LATENT SPACE MODEL FOR ROAD NETWORKS TO PREDICT TIME-VARYING TRAFFIC

• US 20190180612A1
• Filed: 08/14/2017
• Published: 06/13/2019
• Est. Priority Date: 08/15/2016
• Status: Active Grant

• Current Assignee: University of Southern California
• Original Assignee: University of Southern California
• Inventors: Demiryurek, Ugur, Deng, Dingxiong, Shahabi, Cyrus, Zhu, Linhong, Yu, Rose, Liu, Yan
• Granted Patent: US 11,715,369 B2
• Time in Patent Office: Days
• CPC Class Codes: G01C 21/20 Instruments for performing navigational calculations G01C21/24, G01C21/26 take precedence
• CPC Class Codes: G01C 21/3691 Retrieval, searching and output of information related to real-time traffic, weather, or environmental conditions arrangements for giving variable traffic instructions G08G1/09
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G08G 1/0116 from roadside infrastructure, e.g. beacons
• CPC Class Codes: G08G 1/012 from other sources than vehicle or roadside beacons, e.g. mobile networks
• 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 1/0133 for classifying traffic situation
• CPC Class Codes: G08G 1/02 using treadles built into the road pads or other sensitive devices responsive to passage of vehicles E01F11/00
• CPC Class Codes: G08G 1/096833 where different aspects are considered when computing the route

Abstract

A method for traffic prediction of a road network includes receiving past traffic information corresponding to multiple locations on the road network. The method further includes determining, by a processor and based on the past traffic information, temporal characteristics of the past traffic information corresponding to changes of characteristics over time and spatial characteristics of in the past traffic information corresponding to interactions between locations on the road network. The method further includes predicting predicted traffic information corresponding to a later time based on the determined temporal and spatial characteristics of the past traffic information. The method further includes receiving detected additional traffic information corresponding to the later time. The method further includes updating the temporal characteristics of the traffic information and the spatial characteristics of the traffic information based on the predicted traffic information and the detected additional traffic information.

Claims

1. A method for traffic prediction of a road network comprising:

receiving, from a plurality of sensors, past traffic information corresponding to multiple locations on the road network;

determining, by a processor and based on the past traffic information, temporal characteristics of the past traffic information corresponding to changes of characteristics of at least some of the multiple locations on the road network over time and spatial characteristics of the past traffic information corresponding to interactions between at least some of the multiple locations on the road network;

predicting, by the processor, predicted traffic information corresponding to a later time based on the determined temporal characteristics of the past traffic information and the determined spatial characteristics of the past traffic information;

receiving, from the plurality of sensors, detected additional traffic information corresponding to the later time for at least some of the multiple locations on the road network; and

updating, by the processor, the temporal characteristics of the traffic information and the spatial characteristics of the traffic information based on the predicted traffic information and the detected additional traffic information.

2. The method of claim 1 wherein receiving the past traffic information further includes receiving at least one of a vehicle speed or a vehicle density for at least some of the multiple locations at multiple time intervals.

3. The method of claim 1 wherein each of the plurality of sensors includes a traffic loop sensor configured to detect a vehicle speed and a vehicle density over a particular segment of a road.

4. The method of claim 1 further comprising representing the road network via a plurality of vertices each corresponding to an intersection or an end of a road and a plurality of edges interconnecting each of the plurality of vertices and wherein the past or current traffic information includes a vehicle speed along an edge.

5. The method of claim 1 wherein updating the temporal characteristics of the traffic information and the spatial characteristics of the traffic information based on the predicted traffic information and the detected additional traffic information is performed using a combination of global learning and incremental learning.

6. The method of claim 1 further comprising predicting, by the processor, additional predicted traffic information corresponding to traffic information at locations other than locations associated with sensors based on the determined temporal characteristics of the past traffic information and the determined spatial characteristics of the past traffic information.

7. The method of claim 1 wherein the method can be used to predict traffic for periodic road network events and irregular road network events.

8. The method of claim 1 wherein the road network is modeled using a first matrix to represent the temporal characteristics of the past traffic information and a second matrix to represent the spatial characteristics of the past traffic information.

