A method and device for smoothing a map trajectory

Abstract

The present application provides a method and device for smoothing a map trajectory, including: sampling the original map data by using a preset distance threshold to obtain a sampling data set, where the preset distance threshold is used to define the distance between adjacent two points; performing polynomial fitting on the sampling data set by using the least square method according to the moving average idea to obtain a fitting data set; performing exponential smoothing on the fitting data set to obtain trajectory smoothing data. By adopting the idea of moving average and performing step-by-step smoothing, the best function matching can be carried out on the data, ensuring the true expression of the data by the fitting curve, being able to reflect the comprehensive trend of the data, making the positioning data provided by the produced map smoother and more accurate, and greatly improving the smoothness of the operation of intelligent driving vehicles.

Description

Technical Field

[0001] The present application relates to the technical field of data processing, and in particular, to a method and device for smoothing a map trajectory. Background Art

[0002] Autopilot has high requirements for maps, and requires the map to cooperate with modules such as positioning, perception, motion planning, and control. High-precision maps have higher accuracy and more accurate positioning, can provide peripheral information beyond the perception distance, and provide a richer semantic information layer, including road markings, signs, etc., which can assist the vehicle in motion planning and can better serve autonomous vehicles and machines.

[0003] Generally speaking, there will be certain errors in the path trajectory positioning data provided by the basic layer of high-precision maps. If abnormal noise points appear in the output trajectory points, or the curve of the trajectory is not smooth enough, it will cause problems such as jerks and inaccurate positioning during vehicle operation, resulting in driving failures and sudden stops, affecting the vehicle's autopilot, driving experience, and safety. Using Lagrange polynomials for function fitting of data is prone to overfitting and will form oscillations at the data boundaries, making it difficult to see the trend of the data; while using the cubic spline method, connecting each data point with a piecewise cubic function for curve fitting will make it difficult to perform iterative calculations on its piecewise function and cannot reflect the overall trend of the data.

[0004] Therefore, how to make the smoothing process of the map trajectory achieve better results is a technical problem that needs to be solved urgently by those skilled in the art. Summary of the Invention

[0005] In view of this, embodiments of the present application provide a method and device for smoothing a map trajectory, aiming to make the smoothing process of the map trajectory achieve better results.

[0006] In a first aspect, embodiments of the present application provide a method for smoothing a map trajectory, including:

[0007] Sampling the original map data using a preset distance threshold to obtain a sampling data set, where the preset distance threshold is used to specify the distance between adjacent points;

[0008] According to the moving average idea, using the least squares method to perform polynomial fitting on the sampling data set to obtain a fitting data set;

[0009] Performing exponential smoothing on the fitting data set to obtain trajectory smoothing data.

[0010] Optionally, the original map data includes Global Positioning System (GPS) positioning data and Real-Time Kinematic (RTK) positioning data.

[0011] Optionally, according to the moving average idea, using the least squares method to perform polynomial fitting on the sampling data set to obtain a fitting data set, including:

[0012] Set the smoothing interval length;

[0013] According to the smoothing interval length, select a first moving data set from the sampling data set;

[0014] Use the least squares method to perform polynomial fitting on the first moving data set to obtain first fitting data;

[0015] According to the smoothing interval length, select a second moving data set from the sampling data set;

[0016] Use the least squares method to perform polynomial fitting on the second moving data set to obtain second fitting data;

[0017] Generate a fitting data set according to the first fitting data and the second fitting data.

[0018] Optionally, the use of the least squares method to perform polynomial fitting on the first moving data set to obtain first fitting data includes:

[0019] Use the least squares method to obtain the fitting parameters for performing polynomial fitting on the first moving data set;

[0020] Use the fitting parameters for performing polynomial fitting on the first moving data set to perform polynomial fitting on the first moving data set to obtain first fitting data.

