moving object;

[0019] 5) Find the corresponding multiple sets of feature points, map the two-dimensional trajectory to the xoy plane in the SLAM three-dimensional space through the rotation and translation equations, and obtain multiple sets of translation and rotation equations;

[0020] 6) Multiple sets of translation and rotation equations are obtained through multivariate linear regression to obtain the equation of the overall trajectory mapping in the xoy plane in a three-dimensional coordinate system. The coordinate data in the three-dimensional slam model of the track is used to interpolate and fit the two-dimensional trajectory equation in the xoy plane to complete the height information of the moving object, and finally realize the generation of three-dimensional continuous trajectory of high-speed moving objects based on high-frequency two-dimensional data acquisition.

[0021] The key areas in step 2) are the starting point, the end point, the junction of curves, the guardrail points and the complex areas.

[0022] The feature point selection process in step 3) is as follows;

[0023] 301) Select multiple points that can be accurately located, including the starting point, end point, curve junction, and railing point of the track to ensure the accuracy of feature point positioning, and the moving object will definitely pass through these feature points;

[0024] 302) In the preparation stage, after multiple sliding movements, the scratch depth information is recorded along the track, and the points where the depth exceeds the specified threshold are recorded and selected as feature points;

[0025] 303) At the same time, in the preparation stage, after multiple sliding movements, the three-dimensional data of multiple sliding movement trajectories are compared through UWB three-dimensional positioning technology to find multiple overlapping points and select them as feature points;

[0026] 304) The feature points selected by the above three schemes are collected together, and inappropriate points are filtered out, and the rest are used as the final selected feature point set.

[0027] The data processing module in step 4) adopts the TDOA positioning algorithm to infer the relative position of the object to be located relative to each UWB positioning base station by solving a group of nonlinear hyperbolic equations based on the difference in distance between the UWB positioning base station and the mobile UWB positioning tag.

[0028] Infer the relative position of the object to be located relative to each UWB positioning base station, specifically:

[0029] 401) i positioning base stations are set, and the coordinates of the positioning base station i are (xi, yi). The position of each positioning base station i is fixed during installation and deployment, and the coordinates are known;

[0030] 402) The coordinates of the desired positioning tag are Ro (x0, y0);

[0031] 403) The unique intersection point can be calculated using i circle equations, and the real-time position coordinates of the tag can be obtained through the calculation formula.

[0032] The relative position of the object to be located relative to each UWB positioning base station is inferred and calculated by the following formula:

[0033]

[0034] Among them, a, b, c belong to i, and i is an integer greater than or equal to 3.

[0035] Step 5) is as follows:

[0036] 501) obtaining continuous sampling data of the trajectory of a high-speed moving object on the track at a high sampling frequency through UWB positioning technology, selecting the two-dimensional coordinates of multiple feature points, and corresponding them one by one with the feature points of the three-dimensional coordinate system in the known three-dimensional slam model;

[0037] 502) Mapping multiple sets of two-dimensional trajectories to the xoy plane in the SLAM three-dimensional coordinate system, calculating the coordinate value of the point P'(x', y') in the plane of the SLAM three-dimensional coordinate system by knowing a point P'(x', y') in the coordinate system X'O'Y' after rotation and translation, and further obtaining the rotation and translation equations of multiple sets of feature points;

[0038] 503) Find the corresponding multiple sets of feature points, map the two-dimensional trajectory to the xoy plane in the SLAM three-dimensional space through the rotation and translation equation, and obtain multiple sets of translation and rotation equations;

[0039] 504) Based on the rotation and translation equations of multiple sets of feature points, the multivariate linear regression model is used to solve the overall rotation equation of the xoy coordinate system of the CAD 2D to 3D SLAM model.

