Segmentation of Track Data

Abstract

A method for segmenting trajectory data, particularly by a control device, is disclosed, in which trajectory data comprising a plurality of points is received, a curve smoothing and / or generalization algorithm, e.g. the Douglas Pucker algorithm, is applied to the trajectory data in order to identify substantially linear portions bounded by segmentation points, an algorithm for detecting curves, particularly in the form of start and end points of the curve, is applied to the trajectory data, at least one segmentation point identified in an interval between the start and end points of the curve is replaced by the start and end points, and the trajectory data is segmented at the identified segmentation points, the start and the end points. Further disclosed are a control device, a computer program, and a machine-readable storage medium.

Description

[Technical field]

[0001] The present invention relates to a method for segmenting trajectory data.Furthermore, the present invention relates to a control device, a computer program and a machine-readable storage medium. [Background technology]

[0002] A trajectory refers to a sequence of time-stamped positions recorded in a particular coordinate system. An example is a vehicle trajectory recorded in continuous global navigation satellite system (GNSS) based geographic coordinates. Trajectory segmentation aims to divide a given trajectory into parts or blocks that have consistent geometric properties, e.g. direction and straightness. Segmentation has been applied in various fields, e.g. creating planning maps using crowdsourced trajectory data, where trajectory segmentation can be scaled to any area size.

[0003] Typically, trajectory segmentation focuses on finding trajectory points or locations whose geometric properties are inconsistent compared to neighboring points, and these points are registered as segmentation points. Segmentation points form the boundaries of two segmentations or blocks. Methods for segmenting trajectories are known. For example, a trajectory can be divided or segmented into segments of equal length. Furthermore, the linearity of a given trajectory can be exploited to calculate the minimum number of segmentation points.

[0004] However, a problem with the known methods is that segmentation points are often determined within curves, resulting in linear trajectory parts being divided into an excessive number of segments. Summary of the Invention [Problem to be solved by the invention]

[0005] The problem underlying the present invention is to propose a method for subdividing trajectory data into segments, in which segmentation of the curve portions is avoided.

Means for Solving the Problem

[0006] This problem is solved by the respective subject matters of the independent claims. Advantageous embodiments of the invention are the subject matters of the respective dependent claims. According to one aspect of the present invention, there is provided a method for segmenting trajectory data, particularly by a control device. In one step, trajectory data including a plurality of points is received. The points may each be position points having a time stamp.

[0007] In order to identify a substantially straight portion delimited by segmentation points, a curve smoothing and / or generalization algorithm, such as the Douglas-Peucker algorithm, for example, is applied to the trajectory data. Thereby, the shape or geographical characteristics of the trajectory data are simplified, segmentation points are formed, and the trajectory data is divided into curve portions.

[0008] In a further step, an algorithm for detecting curves, particularly in the form of the start and end points of the curve, is applied to the trajectory data. By such an algorithm, the lengths of the curve portions are detected.

[0009] At least one segmentation point identified in the interval between the start and end points of the curve is replaced by the start and end points. Thereby, segmentation within the curve portion of the trajectory data is prevented. Thus, such segmentation points within the curve portion are deleted and replaced by the start point and the end point of the curve portion, respectively.

[0010] The trajectory data is segmented at the identified segmentation points, start points, and end points. By this method, a curve smoothing and / or generalization algorithm is combined with a curve or curve segment detection algorithm. At this time, in order to avoid the formation of segmentation points within the curve part and the straight line part of the track data, the results of the respective algorithms may be integrated. In particular, by considering the consistency of the route changes in the track data, the curve detection algorithm can prevent the curve from being segmented halfway.

[0011] Since the method according to the present invention is applied individually to each set of track data, a plurality of sets of track data can be processed simultaneously in parallel. Here, the set of track data may be a group of measurement data of a vehicle, a group of measurement data of a geographical area, a predefined number of measurement data, etc.

[0012] According to a further aspect of the present invention, a control device is provided, and the control device is configured to execute this method. The control device may be, for example, an in-vehicle control device, an out-of-vehicle control device, or an out-of-vehicle server unit such as a cloud system. The control device may preferably receive and process the track data. For this purpose, the control device may have an internal memory or an external memory for temporarily or permanently storing the track data and the results of the method.