[0021] Optionally, the use of the least squares method to perform polynomial fitting on the second moving data set to obtain second fitting data includes:

[0022] Use the least squares method to obtain the fitting parameters for performing polynomial fitting on the second moving data set;

[0023] Use the fitting parameters for performing polynomial fitting on the second moving data set to perform polynomial fitting on the second moving data set to obtain second fitting data.

[0024] Optionally, the exponential smoothing of the fitting data set to obtain trajectory smoothing data includes:

[0025] Perform weighted average on the fitting data set to obtain the trajectory smoothing data.

[0026] In a second aspect, an embodiment of the present application provides a map trajectory smoothing device, including:

[0027] A sampling module, configured to sample the original map data by using a preset distance threshold to obtain a set of sampled data, where the preset distance threshold is used to define the distance between two adjacent points;

[0028] A fitting module, configured to perform polynomial fitting on the set of sampled data by using the least squares method according to the moving average idea to obtain a set of fitted data;

[0029] A smoothing module, configured to perform exponential smoothing on the set of fitted data to obtain trajectory smoothed data.

[0030] Optionally, the fitting module includes:

[0031] A length setting unit, configured to set the length of the smoothing interval;

[0032] A first selection unit, configured to select a first set of moving data from the set of sampled data according to the length of the smoothing interval;

[0033] A first fitting unit, configured to perform polynomial fitting on the first set of moving data by using the least squares method to obtain first fitted data;

[0034] A second selection unit, configured to select a second set of moving data from the set of sampled data according to the length of the smoothing interval;

[0035] A second fitting unit, configured to perform polynomial fitting on the second set of moving data by using the least squares method to obtain second fitted data;

[0036] A result unit, configured to generate a set of fitted data according to the first fitted data and the second fitted data.

[0037] In a third aspect, an embodiment of the present application provides a device, where the device includes a memory and a processor, the memory is used to store instructions or codes, and the processor is used to execute the instructions or codes so that the device executes the map trajectory smoothing method according to any one of the foregoing first aspects.

[0038] In a fourth aspect, an embodiment of the present application provides a computer storage medium, where code is stored in the computer storage medium, and when the code is run, the device running the code implements the map trajectory smoothing method according to any one of the foregoing first aspects.

[0039] The embodiments of the present application provide a method and device for smoothing a map trajectory. When executing the method, first, the original map data is sampled using a preset distance threshold to obtain a set of sampled data, where the preset distance threshold is used to define the distance between adjacent two points; then, according to the idea of moving average and using the least squares method, polynomial fitting is performed on the set of sampled data to obtain a set of fitted data; finally, exponential smoothing is performed on the set of fitted data to obtain trajectory smoothing data. In this way, by adopting the idea of moving average and performing step-by-step smoothing, the best function matching can be performed on the data, ensuring the true expression of the data by the fitted curve, being able to reflect the comprehensive trend of the data, making the positioning data provided by the produced map smoother and more accurate, and greatly improving the smoothness of the operation of intelligent driving vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] To more clearly illustrate the technical solutions in the embodiments or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0041] Figure 1 It is a flowchart of a method for the map trajectory smoothing method provided by the embodiments of the present application;

[0042] Figure 2 It is another flowchart of a method for the map trajectory smoothing method provided by the embodiments of the present application;

[0043] Figure 3 It is an original data graph of the map trajectory smoothing method provided by the embodiments of the present application;

[0044] Figure 4 It is a result data graph of the map trajectory smoothing method provided by the embodiments of the present application;

[0045] Figure 5 It is a schematic structural diagram of a map trajectory smoothing device provided by the embodiments of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] Generally speaking, there will be certain errors in the path trajectory positioning data provided by the basic layer of the high-precision map. If there are abnormal noise points in the output trajectory points, or the curve of the trajectory is not smooth enough, it will cause problems such as jerks and inaccurate positioning during vehicle operation, resulting in driving failures and sudden stops, affecting the vehicle's autonomous driving, as well as the driving experience and safety. Using Lagrange polynomials for function fitting of data is prone to overfitting, and oscillations will occur at the data boundaries, making it difficult to see the trend of the data; while using the cubic spline method, the piecewise cubic functions are used to connect each data point for curve fitting, which will make it difficult to perform iterative calculations on its piecewise functions and unable to reflect the overall trend of the data.