[0040] The present invention has the following beneficial effects and advantages:

[0041] 1. The present invention proposes a method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data. Based on the characteristics of extremely high ranging accuracy and extremely high sampling frequency of UWB technology, it can realize high-precision continuous trajectory data collection of high-speed moving objects, which is beneficial to improving the accuracy and continuity of data during data processing or trajectory prediction;

[0042] 2. Based on the measured two-dimensional continuous data at a high sampling frequency, the present invention realizes the conversion of two-dimensional trajectory data to three-dimensional trajectory data through the mapping rotation equation of feature points, multivariate linear regression and interpolation fitting, and finally obtains the continuous three-dimensional trajectory positioning of high-speed moving objects;

[0043] 3. The present invention selects feature points through a variety of schemes to select a set of feature points. By selecting multiple sets of feature points, a more accurate regression fit is achieved, thereby ensuring the accuracy of three-dimensional positioning data calculation;

[0044] 4. The present invention is particularly effective in high-precision three-dimensional continuous trajectory positioning of high-speed moving objects in complex indoor and outdoor scenes (such as three-dimensional stereoscopic cyclotron tracks, etc.), which is specifically reflected in the continuity and accuracy of data collection;

[0045] 5. The method of the present invention can flexibly arrange positioning base stations and select important sections, such as curved sections, sprint sections, complex sections, etc., to achieve accurate analysis of key sections while monitoring the entire track. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a flow chart of a method for generating a three-dimensional trajectory of a high-speed moving object according to the present invention;

[0047] Figure 2 A system architecture for application of the present invention;

[0048] Figure 3 A schematic diagram of positioning base station settings in the method of the present invention;

[0049] Figure 4 It is an Xoy plane diagram of mapping the two-dimensional trajectory of the present invention to the three-dimensional space of SLAM. DETAILED DESCRIPTION

[0050] The present invention will be further described below in conjunction with the accompanying drawings.

[0051] The present invention provides a method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data. Figure 1 As shown, the following steps are included:

[0052] 1) Build the system architecture (such as Figure 2 As shown), including a UWB positioning module and a data processing module, wherein the UWB positioning module includes a positioning tag and a UWB positioning base station; the UWB positioning tag communicates with the UWB positioning base station through UWB technology, and the UWB positioning module and the data processing module use a network cable for data transmission; the data processing center module uses a physical server;

[0053] 2) Install UWB positioning tags on moving objects and set up UWB positioning base stations in key areas. Use three-dimensional coordinate measuring instruments to obtain the full track trajectory data and establish a three-dimensional SLAM model of the track.

[0054] 3) Let the object slide on the track for multiple times, select points with deep scratches on the track (several relatively deep points are sufficient) and points that can be accurately located, remove inappropriate points after filtering, and use the rest as the final feature points;

[0055] 4) The UWB positioning module sends the continuous sampling data of the high-speed moving object to the data processing module. The data processing module infers the relative position of the object to be positioned relative to each UWB positioning base station and obtains the two-dimensional continuous trajectory coordinates of the high-speed moving object;

[0056] 5) Find the corresponding multiple sets of feature points, map the two-dimensional trajectory to the xoy plane in the SLAM three-dimensional space through the rotation and translation equations, and obtain multiple sets of translation and rotation equations;

[0057] 6) Multiple sets of translation and rotation equations are obtained through multivariate linear regression to obtain the equation of the overall trajectory mapping in the xoy plane in a three-dimensional coordinate system. The coordinate data in the three-dimensional slam model of the track is used to interpolate and fit the two-dimensional trajectory equation in the xoy plane to complete the height information of the moving object, and finally realize the generation of three-dimensional continuous trajectory of high-speed moving objects based on high-frequency two-dimensional data acquisition.