[0013] Furthermore, the control device may have an integrated communication unit or an external communication unit for receiving the track data and transmitting the segmented track data as a possible result of the method.

[0014] Furthermore, according to one aspect of the present invention, a computer program is provided that includes instructions for causing a computer or a control device to execute the method according to the present invention when the computer program is executed by the computer or the control device. According to a further aspect of the present invention, a machine-readable storage medium storing the computer program according to the present invention is provided.

[0015] In one embodiment, the path direction of each point is specified by a curve detection algorithm, and subsequently, the difference in the path direction between two consecutive points is specified. Points of the trajectory data are assigned to a curve or a curve portion if the difference in the path direction between a point and the next point exceeds a predefined threshold. This enables a technically easy recognition of the curve portion. Points of the trajectory data on a straight portion are filtered as points with a relative path direction below the threshold.

[0016] According to a further embodiment, the path direction is specified in the form of a vector between a first point and a second point having a yaw angle following it. In the trajectory data of a vehicle, this corresponds to the traveling direction of the vehicle from the first point to the second point. The yaw angle may be specified with respect to a predefined orientation or reference direction.

[0017] According to a further example, the difference in the path direction is standardized by the distance between the first point and the second point. By this means, additional curve smoothing and comparison of the specified path directions of different points become possible.

[0018] According to a further embodiment, a point that does not exceed the threshold and is adjacent to a point that exceeds the predefined threshold is defined as the start point or the end point of the curve and used as a segmentation point for dividing the trajectory data into parts. In this way, the unique start point and the unique end point of the curve or the curve portion may be specified. If there is no difference or only a slight difference in the path direction between two adjacent points of the trajectory data, since these do not exceed the threshold, they are assigned to the straight portion by this method and are not assigned to the curve portion.

[0019] According to a further example, the trajectory data segmented at the specified segmentation points, start points, and end points is received for at least some segments and used to control the vehicle. At this time, the trajectory data may be received segment by segment and used for vehicle control. A plurality of segments may be received simultaneously to map, for example, a planned trajectory or a part of a planned trajectory.

Brief Description of the Drawings

[0020] Hereinafter, preferred embodiments of the present invention will be described in detail with highly simplified schematic diagrams.

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Modes for Carrying Out the Invention

[0021] FIG. 1 is a diagram of segmented track data according to the prior art. Here, the straight portion 2 and the curved portion 4 of the track data 1 are divided into different segments 6. Thereafter, the track data 1 can be further processed in the form of the segment 6 or provided to road users (not shown).

[0022] In order to divide the track data 1, it is necessary to detect a segmentation point 8 at which a boundary is drawn between two segments 6. As can be seen from FIG. 1, the segmentation of the track data 1 is performed with uniform elongated segments 6, and these segments 6 also have boundaries within the curved portion 4. When road users use such segmented track data 1, processing delays may occur. This may have an adverse effect on traffic safety in complex traffic situations such as curves. Furthermore, the straight portion 2 is unnecessarily divided into a large number of segments 6.

[0023] Figures 2 to 4 are schematic plan views showing a method according to an embodiment of the present invention. Here, in Figure 2, arrow P indicates the moving direction in which the track data 1 is specified and the track data 1 is processed by this method. This method can be executed by a control device (not shown), and this control device may be designed inside or outside the vehicle.

[0024] In the first step, track data 1 having a plurality of points 3 is received. Regarding this, Figures 2 to 4 show detailed views showing only one track. However, the track data 1 may be composed of a plurality of tracks and may have a plurality of overlapping points 3.

[0025] In order to identify substantially straight-line portions delimited by segmentation points 8, 9, a curve smoothing and / or generalization algorithm such as the Douglas-Peucker algorithm is applied to the track data 1. The corresponding result is shown in Figure 2. Exemplarily, two segmentation points 8, 9 are shown.

[0026] In the second step shown in Figure 3, an algorithm for detecting a curve or a curve portion 4, particularly in the form of a start point 10 and an end point 12 of the curve portion 4, is applied to the track data 1. By the algorithm for detecting the curve portion 4, the course direction of each point 3 is identified, and subsequently, the difference in the course direction between two consecutive points is identified. The point 3 of the track data 1 is assigned to a curve or a curve portion 4 when the difference in the course direction between the point 31 and the next point 32 exceeds a predefined threshold value. This will be described in detail in Figure 5.

