Systems and methods for traffic pattern prediction through collaborative knowledge transferring from node to node
Disclosed are systems and methods for traffic pattern prediction under abnormal behavior through collaborative knowledge transferring from node to node. In one example, a system includes a processor and a memory having instructions that cause the processor to determine vehicle traffic flows at each of a plurality of nodes using a general model that utilizes hyperparameters that derive relationships between and within each node of the plurality of nodes and observed data from sensors monitoring the plurality of nodes. The observed data includes real-world traffic data affected by hidden parameters. Using an understandable algorithm, the general model derives correlations between the hidden parameters from the observed data at multiple levels.
1 . A system comprising:
a processor; and
a memory in communication with the processor, the memory having a mathematical traffic flow module with instructions that, when executed by the processor, cause the processor to:
determine vehicle traffic flows at each of a plurality of nodes using a hierarchical Bayesian general model in which node-level traffic arrivals are modeled by a Poisson distribution with a Gamma distribution modeling hyperparameters (α, β) that link parameters between and within each node of the plurality of nodes and observed data from sensors monitoring the plurality of nodes, wherein the observed data includes real-world traffic data affected by hidden parameters, the hidden parameters comprising road geometry and weather conditions, and wherein the hierarchical Bayesian general model computes posterior distributions over the hidden parameters to derive correlations across multiple hierarchical levels and updates the hyperparameters (α, β) based on data received from the plurality of nodes; and
based on the vehicle traffic flows, control at least one of a traffic signal and an autonomous vehicle system.
2 . The system of claim 1 , wherein the hidden parameters further include one or more distances between each of the plurality of nodes, geometrical similarity of each of the plurality of nodes, and road condition at each of the plurality of nodes.
3 . The system of claim 1 , wherein the mathematical traffic flow module further comprises instructions that, when executed by the processor, cause the processor to update the hyperparameters following computations regarding statistical models at individual nodes.
4 . The system of claim 3 , wherein the mathematical traffic flow module further comprises instructions that, when executed by the processor, cause the processor to model the vehicle traffic flow at each of the plurality of nodes using the observed data.
5 . The system of claim 1 , wherein:
the observed data includes observations of events occurring at each node; and
the observed data includes at least one of loop detector sensor locations, number of loop detector events, loop detector order, and loop detector movement assignment.
6 . The system of claim 1 , wherein the observed data further includes an anomaly score for each of the plurality of nodes, the anomaly score indicating abnormal behavior at a node.
7 . The system of claim 6 , wherein the abnormal behavior indicates anomalous traffic volume data regarding known or unknown situations.
8 . The system of claim 1 , wherein the sensors include one or more of open loop detectors, camera sensors, radar sensors, and sonar sensors.
9 . The system of claim 1 , wherein the observed data includes data collected from vehicles located near the plurality of nodes.
10 . The system of claim 1 , wherein at least one of the plurality of nodes is an intersection.