[0058] The present invention designs a method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data. The method uses a positioning base station fixedly installed near the track and UWB positioning technology to obtain continuous sampling data of a high-speed moving object on the track at a high sampling frequency. Based on the continuous sampling data, the two-dimensional continuous trajectory coordinates of the moving object are obtained through the corresponding algorithm and the coordinate information of the base station; a feature point set is selected as a mapping standard through a variety of schemes, and the track three-dimensional SLAM model obtained by scanning the two-dimensional continuous trajectory coordinates of the moving object and the point cloud is used. Based on the feature point set, the two-dimensional trajectory is mapped to the xoy plane in the SLAM three-dimensional space, and the rotation equation of the CAD two-dimensional coordinate to the SLAM three-dimensional coordinate system is solved by a multivariate linear regression model, and the continuous trajectory information and height information of the moving object are supplemented by an interpolation fitting method. Based on the characteristic that the frequency of collecting two-dimensional data by UWB technology is much higher than that of three-dimensional data, the trajectory generation technology can obtain continuous and accurate trajectory positioning of high-speed moving objects, and the mapping conversion of two-dimensional data to three-dimensional data is completed by using a conversion equation, thereby obtaining three-dimensional continuous trajectory data at a high sampling frequency, so as to achieve real-time, continuous and high-precision positioning of high-speed moving objects. This technology can be used in complex scenes where indoor and outdoor objects move at high speed on three-dimensional spiral tracks, and can obtain continuous, real-time, and high-precision three-dimensional trajectory positioning of high-speed moving objects.

[0059] In step 2), the key areas refer to the starting point, the end point, the intersection of curves, the guardrail point and the complex areas. For example, if a track is a three-dimensional spiral track, the same xy value may correspond to multiple z values, and such a position belongs to a complex area.

[0060] Feature point selection:

[0061] The feature point selection process in step 3) is as follows;

[0062] 301) Select multiple points that can be accurately located, including the starting point, end point, curve junction, and railing point of the track to ensure the accuracy of feature point positioning, and the moving object will definitely pass through these feature points;

[0063] 302) In the preparation stage (that is, the preparation stage before the trajectory generation technology is officially used, in order to find the feature points needed for later use), after multiple sliding movements, the scratch depth information is recorded along the track, and the points with a larger depth (i.e., more passes) exceeding the specified threshold are recorded and selected as feature points; in this embodiment, the specified threshold is set to exceed the average value;

[0064] 303) At the same time, in the preparation stage, after multiple sliding movements, the three-dimensional data of multiple sliding movement trajectories are compared through UWB three-dimensional positioning technology to find multiple overlapping points and select them as feature points;

[0065] 304) The feature points selected by the above three schemes are collected together, and inappropriate points are filtered out, and the rest are used as the final selected feature point set.

[0066] UWB positioning algorithm:

[0067] The data processing module in step 4) adopts the TDOA positioning algorithm to infer the relative position of the object to be located relative to each UWB positioning base station by solving a group of nonlinear hyperbolic equations based on the difference in distance between the UWB positioning base station and the mobile UWB positioning tag.

[0068] The data processing module in step 4) adopts the TDOA positioning algorithm to infer the relative position of the object to be located relative to each UWB positioning base station by solving a group of nonlinear hyperbolic equations based on the difference in distance between the UWB positioning base station and the mobile UWB positioning tag.

[0069] Infer the relative position of the object to be located relative to each UWB positioning base station, specifically:

[0070] 401) i positioning base stations are set, and the coordinates of the positioning base station i are (xi, yi). The position of each positioning base station i is fixed during installation and deployment, and the coordinates are known;

[0071] 402) The coordinates of the desired positioning tag are Ro (x0, y0);

[0072] 403) The unique intersection point can be calculated using i circle equations, and the real-time position coordinates of the tag can be obtained through the calculation formula.

[0073] The relative position of the object to be located relative to each UWB positioning base station is inferred and calculated by the following formula:

[0074]

[0075] Among them, a, b, c belong to i, and i is an integer greater than or equal to 3.