[0047] The method provided by the embodiments of the present application is executed by a computer device to achieve a better effect on the smoothing process of the map trajectory.

[0048] Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0049] See Figure 1 , Figure 1 which is a flowchart of a method for smoothing a map trajectory provided by an embodiment of the present application, including:

[0050] Step S101: Sample the original map data using a preset distance threshold to obtain a set of sampled data.

[0051] The preset distance threshold is used to specify the distance between adjacent points. If the data format of the current trajectory point is set to (o i , p i ), and the data format of the previous trajectory point is set to (o i-1 , p i-1 ), then the distance between this point and the previous point is the preset distance threshold, that is:

[0052]

[0053] Using the preset distance threshold, invalid and duplicate trajectory points in the original map data can be removed, and sampling can be completed to obtain a set of sampled data.

[0054] As a possible implementation manner, the original map data includes Global Positioning System positioning data (i.e., GPS positioning data) and Real-Time Kinematic positioning data (i.e., RTK positioning data). Among them, the GPS positioning data is obtained by the Global Positioning System through resolving the three-dimensional position and time of the GPS receiver by receiving signals from at least 4 satellites. The RTK positioning data is the coordinates obtained by using the carrier phase differential technology, a differential method for real-time processing of the carrier phase observations of two measurement stations, sending the carrier phase collected by the reference station to the user receiver, and performing differential resolution.

[0055] Step S102: According to the moving average idea, use the least squares method to perform polynomial fitting on the sampling data set to obtain a fitting data set.

[0056] The moving average number refers to a series of averages obtained by arithmetically averaging several items of data in a time series by using a step-by-step progressive method. If the number of averaged data items is N, it is called the N-period (item) moving average. The moving average prediction method slides a certain number of average items on the time series to obtain a series of averages (also called smoothed values). These averages can not only eliminate or weaken the irregular fluctuations in the time series, but also reveal the change trend of the phenomenon.

[0057] Using the above idea, a step-by-step progressive method can be adopted to perform polynomial fitting on the sampling data set to obtain a fitting data set. During the fitting process, the least squares method can be used to determine the fitting parameters.

[0058] Adopting the moving average idea, setting a smoothing interval, and performing step-by-step smoothing can perform the best function matching on the data, ensure the true expression of the fitting curve for the data, can reflect the comprehensive trend of this section of data, and has a strong geometric meaning.

[0059] Step S103: Perform exponential smoothing on the fitting data set to obtain trajectory smoothing data.

[0060] The first-order exponential smoothing is actually a weighted average of historical data, and it can be used for short-term prediction of any time series that has no obvious function law but does have a certain correlation before and after. By means of exponential smoothing, the trajectories at previous moments in the fitting data set are weighted and averaged, which can effectively smooth the curve.

[0061] In summary, this embodiment adopts the moving average idea and performs step-by-step smoothing, which can perform the best function matching on the data, ensure the true expression of the fitting curve for the data, can reflect the comprehensive trend of the data, make the positioning data provided by the produced map smoother and more accurate, and greatly improve the smoothness of the operation of intelligent driving vehicles.

[0062] In the embodiment of the present application, the aboveFigure 1 There are multiple possible implementation manners for the described steps, which are introduced separately below. It should be noted that the implementation manners given in the following introduction are only exemplary descriptions and do not represent all the implementation manners of the embodiments of this application.

[0063] See Figure 2 , which is another method flowchart of the map trajectory smoothing method provided by the embodiments of this application, including:

[0064] Step S201: Sample the original map data by using a preset distance threshold to obtain a set of sampled data.

