Identification of vehicle parking using data from vehicle sensor network

• US 9,607,509 B2
• Filed: 04/08/2015
• Issued: 03/28/2017
• Est. Priority Date: 04/08/2015
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP SE
• Inventors: Hu, Junpeng, Li, Yongxun, Chen, Siyi, Liu, Jing, Ni, Yinling, Liu, Jun
• Primary Examiner(s): Sample, Jonathan L
• Application Number:
  US14/681,122
• Publication Number:
  US 20160300486A1
• Time in Patent Office: 720 Days
• US Class Current: 1/1
• CPC Class Codes: G01S 1/026 of associated receivers
• CPC Class Codes: G08G 1/0116 from roadside infrastructure, e.g. beacons
• CPC Class Codes: G08G 1/0125 Traffic data processing
• CPC Class Codes: G08G 1/0129 for creating historical data or processing based on historical data
• CPC Class Codes: G08G 1/017 identifying vehicles G08G1/015, G08G1/054 take precedence
• CPC Class Codes: G08G 1/0175 by photographing vehicles, e.g. when violating traffic rules
• CPC Class Codes: G08G 1/144 on portable or mobile units, e.g. personal digital assistant [PDA]
• CPC Class Codes: G08G 1/147 where the parking area is within an open public zone, e.g. city centre

Abstract

A framework for identifying parking areas from vehicle trajectory data is described herein. Vehicle trajectory data collected from a sensor network having a plurality of sensor stations for detecting vehicles is provided. The vehicle trajectory data is pre-processed to generate a tracking table of vehicles and analyzed to determine an interstation parking area between first and second sensor stations of interest based on input parameters from a user.

Claims

1. A computer-implemented method for identifying parking areas comprising:

providing vehicle trajectory data of vehicles collected from a sensor network having a plurality of sensor stations for detecting vehicles, the sensor stations of the sensor network are distributed in a geographical area of interest, the vehicle trajectory data is stored as vehicle trajectory data records in a vehicle trajectory table, wherein a vehicle trajectory data record of the vehicle trajectory data comprises,

a vehicle id identifying the vehicle detected,

a sensor station id identifying the sensor station detecting the vehicle, and

a captured time which is the time that the sensor station detected the vehicle;

pre-processing the vehicle trajectory data records by an analysis system to generate a tracking table of vehicles;

storing the tracking table in a data model of an analyzer of the analysis system;

analyzing the data model to determine an interstation parking area between first and second sensor stations of interest identified based on input parameters from a user, wherein the interstation parking area represents a probable stopping region of a vehicle of interest and is determined by calculating, from the data model, transition probabilities from the first sensor station to the second sensor station via one or more intermediate sensor stations; and

displaying the interstation parking area at a user interface.

2. The computer-implemented method of claim 1 wherein pre-processing the vehicle trajectory data comprises discarding dirty data records of the vehicle trajectory data, wherein a dirty data record is a vehicle trajectory record having a difference in capture time with a capture time of a previous vehicle trajectory data record with the same vehicle id is greater than a threshold value.

3. The computer-implemented method of claim 2 wherein the tracking table comprises a plurality of tracking tables based on vehicle id.

4. The computer-implemented method of claim 1 wherein pre-processing the vehicle trajectory data comprises:

discarding dirty data records, wherein a dirty data record is a vehicle trajectory record having a difference in capture time with a capture time of a previous vehicle trajectory data record with the same vehicle id is greater than a threshold value; and

wherein the tracking table comprises a plurality of tracking tables based on vehicle id.

5. The computer-implemented method of claim 1 wherein a process of generating the tracking table comprises:

caching a current vehicle trajectory data record received;

searching the vehicle trajectory table to determine if a previous vehicle trajectory data record with the same vehicle id as that of the current vehicle trajectory data record exist;

if the previous vehicle trajectory data record does not exist, terminate the process;

if the previous vehicle trajectory data record does exist, calculate the time difference between a current_captured time of the current vehicle trajectory data record with a last_captured time of the previous vehicle trajectory data record;

if the difference between the current_captured time and the last_captured time is not less than a threshold value, the process terminates; and

if the difference between the current_captured time and the last captured time is less than a threshold value,

the current vehicle trajectory data record is processed to form a tracking table record,

inserting the tracking table record into the tracking table, and

the process terminates.

6. The computer-implemented method of claim 5 wherein the tracking table comprises a plurality of tracking tables based on vehicle id.

7. The computer-implemented method of claim 5 wherein a tracking record in the tracking table comprises:

the vehicle id identifying the vehicle detected;

a previous sensor station id of the previous vehicle trajectory data record;

the previous_captured time of the previous vehicle trajectory data record;

the current_sensor station id of the current vehicle trajectory data record; and

the current_captured time of the current vehicle trajectory data record.

8. The computer-implemented method of claim 7 wherein the tracking table comprises a plurality of tracking tables based on vehicle id.