9. The method of claim 8 wherein predicting the predicted traffic information includes performing matrix operations using the first matrix and the second matrix.

10. The method of claim 8 wherein updating the temporal characteristics of the traffic information and the spatial characteristics of the traffic information includes performing matrix operations using the first matrix and the second matrix.

11. The method of claim 1 further comprising determining, by the processor, aggregated past traffic information by averaging the past traffic information over a predetermined time interval for each of the plurality of sensors, wherein determining the temporal characteristics and the spatial characteristics is based on the aggregated past traffic information.

12. The method of claim 1 further comprising:

receiving, by a navigation processor of a navigation server, a route request from a navigation interface provider including a starting location and a destination;

receiving, by the navigation processor, the predicted traffic information;

detetiiiining, by the navigation processor, a route from the starting location to the destination based on the predicted traffic information; and

transmitting, by the navigation processor, the route to the navigation interface provider.

13. The method of claim 12 further comprising outputting, by the navigation interface provider, navigation instructions corresponding to the route.

14. A system for predicting traffic of a road network, the system comprising:

a plurality of sensors positioned at multiple locations on the road network and configured to detect past traffic information;

a network access device configured to receive the past traffic information from the plurality of sensors; and

a prediction processor coupled to the network access device and configured to;

determine temporal characteristics of the past traffic information corresponding to changes of characteristics of at least some of the multiple locations on the road network over time and spatial characteristics of the past traffic information corresponding to interactions between at least some of the multiple locations on the road network,

predict predicted traffic information corresponding to a later time based on the determined temporal characteristics of the past traffic information and the determined spatial characteristics of the past traffic information,

receive detected additional traffic information corresponding to the later time for at least some of the multiple locations on the road network, and

update the temporal characteristics of the traffic information and the spatial characteristics of the traffic information based on the predicted traffic information and the detected additional traffic information.

15. The system of claim 14 wherein each of the plurality of sensors includes at least one of a traffic loop sensor, an optical sensor, a camera, a line of sight sensor, or a global positioning system (GPS) sensor and is configured to detect at least one of a volume of vehicles, an occupancy of the vehicles, or a speed of the vehicles.

16. The system of claim 14 wherein the prediction processor is further configured to update the temporal characteristics of the traffic information and the spatial characteristics of the traffic information based on the predicted traffic information and the detected additional traffic information by using a combination of global learning and incremental learning.

17. The system of claim 14 wherein the prediction processor is further configured to predict additional predicted traffic information corresponding to traffic information at locations other than locations associated with sensors based on the determined temporal characteristics of the past traffic information and the determined spatial characteristics of the past traffic information.

18. The system of claim 14 further comprising a navigation server having a navigation network access device configured to communicate with the prediction processor, and a navigation processor configured to:

receive a route request from a navigation interface provider including a starting location and a destination via the network access device;

receive the predicted traffic information from the prediction processor via the network access device;

determine a route from the starting location to the destination based on the predicted traffic information; and

transmit the route to the navigation interface provider.

19. The system of claim 18 further comprising the navigation interface provider having an input device configured to receive the starting location and the destination and an output device configured to output navigation instructions corresponding to the route.

20. A method for traffic prediction of a road network comprising:

receiving, from a plurality of sensors, past traffic information corresponding to multiple locations on the road network;

determining, by a processor and based on the past traffic information, temporal characteristics of the past traffic information corresponding to changes of characteristics of at least some of the multiple locations on the road network over time, and spatial characteristics of the past traffic information corresponding to interactions between at least some of the multiple locations on the road network;

predicting, by the processor, predicted traffic information corresponding to traffic information at locations other than those associated with sensors at a later time based on the determined temporal characteristics of the past traffic information and the determined spatial characteristics of the past traffic information by performing matrix operations;

receiving, from the plurality of sensors, detected additional traffic information corresponding to the later time for at least some of the multiple locations on the road network; and

updating, by the processor, the temporal characteristics of the traffic information and the spatial characteristics of the traffic information based on the predicted traffic information and the detected additional traffic information.