[0065] Sample the original map data by using a preset distance threshold to obtain a set of sampled data. The set of sampled data can be expressed as J = {(x1, y1), (x2, y2),...(x i , y i )}. See Figure 3 , Figure 3 , which is the original data graph of the map trajectory smoothing method provided by the embodiments of this application, and the threshold setting range for sampling is (0.05m, 1m).

[0066] Step S202: Set the length of the smoothing interval.

[0067] The length of the smoothing interval is used to represent the length of the data that needs to be moved for fitting. For example, select a point in the J set, and set the distance of z consecutive points before and after this point as the length of the smoothing interval, that is, the length of the smoothing interval n = 2z + 1.

[0068] Step S203: Select a first set of moving data in the set of sampled data according to the length of the smoothing interval.

[0069] Select the data with the length of the smoothing interval in the set of sampled data as the first set of moving data. For example, select a point x m in the J set, and record z consecutive points before and after this point as the first set of moving data: Z = (x m-z , x m-z+1 , x m-z+2 ,..., x m , x m+1 ,...x m+z ).

[0070] Step S204: Perform polynomial fitting on the first set of moving data by using the least square method to obtain the first fitted data.

[0071] First, use the least squares method to obtain the fitting parameters for polynomial fitting of the first moving data set; then, use the fitting parameters for polynomial fitting of the first moving data set to perform polynomial fitting on the first moving data set to obtain the first fitting data.

[0072] Specifically, the k-th degree polynomial fits the data points in the interval as y = a0 + a1x + a2x 2 +... + a k x k . More than n equations are combined into a k-element linear equation system. For the equation system to have a solution, n should be greater than or equal to k, and the fitting parameter A is determined by least squares fitting. Thus, we get:

[0073]

[0074] Represented by a matrix as Y (2z+1)×1 = X (2z+1)×k ·A (k+1)×1 + E (2z+1)×1 .

[0075] The residual e should be as small as possible, and the least squares solution of A is obtained as A = (X T ·X) -1 ·X T ·Y.

[0076] The model predicted value of Y is Y = X·A = X·(X T ·X) -1 ·Y = B·Y.

[0077] At this time, the coefficient B = X·(X T ·X) -1 ·X T .

[0078] Step S205: According to the smoothing interval length, select a second moving data set from the sampling data set.

[0079] Combining the idea of moving average, perform sequential shifting of the J set period by period, that is, move one position to the left or right in the sampling data set to select the second moving data set. As a possible implementation, it can be moved multiple times according to the data length until the J set is traversed.

[0080] Step S206: Use the least squares method to perform polynomial fitting on the second moving data set to obtain the second fitting data.

[0081] Using the least squares method, obtain the fitting parameters for polynomial fitting of the second moving data set; use the fitting parameters for polynomial fitting of the second moving data set to perform polynomial fitting on the second moving data set to obtain second fitting data. The specific steps are similar to those in step S204 and will not be elaborated here.

[0082] Step S207: Generate a fitting data set according to the first fitting data and the second fitting data.

[0083] By combining the results of multiple movements until the J set is traversed in a single set, a fitting data set can be obtained, effectively smoothing the curve.

[0084] Step S208: Perform weighted averaging on the fitting data set to obtain the trajectory smoothing data.

[0085] Set the fitting data set as F and use the formula F t+1 = aY t +(1 - a)F t to perform exponential smoothing on it, that is, F t+1 = aY t + a(1 - a)Y t-1 + a(1 - a) 2 Y t-2 + a(1 - a) 3 Y t-3 +...+(1 - a) t-1 Y1.

[0086] Among them, F t+1 is the predicted value at time t + 1, Ft is the predicted value at time t, a is the coefficient of exponential smoothing, Y t is the actual value at time t, Y t-1 is the actual value at time t - 1, Y t-2 is the actual value at time t - 2, Y t-3 is the actual value at time t - 3, and Y1 is the starting value.

