Muck truck track prediction method

Abstract

The invention provides a slag car track prediction method, which comprises the steps of S1, acquiring historical positioning data of a slag car, and displaying a historical driving track of the slag car by the historical positioning data; s2, performing clustering analysis on the track points on the historical driving track to obtain a track cluster center; and S3, connecting each track cluster center and generating a muck truck prediction track. Therefore, the random points in the driving track of the muck truck can be processed, and the track of the muck truck can be accurately predicted. The invention also provides another muck truck track prediction method, which comprises the following steps: S1, acquiring historical positioning data of the muck truck, the historical positioning data displaying the historical driving track of the muck truck; s2, the historical driving track of the muck truck is segmented, the feature vector of each track segment is obtained, and the feature vectors comprise the driving distance and the driving direction of the muck truck; s3, performing clustering analysis on each track segment according to the feature vectors to obtain a central track segment; and S4, connecting the central track sections and generating a predicted track of the muck truck.

Description

Technical Field

[0001] The present invention relates to the field of construction waste transport vehicles, and particularly to a method for predicting the trajectory of a muck truck. Background Art

[0002] A muck truck is a transport vehicle for hauling earthwork, sand and gravel, lime, asphalt, cement and other building materials and construction waste in urban construction. The muck truck shuttles back and forth between construction sites and disposal sites, or between building material markets and construction sites. Its driving trajectory has certain rules, but also has a certain degree of randomness. The muck truck cannot pass through the same point repeatedly, but there will be some similar random points near that point. Currently, there is a lack of a method for processing these random points and predicting the trajectory of the muck truck based on them. Summary of the Invention

[0003] In view of this, the present invention provides a method for predicting the trajectory of a muck truck that can process the random points in the driving trajectory of the muck truck and can accurately predict the trajectory of the muck truck based on this.

[0004] The above-mentioned method for predicting the trajectory of a muck truck according to the present invention includes the following steps:

[0005] S1: Obtain the historical positioning data of the muck truck, where the historical positioning data shows the historical driving trajectory of the muck truck;

[0006] S2: Perform cluster analysis on the trajectory points on the historical driving trajectory to obtain the center of the trajectory cluster;

[0007] S3: Connect the centers of each trajectory cluster to generate the predicted trajectory of the muck truck.

[0008] Further, the trajectory points on the historical driving trajectory for cluster analysis are the trajectory points in the data collection points.

[0009] Further, perform cluster analysis on each data collection point on each historical driving trajectory.

[0010] Further, perform cluster analysis on the data collection points on each historical driving trajectory in a manner of spacing 1, 2,... or s data collection points.

[0011] Further, in S2, draw a circle with each similar point on each trajectory point as the center, and use the center of the circle that includes all similar points as the center of the trajectory cluster.

[0012] Further, in S2, draw a circle with each similar point on each trajectory point as the center, and use the center of the circle that includes the most similar points as the center of the trajectory cluster.

[0013] Further, before S2, perform similarity measurement on the trajectory points on the historical driving trajectory and remove the trajectory points with low similarity.

[0014] Furthermore, similarity measurement is performed in a human - observation manner, and trajectory points with low similarity are removed.

[0015] Furthermore, similarity measurement is performed as follows: Obtain the geographical coordinates of each similar point on each trajectory point, and identify the trajectory points with large deviations in geographical coordinate values as trajectory points with low similarity.

[0016] Furthermore, similarity measurement is performed as follows:

[0017] Suppose there are n trajectory points on the historical driving trajectory of the first muck truck. For the h - th trajectory point, calculate the distance L1h from it to the first trajectory point and the distance Lhn from it to the n - th trajectory point, and then calculate the sum Lh = L1h+Lhn;

[0018] Suppose there are n trajectory points on the historical driving trajectory of the second muck truck. For the h - th trajectory point, calculate the distance L1h from it to the first trajectory point and the distance Lhn from it to the n - th trajectory point, and then calculate the sum Lh' = L1h'+Lhn';

[0019] And so on...