9. The computer-implemented method of claim 1 wherein a tracking record in the tracking table comprises:

the vehicle id identifying the vehicle detected;

a from_sensor station id identifying a from sensor station detecting the vehicle;

a from_captured time which is the time that the from sensor station detected the vehicle;

a to_sensor station id identifying a to sensor station detecting the vehicle; and

a to_captured time which is the time that the to sensor station detected the vehicle.

10. The computer-implemented method of claim 9 wherein tracking records in the tracking table comprises a difference between the from_capture time and the to_capture time is less than a threshold value.

11. The computer-implemented method of claim 10 wherein the tracking table comprises a plurality of tracking tables based on vehicle id.

12. The computer-implemented method of claim 1 wherein the sensor stations comprise fixed traffic sensors, wherein a fixed sensor station comprises a radio-frequency identification (RFID) sensor, a license plate recognition (LPR) sensor or both.

13. The computer-implemented method of claim 1 wherein identifying parking areas further comprises identifying interstation parking areas between first and second contiguous sensor stations of a set of stations.

14. The computer-implemented method of claim 13 comprises determining a convex hull surrounding the first and second contiguous sensor stations, wherein the parking areas are disposed within the convex hull.

15. The computer-implemented method of claim 13 wherein identifying the interstation parking areas between the first and second contiguous sensor stations of a set of stations comprises:

calculating probabilities Pr(b_i|S₁,S₂) for b_iε

A₂, where i=1, . . . N−

2;

sorting the probabilities Pr(b_i|S₁,S₂); and

selecting first K stations based on

K=max{θ

₁,argmin_K[Σ

_k=0^KPr(b_iN-k-2|S1,S2)≧

θ

₂]}.

16. The computer-implemented method of claim 15 comprises determining a convex hull surrounding the first and second contiguous sensor stations in which the parking areas are disposed, where determining the convex hull comprises determining a set of points by {S₁,S₂}∪

B_K, where B_K={b<sub>iN−

2</sub>, . . . ,b<sub>n−

K−

2</sub>}, wherein the set of points is used to generate the convex hull to form a convex polygon enclosing the set of points.

17. A parking area analysis system comprising:

a data storage module for storing a vehicle trajectory table, the vehicle trajectory table includes vehicle trajectory data records collected from a sensor network having a plurality of sensor stations for detecting vehicles, the sensor stations of the sensor network are distributed in a geographical area of interest, wherein a vehicle trajectory data record comprises,

a vehicle id identifying the vehicle detected,

a sensor station id identifying the sensor station detecting the vehicle, and

a captured time which is the time that the sensor station detected the vehicle;

a pre-processor for analyzing the vehicle trajectory data set to generate a tracking table of vehicles, wherein the tracking table is stored in a data model;

an analyzer for analyzing the data model to determine an interstation parking area between first and second sensor stations of interest identified based on input parameters from a user, wherein the interstation parking area is disposed within a convex hull surrounding the first and second sensor stations, wherein the interstation parking area represents a probable stopping region of a vehicle of interest and is determined by calculating, from the data model, transition probabilities from the first sensor station to the second sensor station via one or more intermediate sensor stations; and

a display device that displays the interstation parking area.

18. The system of claim 17 wherein a tracking record in the tracking table comprises:

the vehicle id identifying the vehicle detected;

a from_sensor station id identifying a from sensor station detecting the vehicle;

a from_captured time which is the time that the from sensor station detected the vehicle;

a to_sensor station id identifying a to sensor station detecting the vehicle; and

a to_captured time which is the time that the to sensor detected the vehicle.

19. A non-transitory computer-readable medium having stored thereon a program code, the program code executable by a computer for identifying parking areas comprising:

providing vehicle trajectory data of vehicles collected from a sensor network having a plurality of sensor stations for detecting vehicles, the sensor stations of the sensor network are distributed in a geographical area of interest, the vehicle trajectory data is stored as vehicle trajectory data records in a vehicle trajectory table, wherein a vehicle trajectory data record of the vehicle trajectory data comprises,

a vehicle id identifying the vehicle detected,

a sensor station id identifying the sensor station detecting the vehicle, and

a captured time which is the time that the sensor station detected the vehicle;

pre-processing the vehicle trajectory data records to generate a tracking table of vehicles;

storing the tracking table in a data model of an analyzer of the analysis system;

analyzing the data model to determine an interstation parking area between first and second sensor stations of interest identified based on input parameters from a user, wherein the interstation parking area is disposed within a convex hull surrounding the first and second sensor stations, wherein the interstation parking area represents a probable stopping region of a vehicle of interest and is determined by calculating, from the data model, transition probabilities from the first sensor station to the second sensor station via one or more intermediate sensor stations; and

displaying the interstation parking area at a user interface.

20. The non-transitory computer-readable medium of claim 19 wherein the convex hull is determined by:

calculating probabilities Pr(b_i|S₁,S₂) for b_iε

A₂, where i=1, . . . N−

2;

sorting the probabilities Pr(b_i|S₁,S₂); and

selecting first K stations based on

K=max{θ

₁,argmin_K[Σ

_k=0^KPr(b<sub>iN−

k−

2</sub>|S1,S2)≧

θ

₂]}.