[0087] In summary, this embodiment further describes a method for polynomial fitting using the idea of moving average, which can perform the best function matching on data, ensure the true expression of the fitting curve for data, reflect the comprehensive trend of data, and make the positioning data provided by the produced map smoother and more accurate, greatly improving the smoothness of the operation of intelligent driving vehicles.

[0088] The above are some specific implementation manners of the map trajectory smoothing method provided by the embodiments of the present application. Based on this, the present application also provides a corresponding device. The device provided by the embodiments of the present application will be introduced from the perspective of functional modularization below.

[0089] See Figure 5 the structural schematic diagram of the map trajectory smoothing device shown in Figure 5 . The device includes a sampling module 501, a fitting module 502, and a smoothing module 503.

[0090] The sampling module 501 is used to sample the original map data by using a preset distance threshold to obtain a sampling data set, and the preset distance threshold is used to define the distance between adjacent two points;

[0091] The fitting module 502 is used to perform polynomial fitting on the sampling data set by using the moving average idea and the least square method to obtain a fitting data set;

[0092] The smoothing module 503 is used to perform exponential smoothing on the fitting data set to obtain trajectory smoothing data.

[0093] As a possible implementation manner, the original map data includes global positioning system positioning data and real-time kinematic carrier phase differential positioning data.

[0094] As a possible implementation manner, the fitting module 502 includes:

[0095] a length setting unit for setting the length of the smoothing interval;

[0096] a first selection unit for selecting a first moving data set from the sampling data set according to the length of the smoothing interval;

[0097] a first fitting unit for performing polynomial fitting on the first moving data set by using the least square method to obtain first fitting data;

[0098] a second selection unit for selecting a second moving data set from the sampling data set according to the length of the smoothing interval;

[0099] a second fitting unit for performing polynomial fitting on the second moving data set by using the least square method to obtain second fitting data;

[0100] a result unit for generating a fitting data set according to the first fitting data and the second fitting data. The embodiments of the present application also provide corresponding devices and computer storage media for implementing the solutions provided by the embodiments of the present application.

[0101] As a possible implementation manner, the first fitting unit includes:

[0102] a first parameter component for obtaining fitting parameters for performing polynomial fitting on the first moving data set by using the least square method;

[0103] A first fitting component, configured to perform polynomial fitting on the first mobile data set by using fitting parameters for polynomial fitting of the first mobile data set, so as to obtain first fitting data.

[0104] As a possible implementation manner, the second fitting unit includes:

[0105] A second parameter component, configured to obtain fitting parameters for polynomial fitting of the second mobile data set by using the least squares method;

[0106] A second fitting component, configured to perform polynomial fitting on the second mobile data set by using fitting parameters for polynomial fitting of the second mobile data set, so as to obtain second fitting data.

[0107] As a possible implementation manner, the smoothing module 503 includes:

[0108] A weighted average module, configured to perform weighted average on the fitting data set to obtain the trajectory smoothing data.

[0109] Wherein, the device includes a memory and a processor, the memory is configured to store instructions or codes, and the processor is configured to execute the instructions or codes, so that the device executes the map trajectory smoothing method according to any embodiment of the present application.

[0110] Codes are stored in the computer storage medium, and when the codes are run, the device running the codes implements the map trajectory smoothing method according to any embodiment of the present application.

[0111] In the embodiments of the present application, the "first", "second" (if any) in the names such as "first" and "second" are only used as name identifiers and do not represent the first and second in order.

[0112] Through the description of the above embodiments, those skilled in the art can clearly understand that all or part of the steps in the above embodiment methods can be implemented by means of software plus a general hardware platform. Based on such an understanding, the technical solution of the present application can be embodied in the form of a software product, and the computer software product can be stored in a storage medium, such as read-only memory (ROM) / RAM, magnetic disk, optical disc, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network communication device such as a router) to execute the methods according to various embodiments or some parts of the embodiments of the present application.

