System and method for predicting and maximizing traffic flow
AssigneesVolkswagen Aktiengesellschaft, Audi AG, Porsche AG
InventorsFlorian NEUKART, Gabriele COMPOSTELLA, Christian SEIDEL, David VON DOLLEN
Life of this patent
Estimating the term…
Abstract
A traffic flux maximization method and system to control traffic flow by combining classical computing machine learning to predict traffic flux minimization before its occurrence, with quantum annealing to optimize future positions of vehicles. Vehicles are redirected to minimize the travel time for each vehicle, taking into account other vehicles in the road network.
Claims18
A vehicle traffic control system comprising:
- a traffic monitoring system configured using classic computing to monitor traffic flow in a selected area as a point in time;
- a traffic prediction system configured using classic computing to predict traffic flow minimization prior to the occurrence of the traffic minimization flow based on information from the traffic monitoring system;
- and
- a means for combining machine learning on classical computing to predict traffic flux minimization before occurrence, with quantum annealing to optimize future positions of vehicles such that a redirection minimizes the travel time for each vehicle under consideration of all the other cars in the road network.
A method for controlling traffic flow comprising:
- predicting traffic flux minimization before occurrence using machine learning on classical computing;
- and
- optimizing future positions of vehicles such that a redirection maximizes traffic flux to minimize travel time for each vehicle, wherein optimization includes accounting for other vehicles in a road network using quantum annealing.
A non-transitory computer-readable storage medium in which is stored computer code that when executed on one or more central processing units and one or more quantum processing units causes a traffic flow maximization system to perform the method of:
- using classical computing:
- identifying areas where traffic flux minimization occurs;
- determining spatially and temporally valid alternative routes;
- and
- formulating the problem as an energy matrix wherein the elements of the matrix are the coefficients of terms of a cost function;
- using quantum annealing:
- finding a minimum energy solution, which represents an ideal distribution of vehicles by minimizing the time in a selected whole traffic graph;
- using classical computing:
- redistributing the vehicle based on the matrix produced;
- and
- repeating the above steps to maximize traffic flow by approaching prediction of no traffic flow minimization for a given time window.
Dependencies are inferred from the claim text.