Real-time traffic safety management system

A proactive method of managing traffic patterns, the method comprising the steps of:

• receiving, on a central server, and aggregating, via the server, a plurality of real-time data including traffic data, weather data, and social media data, the plurality of real-time data relating to and captured from a segment of a traffic system;
• overlaying graphical depictions of the plurality of real-time data on a base map representing the segment of the traffic system;
• identifying, based on the graphical depictions, one or more likely areas of the base map that represent a likelihood of a traffic event by:
• calculating a statistical likelihood of a plurality of traffic events based on a statistical model selected from a group consisting of a binary logistic model, a conditional logistic model, a Bayesian conditional logistic model, a Bayesian random parameter/effect logistic model and a combination of thereof;
• calibrating the statistical model using an under-sampling method to reduce a sample size of a plurality of non-traffic events;
• simultaneously utilizing a deep learning model to identify a pattern-based likelihood of the traffic event, the deep learning model selected from the group consisting of an artificial neural network, a recurrent neural network, and a long short-term memory model;
• training, retraining, or both, the deep learning model using an oversampling method to increase the sample size of the plurality of traffic events;
• and
• combining a result of the statistical likelihood of the traffic event with a result of the deep learning model to generate a hybrid model representing the likelihood of the traffic event on the base map;
• outputting at least one alteration of a traffic route to reduce the likelihood of the traffic event by:
• utilizing the hybrid model as an input to a calculation of the at least one alteration;
• fitting a plurality of countermeasures stored in a database over the base map;
• calculating a difference between each of the plurality of countermeasures and the hybrid model by subjecting each of the plurality of countermeasures to one or more virtual simulations;
• and
• selecting at least one of the plurality of countermeasures as the at least one alteration of the traffic route to reduce the likelihood of the traffic event;
• and
• graphically displaying the at least one alteration of the traffic route on the base map.

A proactive traffic management system comprising:

• a data collection server in communication with a first raw data server, wherein the first raw data server receives a plurality of raw data from the data collection server;
• a first data filter server in communication with the first raw data server, the first data filter server configured to receive the plurality of raw data from the first raw data server and filter the plurality of raw data to create a plurality of filtered data wherein the first data filter server is configured to calibrate using an under-sampling method the plurality of raw data, whereby at least one non-traffic event is removed from the plurality of filtered data;
• an action database server in communication with the first data filter server, wherein the action database server receives the plurality of filtered data from the first data filter server and compares the plurality of filtered data with a plurality of historic data to output a suggested dataset from the plurality of filtered data, wherein the action database server is configured to be trained, retrained, or both, via a deep learning model using an oversampling method to input at least one additional traffic event into the plurality of filtered data;
• a second raw data server in communication with each of the action database server and a second data filter server, wherein the second raw data server receives the suggested dataset from the action database server and transmits the suggested dataset to the second data filter server;
• a real-time database server including a base map of a segment of a traffic system, the real-time database server in communication with the second data filter server, wherein the real-time database server receives a filtered portion of the suggested dataset from the second data filter server;
• and
• a web server in communication with the real-time database server, the web server including a graphical user interface having a real-time display controller that is configured to receive an input from a user, wherein the web server receives a final dataset from the real-time database, the final dataset including one or more real-time data points related to the segment of the traffic system, and wherein the web server displays the final dataset overlayed on the base map,
• wherein the final dataset is overlayed on the base map, whereby the base map is configured to displays at least one alteration of the traffic route, thereby improving a traffic pattern thereof.