[0027] In a further step of this method, at least one segmentation point 8, 9 identified in the section between the start point 10 and the end point 12 of the curve portion 4 is replaced by the start point 10 and the end point 12. At this time, the corresponding segmentation point 9 is deleted or at least not used as the segmentation point 9.

[0028] Subsequently, the trajectory data 1 is segmented into segments 6 at the identified segmentation points 8, start point 10, and end point 12. Figure 5 is a schematic plan view showing the detection algorithm for the curved portion 4. Here, at the first point 31 having the first time stamp t n the travel direction is specified in the form of the yaw angle yaw. The second point 32 has the second time stamp t n+1 .

[0029] The trajectory data 1 is specified, for example, by a vehicle 14 traveling along a route. At this time, different points 31, 32 are traveled at different times t and corresponding time stamps are assigned. In the illustrated embodiment, the yaw angle yaw is measured with respect to the horizontal axis x, which may be, for example, a line connecting east and west. At this time, the yaw angle yaw in the illustrated embodiment may be calculated by the arctangent of the quotient of the vertical offset Δy and the horizontal offset Δx with respect to the subsequent second point 32.

[0030] The yaw angle yaw is calculated for each point 3 having a subsequent point. In a further step, the relative path change Δyaw between each point 3 is calculated. This is done by forming the difference between the respective yaw angles yaw.

[0031] The relative path change Δyaw of each trajectory point 3 may further be accumulated with the values of the points 3 within a predefined distance parameter. The accumulated angle values of each trajectory point may further be smoothed with a predefined averaging width. By these two steps, the points 3 having a continuous path change may be emphasized and the points 3 of the trajectory data 1 having a relatively constant moving direction like the straight portion 2 may be suppressed. By emphasizing the points where the smoothed relative path direction or path change Δyaw exceeds a predetermined angle threshold, the curve of a predetermined trajectory may be detected.

Claims

1. A method for segmenting trajectory data (1) by a control device, comprising: - receiving trajectory data (1) including a plurality of points (3); - applying a curve smoothing and / or generalization algorithm to the trajectory data (1) to identify substantially linear portions (2) delimited by segmentation points (8, 9); - applying an algorithm to the trajectory data (1) to detect a curve (4) in the form of a start point (10) and an end point (12) of the curve (4); - at least one segmentation point (9) identified in the interval between the start point (10) and the end point (12) of the curve (4) is replaced by the start point (10) and the end point (12), and the trajectory data (1) is segmented at the identified segmentation point (8), the start point (10), and the end point (12); wherein in the method, the routing direction of each point (3) of the trajectory data (1) is identified by the algorithm for detecting the curve (4), and subsequently, the difference in the routing direction between two consecutive first points (31) and second points (32) is identified. A point (3) of the trajectory data (1) is assigned to a curve (4) if the difference in the routing direction between the first point (31) and the second point (32) exceeds a predefined threshold; points (3) that do not exceed the threshold and are adjacent to the points (3) of the trajectory data (1) that exceed the predefined threshold are defined as the start point (10) and the end point (12) of the curve (4), and are used as segmentation points (8) for dividing the trajectory data (1) into segments (6); a method.

2. The method according to claim 1, wherein the curve smoothing and / or generalization algorithm is a Douglas-Peucker algorithm.

3. The method according to claim 1, wherein the routing direction is identified in the form of a vector between a first point (31) and a second point (32) having a subsequent yaw angle.

4. The method according to claim 1, wherein the difference in the routing direction is normalized by the distance between the first point (31) and the second point (32).

5. The method according to claim 1, wherein the trajectory data (1) segmented at the identified segmentation point (8), the start point (10), and the end point (12) is received for at least some segments and used to control a vehicle (14).

6. A control device configured to execute the method according to claim 1.

7. A computer program including instructions for causing a computer or a control device to execute the method according to claim 1 when the computer program is executed by the computer or the control device.

8. A machine-readable storage medium storing the computer program according to claim 7.

Images