[0020] Suppose there are n trajectory points on the historical driving trajectory of the i - th muck truck. For the h - th trajectory point, calculate the distance L1h'' from it to the first trajectory point and the distance Lhn'' from it to the n - th trajectory point, and then calculate the sum Lh'' = L1h''+Lhn'';

[0021] Compare the magnitudes of Lh, Lh',...Lh'', and identify the trajectory points with large deviations in L values as trajectory points with low similarity.

[0022] Furthermore, the historical positioning data also shows the driving speed of the muck truck. When the driving speed of the muck truck is greater than a certain threshold, the trajectory points on the historical driving trajectory are removed.

[0023] Furthermore, the historical positioning data is obtained through at least one of the following three methods:

[0024] ① Obtained by the GPS global positioning system;

[0025] ② Obtained by a GPS signal receiving device. Among them, the on - vehicle GPS device sends location information to the GPS signal receiving device, and the GPS signal receiving device analyzes and obtains the geographical coordinates of the muck truck and returns the geographical coordinates;

[0026] ③ Obtained by a geographical coordinate sending device. Among them, the geographical coordinate sending device sends geographical coordinate signals, and the on - vehicle GPS device receives the geographical coordinate signals and obtains the geographical coordinates.

[0027] The trajectory prediction method for muck trucks of the present invention can process the random points in the driving trajectory of muck trucks and accurately predict the muck truck trajectory based on this.

[0028] Furthermore, the present invention also provides another trajectory prediction method for muck trucks, including the following steps:

[0029] S1: Obtain the historical positioning data of the muck truck, where the historical positioning data shows the historical driving trajectory of the muck truck;

[0030] S2: Segment the historical driving trajectory of the muck truck and obtain the feature vectors of each trajectory segment, where the feature vectors include the driving distance and driving direction of the muck truck;

[0031] S3: Perform clustering analysis on each trajectory segment according to the feature vectors to obtain the central trajectory segments;

[0032] S4: Connect each central trajectory segment and generate the predicted trajectory of the muck truck.

[0033] Furthermore, the time period for segmenting is determined based on the time period for data collection.

[0034] Furthermore, the time period for segmenting is the same as the time period for data collection.

[0035] Furthermore, the time period for segmenting is 2 times, 3 times... or s times the time period for data collection.

[0036] Furthermore, before S3, similarity measurement is performed by manually observing the driving distance and driving direction of the muck truck on each trajectory segment, and the trajectory segments with low similarity are removed.

[0037] This trajectory prediction method for muck trucks can process the similar trajectory segments in the driving trajectory of muck trucks and accurately predict the muck truck trajectory based on this. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a schematic flowchart of the trajectory prediction method for muck trucks according to an embodiment of the present invention.

[0039] Figure 2 is an example schematic diagram of the trajectory points on the historical driving trajectory for clustering analysis being the trajectory points between data collection points.

[0040] Figure 3 is an example schematic diagram of performing clustering analysis on each data collection point on each historical driving trajectory.

[0041] Figure 4It is an example schematic diagram for performing clustering analysis on data acquisition points on each historical driving trajectory at an interval of one data acquisition point.

[0042] Figure 5 It is an example schematic diagram for performing clustering analysis on data acquisition points on each historical driving trajectory at an interval of two data acquisition points.

[0043] Figure 6 It is a schematic flowchart of the trajectory prediction method for a muck truck according to another embodiment of the present invention.

[0044] Figure 7 It is an example schematic diagram where the time period for segmented processing is different from the time period of data acquisition points.

[0045] Figure 8 It is an example schematic diagram where the time period for segmented processing is the same as the time period of data acquisition.

[0046] Figure 9 It is an example schematic diagram where the time period for segmented processing is twice the time period of data acquisition. Detailed implementation manners

[0047] The following combines the accompanying drawings and embodiments to further describe in detail the specific implementation manners of the present invention. The following embodiments are used to illustrate the present invention, but are not used to limit the scope of the present invention.