[0076] In this embodiment, the coordinates of the moving object are obtained based on the signal arrival time difference ranging algorithm, and the specific implementation method is listed as follows:

[0077] like Figure 3As shown, three UWB positioning base stations are set up, the coordinates of the first UWB positioning base station 1 are (x1, y1), the coordinates of the second UWB positioning base station 2 are (x2, y2), and the coordinates of the third UWB positioning base station 3 are (x3, y3). The first to third UWB positioning base stations 1~3 are fixed in position and have known coordinates during installation and deployment. The coordinates of the desired positioning tag (i.e., the moving object) are Ro (x0, y0); the unique intersection can be calculated using the three circular equations, and the real-time position coordinates of the tag can be calculated using the following calculation formula:

[0078]

[0079] in, v is the speed of light, which is the propagation speed of the message sent by the positioning tag. t 1 ~ t 3 are the times when the three base stations receive the positioning tag information;

[0080] 3D continuous trajectory generation:

[0081] Step 5) If Figure 4 As shown, specifically:

[0082] 501) The continuous sampling data of the trajectory of the high-speed moving object on the track at a high sampling frequency is obtained through UWB positioning technology, and the two-dimensional coordinates of multiple feature points are selected to correspond one by one with the feature points of the three-dimensional coordinate system in the known three-dimensional slam model;

[0083] 502) Mapping multiple sets of two-dimensional trajectories to the xoy plane in the SLAM three-dimensional coordinate system, calculating the coordinate value of the point P'(x', y') in the plane of the SLAM three-dimensional coordinate system by knowing a point P'(x', y') in the coordinate system X'O'Y' after rotation and translation, and further obtaining the rotation and translation equations of multiple sets of feature points;

[0084]

[0085] Among them, x and y are the coordinate values ​​of point P in the required new coordinate system. is the angle of rotation of the coordinate system, m and n are the lengths of translation of the coordinate system in the x and y directions, respectively, as shown in Figure 4;

[0086] 503) Find the corresponding multiple sets of feature points, map the two-dimensional trajectory to the xoy plane in the SLAM three-dimensional space through the rotation and translation equation, and obtain multiple sets of translation and rotation equations;

[0087] 504) Based on the rotation and translation equations of multiple sets of feature points, the multivariate linear regression model is used to solve the overall rotation equation of the xoy coordinate system of the CAD 2D to 3D SLAM model.

[0088] Based on the rotation and translation equations of multiple sets of feature points, the multivariate linear regression model is used to solve the overall rotation equation from CAD two-dimensional to the xoy coordinate system in SLAM. Specifically, the regress function in MATLAB, that is, the least squares method, is used to solve it.

[0089] After obtaining the mapping conversion equation from the two-dimensional trajectory coordinates to the xoy coordinate system in the three-dimensional SLAM, since a known two-dimensional coordinate must be able to obtain a certain z value in the three-dimensional SLAM model, the z value can be obtained through the vertical correspondence. This part can use the griddata function to interpolate the known data points, because its data points (x, y) do not require regular arrangement, and the specific implementation is the scatteredInterpolant function of matlab. In this way, the height information of the mapped function is completed, and a complete and continuous trajectory generation model of mapping two-dimensional coordinates to three-dimensional coordinates is obtained. Based on high-frequency two-dimensional data acquisition, real-time, high-precision three-dimensional continuous trajectory data of high-speed moving objects can be obtained.

[0090] In terms of trajectory data processing, since the processing speed of two-dimensional data is much faster than that of three-dimensional data, the method of the present invention adopts a conversion calculation based on the two-dimensional data to obtain relatively more continuous and complete trajectory data, and then maps the two-dimensional trajectory data to the three-dimensional model coordinate system for data analysis and processing, and interpolates and fits the trajectory data to supplement the integrity of the trajectory data. In addition, for complex scenes of high-speed motion in indoor and outdoor three-dimensional stereoscopic orbits, the present invention can also achieve accurate trajectory positioning.