[0113] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the device embodiments, since they are basically similar to the method embodiments, they are described relatively simply. For the relevant parts, reference can be made to the descriptions in the method embodiments. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative work.

[0114] The above description is only an exemplary embodiment of the present application and is not intended to limit the protection scope of the present application.

Claims

1. A method for smoothing a map trajectory, characterized in that, The method includes: Sampling the original map data by using a preset distance threshold to obtain a set of sampled data, where the preset distance threshold is used to define the distance between adjacent two points; Performing polynomial fitting on the set of sampled data by using the least squares method according to the moving average idea to obtain a set of fitted data; Performing exponential smoothing on the set of fitted data to obtain trajectory smoothing data; The performing polynomial fitting on the set of sampled data by using the least squares method according to the moving average idea to obtain a set of fitted data includes: Setting the length of the smoothing interval; Selecting a first set of moving data from the set of sampled data according to the length of the smoothing interval; Performing polynomial fitting on the first set of moving data by using the least squares method to obtain first fitted data; Selecting a second set of moving data from the set of sampled data according to the length of the smoothing interval; Performing polynomial fitting on the second set of moving data by using the least squares method to obtain second fitted data; Generating a set of fitted data according to the first fitted data and the second fitted data.

2. The method according to claim 1, wherein The original map data includes global positioning system (GPS) positioning data and real-time kinematic (RTK) positioning data.

3. The method according to claim 1, characterized in that, The performing polynomial fitting on the first set of moving data by using the least squares method to obtain first fitted data includes: Obtaining fitting parameters for performing polynomial fitting on the first set of moving data by using the least squares method; Performing polynomial fitting on the first set of moving data by using the fitting parameters for performing polynomial fitting on the first set of moving data to obtain first fitted data.

4. The method according to claim 1, wherein The performing polynomial fitting on the second set of moving data by using the least squares method to obtain second fitted data includes: Obtaining fitting parameters for performing polynomial fitting on the second set of moving data by using the least squares method; Performing polynomial fitting on the second set of moving data by using the fitting parameters for performing polynomial fitting on the second set of moving data to obtain second fitted data.

5. The method according to claim 1, characterized in that, The performing exponential smoothing on the set of fitted data to obtain trajectory smoothing data includes: Performing weighted average on the set of fitted data to obtain the trajectory smoothing data.

6. A map trajectory smoothing device, characterized in that, The device includes: A sampling module, configured to sample the original map data by using a preset distance threshold to obtain a set of sampled data, where the preset distance threshold is used to define the distance between adjacent two points; A fitting module, configured to perform polynomial fitting on the set of sampled data by using the least squares method according to the moving average idea to obtain a set of fitted data; A smoothing module, configured to perform exponential smoothing on the set of fitted data to obtain trajectory smoothing data; The fitting module includes: A length setting unit, configured to set the length of the smoothing interval; A first selection unit, configured to select a first set of moving data from the set of sampled data according to the length of the smoothing interval; A first fitting unit, configured to perform polynomial fitting on the first set of moving data by using the least squares method to obtain first fitted data; A second selection unit, configured to select a second moving data set from the sampling data set according to the smoothing interval length; A second fitting unit, configured to perform polynomial fitting on the second moving data set by using the least square method to obtain second fitting data; A result unit, configured to generate a fitting data set according to the first fitting data and the second fitting data.

7. A device, characterized in that, The device includes a memory and a processor, the memory is configured to store instructions or codes, and the processor is configured to execute the instructions or codes so that the device performs the map trajectory smoothing method according to any one of claims 1 to 5.

8. A computer storage medium, characterized in that, Codes are stored in the computer storage medium, and when the codes are run, the computer storage device running the codes implements the map trajectory smoothing method according to any one of claims 1 to 5.

Images