[0048] As Figure 1 shown, the trajectory prediction method for a muck truck according to an embodiment of the present invention includes the following steps:

[0049] S1: Obtain the historical positioning data of the muck truck, where the historical positioning data shows the historical driving trajectory of the muck truck;

[0050] S2: Perform clustering analysis on the trajectory points on the historical driving trajectory and obtain the trajectory cluster center;

[0051] S3: Connect the centers of each trajectory cluster to generate a vehicle prediction trajectory.

[0052] The muck truck transports back and forth between the construction site and the disposal site, or between the building materials market and the construction site. The GPS trajectory has certain rules but also certain randomness, so that the historical positioning data will show several (for example, i) similar historical driving trajectories of the muck truck. There will be i similar trajectory points at each location on the historical driving trajectory. By performing clustering analysis on the i similar trajectory points at each location and obtaining the trajectory cluster center, and connecting the centers of each trajectory cluster, the muck truck prediction trajectory can be generated. Therefore, the trajectory prediction method for the muck truck of the present invention can process the random points in the muck truck driving trajectory and can accurately predict the muck truck trajectory based on this.

[0053] There are various clustering analysis methods, and the present invention provides two embodiments.

[0054] The first method is to draw a circle with each similar point on each trajectory point as the center, and take the center of the circle that includes all similar points as the trajectory cluster center. The radius of the circle is determined according to the distribution of similar points. It should not be too small, otherwise clustering cannot be performed; nor should it be too large, otherwise the circles drawn with each similar point as the center can all include all similar points. This method requires a very high similarity of each similar point to determine the trajectory cluster center, so that the circle corresponding to the trajectory cluster center can include all similar points.

[0055] The second method is to draw a circle with each similar point on each trajectory point as the center, and take the center of the circle that includes the most similar points as the trajectory cluster center. The radius of the circle is determined according to the distribution of similar points. It should not be too small, otherwise clustering cannot be performed; nor should it be too large, otherwise the circles drawn with each similar point as the center may all include all similar points, making it impossible to determine which circle includes the most similar points. This method does not require a very high similarity of each similar point to determine the trajectory cluster center and has stronger versatility.

[0056] The trajectory points on the historical driving trajectory for clustering analysis can be the trajectory points between data collection points. For example, as Figure 2 shown, the positioning device collected the trajectory point s1 at time t1 and the trajectory point s2 at the next time t2. The trajectory point for clustering analysis is the trajectory point s3 between the trajectory point s1 and the trajectory point s2. However, the trajectory points for clustering analysis are preferably the trajectory points among the data collection points, so as to more accurately predict the trajectory of the muck truck.

[0057] There are two ways for the trajectory points for clustering analysis to be the trajectory points among the data collection points.

[0058] One way is to perform clustering analysis on each data collection point on each historical driving trajectory. As Figure 3 shown, the positioning device collected the trajectory point s1 at time t1, the trajectory point s2 at the next time t2, and the trajectory point s3 at the next time t3. At this time, clustering analysis is performed on the trajectory points s1, s2, and s3. This processing method can obtain a more refined predicted trajectory of the muck truck.

[0059] Another way is to perform clustering analysis on the data collection points on each historical driving trajectory at intervals of 1, 2,... or s data collection points. For example, as Figure 4 shown, performing clustering analysis on the data collection points on each historical driving trajectory at intervals of 1 data collection point means performing clustering analysis on the 1st, 3rd, 5th,... data collection points. As Figure 5As shown, by clustering the data acquisition points on each historical driving trajectory at an interval of two data acquisition points, it is to perform clustering analysis on the 1st, 4th, 7th... data acquisition points. And so on. This processing method can obtain the predicted trajectory of the muck truck more quickly.

[0060] Further, before S2, similarity measurement is also performed on the trajectory points on the historical driving trajectory, and the trajectory points with low similarity are removed. The trajectory points on the historical driving trajectory have a certain degree of randomness. Especially in bad weather such as at night, in fog, or in rain, the randomness of the trajectory points will be stronger, and there will be some trajectory points with low similarity. By performing similarity measurement on the trajectory points on the historical driving trajectory and removing the trajectory points with low similarity, the trajectory points on the historical driving trajectory can be better clustered.