[0091] The two-dimensional data processing part uses the TDOA (Time Difference Of Arrival) positioning algorithm. Based on the distance difference between the positioning base station and the mobile tag, the relative position of the object to be located relative to each reference base station is inferred by solving a set of nonlinear hyperbolic equations. Since the flight speed of electromagnetic waves is known and constant, the flight distance and flight time can be converted to each other. If the transmission time is the same (or the interval is known), the flight time can be further converted to the signal arrival time. Therefore, it is only necessary to measure the difference in arrival time of the signal sent by the mobile tag to each positioning base station to obtain the corresponding distance difference.

[0092] The mapping of two-dimensional data to three-dimensional coordinate system is to map the two-dimensional trajectory to the xoy plane in the SLAM three-dimensional space, and solve the rotation equation of CAD two-dimensional coordinates to SLAM three-dimensional coordinate system through a multivariate linear regression model, where the parameter estimation of the multivariate regression model uses the least squares method.

[0093] The part of trajectory fitting after mapping uses two-dimensional interpolation to complete the trajectory information and height information of the moving object. For some equipment or measuring instruments, the location of the collected data points is not a regularly arranged grid structure. For such data, the scatter plot cannot well show its internal structure, so it is necessary to interpolate it to generate a regular grid image. Interpolation algorithms must first face a common problem - the selection of neighboring points. The selected neighboring points should be evenly distributed around the estimated point as much as possible, because the use of an appropriate amount of known points is very important for the accuracy of interpolation. When there are too many known points, the interpolation accuracy will decrease, because too much information will cover up useful information; the points formed by the selected neighboring points are unevenly distributed. If there are too many data points in a certain direction, or some data points in the point set are concentrated, or the data points are too far away, it will bring a large interpolation error. The griddata function will first perform Delaunay triangulation on the known data points. When the triangulation is completed, according to the values ​​of each triangle vertex, any point in the triangle area can be linearly interpolated. Centroid interpolation can be used for the interpolation points in the triangle, that is, only the influence of the nearest surrounding points on the interpolation point is considered, the relative position relationship between the interpolation point and the nearest surrounding points is determined, and the influence weight factor with the surrounding points is calculated to establish the linear interpolation formula.

[0094] Compared with three-dimensional positioning technology, UWB two-dimensional positioning technology has the advantages of lower hardware cost, lower hardware installation difficulty, less time required for ranging algorithm calculation and higher sampling frequency. The present invention adopts UWB two-dimensional positioning technology to measure the continuous trajectory data of high-speed moving objects, and uses the measured high-precision two-dimensional continuous data to obtain the three-dimensional continuous trajectory of the high-speed moving object in the corresponding SLAM coordinate system through the mapping rotation equation and multivariate linear regression of feature points and interpolation fitting. Finally, it can realize the generation of continuous three-dimensional trajectory of high-speed moving objects based on high-frequency two-dimensional data acquisition, especially for high-precision continuous trajectory positioning of high-speed moving objects in complex indoor and outdoor scenes (such as three-dimensional stereoscopic spiral orbits, etc.), the effect is particularly obvious. With respect to the selection of feature points, the present invention selects a set of feature points through a variety of schemes, and selects multiple groups of feature points to achieve more accurate regression fitting, thereby ensuring the accuracy of three-dimensional positioning data calculation.

Claims

1. A method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data, characterized in that The following steps are involved: 1) Build the system architecture, including the UWB positioning module and the data processing module. The UWB positioning module includes the positioning tag and the UWB positioning base station. The UWB positioning tag and the UWB positioning base station communicate through the UWB technology, and the UWB positioning module and the data processing module use the network cable for data transmission. 2) Install UWB positioning tags on moving objects and set up UWB positioning base stations in key areas. Use three-dimensional coordinate measuring instruments to obtain the full track trajectory data and establish a three-dimensional SLAM model of the track. 3) Let the object slide on the track for multiple times, select the points with deep scratches on the track and the points that can be accurately located, remove the inappropriate points after filtering, and use the rest as the final feature points; 4) The UWB positioning module sends the continuous sampling data of the high-speed moving object to the data processing module. The data processing module infers the relative position of the object to be positioned relative to each UWB positioning base station and obtains the two-dimensional continuous trajectory coordinates of the high-speed moving object; 5) Find the corresponding multiple sets of feature points, map the two-dimensional trajectory to the xoy plane in the SLAM three-dimensional space through the rotation and translation equations, and obtain multiple sets of translation and rotation equations; 6) Multiple sets of translation and rotation equations are obtained through multivariate linear regression to obtain the equation of the overall trajectory mapping in the xoy plane in a three-dimensional coordinate system. The coordinate data in the three-dimensional slam model of the track is used to interpolate and fit the two-dimensional trajectory equation in the xoy plane to complete the height information of the moving object, and finally realize the generation of three-dimensional continuous trajectory of high-speed moving objects based on high-frequency two-dimensional data acquisition.

