# US 12,008,896 B2 · System and method for detection and quantification of irregular traffic congestion

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,008,896 B2
- Document ID: US12008896B2
- Status: Granted
- Kind: B2
- Application number: 17/405,042
- Filed: 2021-08-18
- Published: 2021-12-02
- Granted: 2024-06-11
- Earliest claimed: 2019-01-15
- Assignees: Waycare Technologies Ltd.
- Applicants: Waycare Technologies Ltd.
- Inventors: Idan Hahn; Alex Smolyak
- Web page: https://patentgrep.com/patent/US12008896B2

## CPC

- G08G 1/0112: {from the vehicle, e.g. floating car data [FCD]}
- G08G 1/0145: {for active traffic flow control}
- G06F 17/18: for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)}
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G06N 5/046: Forward inferencing; Production systems
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06N 20/00: Machine learning
- G08G 1/0116: {from roadside infrastructure, e.g. beacons}
- G08G 1/012: {from other sources than vehicle or roadside beacons, e.g. mobile networks}
- G08G 1/0129: {for creating historical data or processing based on historical data}
- G08G 1/0133: {for classifying traffic situation}
- G08G 1/02: using treadles built into the road (pads or other sensitive devices responsive to passage of vehicles E01F11/00)
- G08G 1/04: using optical or ultrasonic detectors
- G08G 1/042: using inductive or magnetic detectors
- G06N 3/045: Combinations of networks
- G08G: TRAFFIC CONTROL SYSTEMS (guiding railway traffic, ensuring the safety of railway traffic B61L; arrangement of road signs or traffic signals E01F9/00; radar or analogous systems, sonar systems, lidar systems specially adapted for traffic control G01S13/91, G01S15/88, G01S17/88; {radar or analogous systems, sonar systems, lidar systems specially adapted for anti-collision purposes G01S13/93, G01S15/93, G01S17/93})
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A system and a method of detecting and handling traffic congestion outliers by at least one processor. Embodiments of the present invention may include: receiving, from a plurality of information sources, data corresponding to at least one temporal, local traffic property; producing a local, time-based traffic profile based on the received data; receiving, from at least one information source, new data corresponding to at least one temporal, local traffic property; analyzing the new data in relation to at least one respective time-based traffic profile, to produce a score of the new data; and if the score surpasses a threshold, then identifying the data as originating from a local, temporal traffic congestion outlier and producing at least one recommendation for handling the traffic congestion outlier.

## Claims

### Claim 1

A method of detecting and handling traffic congestion outliers by at least one processor, the method comprising:
receiving, from a plurality of information sources, data corresponding to at least one temporal, local traffic property;
producing a local, time-based traffic profile based on the received data, wherein the local, time-based traffic profile reflects historic states of the at least one temporal, local traffic property;
receiving, from at least one information source, new data corresponding to the at least one temporal, local traffic property;
generating a score for the new data by:
comparing the new data to the local, time-based traffic profile so as to generate one or more severity rates, each severity rate corresponding to exactly one of the at least one temporal, local traffic property,
assigning a weight to each severity rate, and
accumulating the one or more severity rates according to the respective weights, to produce an overall accumulated weighted score as the score;
if the score surpasses a threshold, then identifying the data as originating from a local, temporal traffic congestion outlier;
clustering a plurality of segments identified as having a traffic congestion outlier to a plurality of clusters;
assigning a cluster score to each cluster, according to at least one severity rate of member congested segments;
producing a traffic report, comprising a geographic representation of the plurality of clustered segments and the clusters' respective cluster score; and
transmitting a recommendation to a user device, based on the traffic report, wherein the recommendation includes at least one of: a geographic location of a cluster of the plurality of clusters, a time-based origin of the cluster, and a cluster score of the cluster.

### Claim 2 (depends on 1)

The method of claim 1, further comprising segmenting a geographic region of interest (ROI) to a plurality of segments based on a granularity parameter, and wherein producing a local, time-based traffic profile comprises:
arranging the received data in a plurality of historic baselines, each associated with a specific segment; and
producing a time-based traffic profile per each segment based on the respective historic baseline.

### Claim 3 (depends on 2)

The method of claim 2, wherein arranging the received data in a historic baseline comprises:
fragmenting time to a plurality of discrete timestamp bins of a predefined duration;
synchronizing the temporal traffic property data of the plurality of information sources according to the discrete timestamp bins; and
aggregating data from the plurality of information sources into respective timestamp bins.

### Claim 4 (depends on 2)

The method of claim 2, wherein producing a time-based traffic profile per each segment comprises:
computing one or more values of one or more statistical parameters of the one or more traffic properties; and
aggregating the computed one or more values to produce a time-based profile, representing a normal evolution of the traffic property over time per a specific segment.

### Claim 5 (depends on 1)

The method of claim 1, further comprising comparing the accumulated weighted score to a threshold, to identify a road segment as having a congestion outlier.

### Claim 6 (depends on 5)

The method of claim 5, wherein computing one or more values of one or more statistical parameters comprises:
producing a distribution histogram of the at least one traffic property, over the plurality of discrete timestamp bins per each segment;
training a Gaussian Mixture Model (GMM) per each segment, based on the distribution histogram; and
labeling each Gaussian distribution of the GMM to represent a probability of distribution of the at least one traffic property according to a timing property.