[0061] There are three methods for similarity measurement.

[0062] The first is to perform similarity measurement in a way of human observation, that is, to observe the trajectory points on the historical driving trajectory through human eyes and remove the trajectory points with large deviations. This method is more intuitive and direct.

[0063] The second is to obtain the geographical coordinates of each similar point on each trajectory point, and identify the trajectory points with large deviations in geographical coordinate values as the trajectory points with low similarity. This method does not require manual operation and can perform similarity measurement automatically by machine.

[0064] The third is to assume that there are n trajectory points on the historical driving trajectory of the first muck truck. For the hth trajectory point, calculate its distance L 1h to the first trajectory point and its distance L hn to the nth trajectory point, and then calculate the sum L 1h of L hn and L h =L 1h +L hn ;

[0065] Assume that there are n trajectory points on the historical driving trajectory of the second muck truck. For the hth trajectory point, calculate its distance L 1h ’ to the first trajectory point and its distance L hn ’ to the nth trajectory point, and then calculate the sum L 1h ’ of L hn ’ and L h ’=L 1h ’+L hn ’;

[0066] And so on...

[0067] Suppose there are n trajectory points on the historical driving trajectory of the i-th muck truck. For the h-th trajectory point, calculate the distance L from it to the first trajectory point 1h ” and the distance L from it to the n-th trajectory point hn ”. Then calculate the sum L 1h ” of L hn ” and L h ” = L 1h ” + L hn ”;

[0068] Compare the magnitudes of L h , L h ’... L h ”, and identify the trajectory points with large deviations in L values as trajectory points with low similarity.

[0069] This method also does not require manual operation, can automatically perform similarity measurement by machine, and has higher accuracy.

[0070] Furthermore, the historical positioning data also shows the driving speed of the muck truck. When the driving speed of the muck truck is greater than a certain threshold, the trajectory points on the historical driving trajectory are removed. The threshold can be determined according to the speed limit of this section of the road. For example, the threshold can be the speed limit of this section of the road, or a certain value larger than the speed limit of this section of the road. When the driving speed of the muck truck is greater than a certain threshold, it indicates that the GPS signal is inaccurate, resulting in the driving speed corresponding to the point located by GPS being greater than the threshold. For example, when the muck truck passes through a tunnel, the signal in the tunnel is poor, and GPS may locate the muck truck outside the tunnel. At this time, the driving speed corresponding to the point located by GPS will be greater than the threshold, and in this case, the GPS located point needs to be removed.

[0071] The historical positioning data is obtained through at least one of the following three methods:

[0072] ① Obtained by the GPS global positioning system;

[0073] ② Obtained by a GPS signal receiving device, where the on-vehicle GPS device sends location information to the GPS signal receiving device, and the GPS signal receiving device analyzes and obtains the geographical coordinates of the muck truck and returns the geographical coordinates;

[0074] ③ Obtained by a geographical coordinate sending device, where the geographical coordinate sending device sends geographical coordinate signals, and the on-vehicle GPS device receives the geographical coordinate signals.

[0075] The historical positioning data is preferably obtained through the GPS global positioning system. GPS can obtain positioning data globally. However, in some sections, such as inside tunnels, the GPS signal may be poor, unable to locate the vehicle, or locate the vehicle in other positions. At this time, the positioning data can be obtained through the GPS signal receiving device or the geographical coordinate sending device.

[0076] As Figure 6 shown, the present invention also provides another method for predicting the trajectory of a muck truck, including the following steps:

[0077] S1: Obtain the historical positioning data of the muck truck, where the historical positioning data shows the historical driving trajectory of the muck truck;

[0078] S2: Segment the historical driving trajectory of the muck truck and obtain the feature vector of each trajectory segment, where the feature vector includes the driving distance and driving direction of the muck truck;

[0079] S3: Perform clustering analysis on each trajectory segment according to the feature vector to obtain the central trajectory segment;

[0080] S4: Connect each central trajectory segment to generate the predicted trajectory of the muck truck.

