System and method for traffic signal timing estimation

1. A system for estimating traffic signal timing, comprising:

a computer processor; and

a non-transitory computer readable medium containing instructions directing the system to execute steps comprising:

extracting one or more trajectories of a vehicle from probe data;

matching the one or more trajectories to a map, wherein the map comprises one or more intersections and the one or more trajectories passes through the one or more intersections;

determining one or more traffic signal cycles of a traffic light at the one or more intersections; and

computing a signal timing estimation for the traffic light, based on an optimization of a score function,

wherein the score function is a function of (1) an offset parameter and (2) the one or more traffic signal cycles for the traffic light at the one or more intersections,

wherein the score function is a product of terms, each term being a sum of exponentials, and

wherein the offset parameter accounts for an unknown starting time of a cycle of the traffic light at the one or more intersections.

2. The system of claim 1 , wherein the probe data comprises information sent a vehicle on a roadway.

3. The system of claim 1 , wherein at least some of the one or more trajectories is associated with the vehicle passing through the one or more intersections.

4. The system of claim 3 , wherein the optimization of the scoring function is a deterministic model.

5. The system of claim 4 , wherein the deterministic model is a gradient descent model.

6. A method of estimating traffic signals, comprising:

constructing trajectories of probe vehicles from GPS data emitted by the probe vehicles passing through one or more intersections;

estimating traffic signal cycles of a traffic light at the one or more intersections based on transition times of the probe vehicles starting after the traffic light turns green; and

combining estimates of the traffic signal cycles; and

computing a traffic signal timing of the traffic light by maximizing a scoring function,

wherein the scoring function is a function of (1) an offset parameter and (2) the estimates of the traffic signal cycles of the traffic light at the one or more intersections,

wherein the scoring function is a product of terms, each term being a sum of exponentials, and

wherein the offset parameter accounts for an unknown starting time of a cycle of the traffic light at the one or more intersections.

7. The method of claim 6 , wherein the scoring function is maximized by a deterministic model.

8. The method of claim 6 , wherein the scoring function is maximized according to a gradient descent.

9. An intelligent transportation system for estimating traffic signal timing, comprising:

a sensor array for detecting probe signals, wherein the probe signals comprise probe position data;

a network connecting the sensor array with a computer processor;

a non-transitory computer readable medium containing instructions directing the system to execute steps comprising:

extracting one or more trajectories from probe data;

matching the one or more trajectories to a map, wherein the map comprises one or more intersections and the one or more trajectories passes through the one or more intersections;

determining one or more traffic signal cycles of a traffic light at the one or more intersections; and

computing a signal timing estimation for the traffic light, based on an optimization of a score function,

wherein the score function is a function of (1) an offset parameter and (2) the one or more traffic signal cycles for the traffic light at the one or more intersections,

wherein the score function is a product of terms, each term being a sum of exponentials, and

wherein the offset parameter accounts for an unknown starting time of a cycle of the traffic light at the one or more intersections.

10. A method of estimating traffic signals, comprising:

extracting one or more trajectories from probe data;

matching the one or more trajectories to a map, wherein the map comprises one or more intersections and the one or more trajectories passes through the one or more intersections;

determining one or more traffic signal cycles of a traffic light at the one or more intersections; and

computing a signal timing estimation for the traffic light, based on an optimization of a scoring function,

wherein the score function is a function of (1) an offset parameter and (2) the one or more traffic signal cycles for the traffic light at the one or more intersections,

wherein the score function is a product of terms, each term being a sum of exponentials, and

wherein the offset parameter accounts for an unknown starting time of a cycle of the traffic light at the one or more intersections.

11. The method of claim 10 , wherein the probe data comprises information sent a vehicle on a roadway.

12. The method of claim 11 , wherein at least some of the one or more trajectories is associated with the vehicle passing through the one or more intersections.

13. The method of claim 10 , wherein the optimization of the scoring function is a deterministic model.

14. The method of claim 13 , wherein the deterministic model is a gradient descent model.

15. The method of claim 11 , wherein the probe data are extracted from GPS data emitted from one or more devices disposed in one or more vehicles.

16. The method of claim 15 , wherein the one or more devices comprise a smart phone.

17. The method of claim 15 , wherein the one or more devices comprise a GPS device.

18. The method of claim 1 , wherein the score function is a product of L sub-functions, where L is a number of incoming links along the one or more trajectories of the vehicle, the incoming links being associated with the one or more intersections.

19. The method of claim 6 , wherein the score function is a product of L sub-functions, where L is a number of incoming links along the one or more trajectories of the vehicle, the incoming links being associated with the one or more intersections.

20. The method of claim 10 , wherein the score function is a product of L sub-functions, where L is a number of incoming links along the one or more trajectories of the vehicle, the incoming links being associated with the one or more intersections.