2. The method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data according to claim 1, characterized in that: The key areas in step 2) are the starting point, the end point, the junction of curves, the guardrail points and the complex areas.

3. The method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data according to claim 1, characterized in that: The process of selecting feature points in step 3) is as follows; 301) Select multiple points that can be accurately located, including the starting point, end point, curve junction, and railing point of the track to ensure the accuracy of feature point positioning, and the moving object will definitely pass through these feature points; 302) In the preparation stage, after multiple sliding movements, the scratch depth information is recorded along the track, and the points where the depth exceeds the specified threshold are recorded and selected as feature points; 303) At the same time, in the preparation stage, after multiple sliding movements, the three-dimensional data of multiple sliding movement trajectories are compared through UWB three-dimensional positioning technology to find multiple overlapping points and select them as feature points; 304) The feature points selected by the above three schemes are collected together, and inappropriate points are filtered out, and the rest are used as the final selected feature point set.

4. The method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data according to claim 1, characterized in that: The data processing module in step 4) adopts the TDOA positioning algorithm to infer the relative position of the object to be located relative to each UWB positioning base station by solving a group of nonlinear hyperbolic equations based on the difference in distance between the UWB positioning base station and the mobile UWB positioning tag.

5. The method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data according to claim 4, characterized in that: Infer the relative position of the object to be located relative to each UWB positioning base station, specifically: 401) i positioning base stations are set, and the coordinates of the positioning base station i are (xi, yi). The position of each positioning base station i is fixed during installation and deployment, and the coordinates are known; 402) The coordinates of the desired positioning tag are Ro (x0, y0); 403) The unique intersection point can be calculated using i circle equations, and the real-time position coordinates of the tag can be obtained through the calculation formula.

6. The method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data according to claim 5, characterized in that: The relative position of the object to be located relative to each UWB positioning base station is inferred and calculated by the following formula: Among them, a, b, c belong to i, and i is an integer greater than or equal to 3.

7. The method for generating a three-dimensional trajectory of a high-speed moving object based on high-frequency two-dimensional data according to claim 1, characterized in that: Step 5) is as follows: 501) obtaining continuous sampling data of the trajectory of a high-speed moving object on the track at a high sampling frequency through UWB positioning technology, selecting the two-dimensional coordinates of multiple feature points, and corresponding them one by one with the feature points of the three-dimensional coordinate system in the known three-dimensional slam model; 502) Mapping multiple sets of two-dimensional trajectories to the xoy plane in the SLAM three-dimensional coordinate system, calculating the coordinate value of the point P'(x', y') in the plane of the SLAM three-dimensional coordinate system by knowing a point P'(x', y') in the coordinate system X'O'Y' after rotation and translation, and further obtaining the rotation and translation equations of multiple sets of feature points; 503) Find the corresponding multiple sets of feature points, map the two-dimensional trajectory to the xoy plane in the SLAM three-dimensional space through the rotation and translation equation, and obtain multiple sets of translation and rotation equations; 504) Based on the rotation and translation equations of multiple sets of feature points, the multivariate linear regression model is used to solve the overall rotation equation of the xoy coordinate system of the CAD 2D to 3D SLAM model.

Images