[0081] The difference between this method for predicting the trajectory of a muck truck and the above-mentioned method for predicting the trajectory of a muck truck is that the above-mentioned method for predicting the trajectory of a muck truck performs clustering analysis on each similarity point and obtains the center of the trajectory cluster, while this method for predicting the trajectory of a muck truck performs clustering analysis on each trajectory segment and obtains the central trajectory segment, and connecting each central trajectory segment can generate the predicted trajectory of the muck truck. Therefore, this method for predicting the trajectory of a muck truck can process the similar trajectory segments in the driving trajectory of the muck truck and can accurately predict the trajectory of the muck truck based on this.

[0082] The time period for segmentation processing can be different from the time period of the data collection points. For example, as Figure 7 shown, the positioning device collects the trajectory point s1 at time t1, collects the trajectory point s2 at the next time t2, and then collects the trajectory point s3 at the next time t3. The starting point of the time period for segmentation processing is t' between t1 and t2, and the ending point of the time period for segmentation processing is t" between t2 and t3. However, it is preferably determined based on the time period of data collection, so that the trajectory of the muck truck can be predicted more accurately.

[0083] There are two ways to determine the time period for segmentation processing based on the time period of data collection.

[0084] One is that the time period for segmentation processing is the same as the time period of data collection. As Figure 8 shown, the positioning device collects the trajectory point s1 at time t1, collects the trajectory point s2 at the next time t2, and then collects the trajectory point s3 at the next time t3. The time period for segmentation processing is t1t2 and t2t3. This processing method can obtain a more refined predicted trajectory of the muck truck.

[0085] The other is that the time period for segmentation processing is 2 times, 3 times... or s times the time period of data collection. For example, asFigure 9 As shown, the positioning device collected the trajectory point s1 at time t1, the trajectory point s2 at the next time t2, the trajectory point s3 at the next time t3, the trajectory point s4 at the next time t4, and the trajectory point s5 at the next time t5. The time period for segmented processing is twice the time period of data collection, and clustering analysis is performed on the time periods t1-t3 and t3-t5. And so on. This processing method can obtain the predicted trajectory of the muck truck more quickly.

[0086] Further, before S3, similarity measurement is also performed by manually observing the driving distance and driving direction of the muck truck for each trajectory segment, and the trajectory segments with low similarity are removed. That is, by observing the driving distance and driving direction of the muck truck for each trajectory segment with the human eye, and removing the trajectory segments with large deviations in driving distance and driving direction. By performing similarity measurement on each trajectory segment and removing the trajectory segments with low similarity, the clustering analysis of the trajectory segments on the historical driving trajectory can be better performed.

[0087] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0088] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for predicting the trajectory of a muck truck, characterized in that: The steps include: S1: Obtain historical positioning data of the muck truck, wherein the historical positioning data shows the historical driving trajectory of the muck truck; S2: cluster analysis of trajectory points on the historical driving trajectory to obtain the trajectory cluster center; S3: Connect the centers of each trajectory cluster and generate the predicted trajectory of the muck truck.

2. The method for predicting the trajectory of a muck truck according to claim 1, characterized in that: The trajectory points on the historical driving trajectory of cluster analysis are the trajectory points in the data collection points.

3. The method for predicting the trajectory of a muck truck according to claim 2, characterized in that: Cluster analysis is performed on each data collection point on each historical driving trajectory.

4. The method for predicting the trajectory of a muck truck according to claim 2, characterized in that: Cluster analysis is performed on the data collection points on each historical driving trajectory in a manner of intervals of 1, 2... or s data collection points.

5. The method for predicting the trajectory of a muck truck according to claim 1, characterized in that: In S2, a circle is drawn with each similar point on each trajectory point as the center, and the center of the circle including all similar points is the trajectory cluster center.