### Claim 7

The method of 6, wherein producing one or more severity rates comprises:
predicting the value of the temporal, local traffic property according to at least one GMM;
comparing the value of the traffic property of the new data with the predicted value of the traffic property; and
producing a first severity rate of the temporal, local traffic property based on the comparison.

### Claim 8 (depends on 1)

The method of claim 1, wherein producing one or more severity rates comprises:
producing a Gaussian distribution function of the at least one local, temporal traffic property per each segment and per each timestamp bin;
calculating the difference between the new data value and the mean value of the Gaussian distribution function in z-score units; and
producing a second severity rate based on the calculation.

### Claim 9 (depends on 1)

The method of claim 1, wherein producing one or more severity rates comprises:
calculating a historical extremum value of the at least one local, temporal traffic property per each segment and per each timestamp bin;
comparing the new data value with the calculated historical extremum value; and
producing a third severity rate based on the comparison.

### Claim 10 (depends on 1)

The method of claim 1, wherein the new data is an indication of a driver behavior within a specific segment and timestamp bin, and wherein producing one or more severity rates comprises:
counting occurrences of indications of driver behavior; and
producing a forth severity rate based on the count.

### Claim 11 (depends on 1)

The method of claim 1, wherein the new data comprises textual traffic property data originating from at least one online textual data source, and wherein producing one or more severity rates comprises:
analyzing the textual data to determine whether the data indicates a condition of a traffic congestion;
cross referencing the analyzed data with geo-location data of the user;
determining if the congestion is a historical first-time occurrence of a congestion at the user's location; and
if the congestion is a historical first-time occurrence of a congestion at the user's location then producing a fifth severity rate.

### Claim 12 (depends on 1)

The method of claim 1, wherein the new data comprises numerical traffic property data originating from one or more online numerical data sources, and wherein producing one or more severity rates comprises:
extracting the numerical data;
analyzing the numerical data to determine whether the numerical data indicates a condition of a traffic congestion at a specific location;
identifying whether the determined congestion is a first-time occurrence of a congestion at the respective location; and
if the congestion is a historical first-time occurrence of a congestion at the respective location then producing a sixth severity rate.

### Claim 13 (depends on 1)

The method of claim 1, further comprising:
clustering the plurality of clusters to form a time-based cluster model of clusters; and
analyzing each such time-based cluster to determine an evolution of an outlier congestion condition over time, wherein the congestion report further comprises the evolution of at least one time-based cluster.

### Claim 14 (depends on 1)

The method of claim 1, wherein the plurality of information sources comprises at least one physical roadway detector selected from a list comprising one or more of: an induction loop, a radar, an infrared camera, a still camera, a video camera, a Light Detection and Ranging (LIDAR) sensor, a pneumatic sensor, a piezoelectric sensor, a magnetic sensor, an acoustic sensor, a doppler microwave sensor and a Bluetooth receiver.

### Claim 15 (depends on 1)

The method of claim 1, wherein the plurality of information sources comprises at least one in-vehicle sensor selected from a list comprising one or more of: a geo-location sensor, a collision detector and a proximity detector.

### Claim 16 (depends on 1)

The method of claim 1, wherein the plurality of information sources comprises a textual traffic report in a social network, and wherein the method further comprises extraction of data relating to traffic congestion from the textual traffic report.

### Claim 17 (depends on 1)

The method of claim 1, wherein the recommendation comprises a prompt for a user's attention to a condition of a congestion outlier, and wherein the method further comprises detecting a source of traffic irregularity using one or more integrated CCTV cameras according to the prompt.

### Claim 18 (depends on 1)

The method of claim 1, further comprising sending the recommendation via a cellular communication network to at least one of a freeway service patrol and a police patrol, and wherein the recommendation comprises an assignment of a task or task list to the patrol force, according to specific characteristics of the congestion outlier.

### Claim 19 (depends on 1)

The method of claim 1, further comprising sending the recommendation to a dynamic message signage (DMS) system and wherein the recommendation comprises a warning and/or an instruction to the public.

### Claim 20 (depends on 1)

The method of claim 1, further comprising training a neural network to predict at least one condition of at least one road segment given an identification of at least one traffic congestion outlier.

### Claim 21

A system for detecting traffic congestion outliers, the system comprising a non-transitory memory device comprising instruction code stored thereon, and a processor associated with the memory device, and configured to execute the instruction code, where upon execution of said instruction code the processor is configured to:
receive, from a plurality of information sources, data corresponding to at least one temporal, local traffic property;
produce a local, time-based traffic profile based on the received data, wherein the local, time-based traffic profile reflects historic states of the at least one temporal, local traffic property;
receive, from at least one information source, new data corresponding to the at least one temporal, local traffic property;
generate a score for the new data by:
comparing the new data to the local, time-based traffic profile so as to generate one or more severity rates, each severity rate corresponding to exactly one of the at least one temporal, local traffic property,
assigning a weight to each severity rate, and
accumulating the one or more severity rates according to the respective weights, to produce an overall accumulated weighted score as the score;
if the score surpasses a threshold, then identify the data as originating from a local, temporal traffic congestion outlier;
cluster a plurality of segments identified as having a traffic congestion outlier to a plurality of clusters;
assign a cluster score to each cluster, according to at least one severity rate of member congested segments;
produce a traffic report, comprising a geographic representation of the plurality of clustered segments and the clusters' respective cluster score; and
transmit a recommendation to a user device, based on the traffic report, wherein the recommendation includes at least one of: a geographic location of a cluster of the plurality of clusters, a time-based origin of the cluster, and a cluster score of the cluster.