6. The method for predicting the trajectory of a muck truck according to claim 1, characterized in that: In S2, a circle is drawn with each similar point on each trajectory point as the center, and the center of the circle including the most similar points is the center of the trajectory cluster.

7. The method for predicting the trajectory of a muck truck according to claim 1, characterized in that: Before S2, the similarity of the trajectory points on the historical driving trajectory is measured, and the trajectory points with low similarity are removed.

8. The method for predicting the trajectory of a muck truck according to claim 7, characterized in that: The similarity is measured in a human-observation manner, and trajectory points with low similarity are removed.

9. The method for predicting the trajectory of a muck truck according to claim 7, characterized in that: The similarity measurement is performed in the following manner: the geographic coordinates of each similar point on each trajectory point are obtained, and trajectory points with large deviations in geographic coordinate values ​​are identified as trajectory points with low similarity.

10. The method for predicting the trajectory of a muck truck according to claim 7, characterized in that: The similarity is measured in the following way: Assume that there are n track points on the historical driving track of the first muck truck. For the hth track point, calculate its distance L to the first track point. 1h and its distance to the nth trajectory point L hn , and then calculate L 1h and L hn and L h =L 1h +L hn ; Assume that there are n track points on the historical driving track of the second muck truck. For the hth track point, calculate its distance L to the first track point. 1h ' and its distance to the nth trajectory point L hn ', then calculate L 1h ' and L hn ' and L h '=L 1h '+L hn '; So recursion... Assume that there are n track points on the historical driving track of the i-th muck truck. For the h-th track point, calculate its distance L to the first track point. 1h " and its distance to the nth trajectory point L hn ", and then calculate L 1h ” and L hn " and L h ”=L 1h ”+L hn ”; Compare L h , L h '...L h " and identify trajectory points with large L value deviation as trajectory points with low similarity.

11. The method for predicting the trajectory of a muck truck according to claim 1, characterized in that: The historical positioning data also shows the driving speed of the muck truck. When the driving speed of the muck truck is greater than a certain threshold, the track points on the historical driving track are removed.

12. The method for predicting the trajectory of a muck truck according to claim 1, characterized in that: The historical positioning data is obtained in at least one of the following three ways: ① Obtained by GPS global positioning system; ② Obtained by a GPS signal receiving device, wherein the vehicle-mounted GPS device sends location information to the GPS signal receiving device, and the GPS signal receiving device analyzes and obtains the geographic coordinates of the muck truck and returns the geographic coordinates; ③ Obtained by a geographic coordinate sending device, wherein the geographic coordinate sending device sends a geographic coordinate signal, the vehicle-mounted GPS device receives the geographic coordinate signal, and obtains the geographic coordinates.

13. A method for predicting the trajectory of a muck truck, characterized in that: The steps include: S1: Obtain historical positioning data of the muck truck, wherein the historical positioning data shows the historical driving trajectory of the muck truck; S2: Segment the historical driving trajectory of the muck truck and obtain the feature vector of each trajectory segment, wherein the feature vector includes the driving distance and driving direction of the muck truck; S3: performing cluster analysis on each trajectory segment according to the feature vector to obtain a central trajectory segment; S4: Connect each center trajectory segment and generate a predicted trajectory of the muck truck.

14. The method for predicting the trajectory of a muck truck according to claim 13, characterized in that: The time period for segmented processing is determined based on the time period for data collection.

15. The method for predicting the trajectory of a muck truck according to claim 14, characterized in that: The time period for segmentation processing is consistent with the time period for data collection.

16. The method for predicting the trajectory of a muck truck according to claim 14, characterized in that: The time period of segmented processing is 2 times, 3 times... or s times the time period of data acquisition.

17. The method for predicting the trajectory of a muck truck according to claim 13, characterized in that: Before S3, similarity was measured by manually observing the driving distance and driving direction of the muck truck in each trajectory segment, and trajectory segments with low similarity were removed.

Images