# US 2025/0148381 A1 · Apparatus and method for estimating traffic volume based on demand of route search

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

- Number: US 2025/0148381 A1
- Document ID: US20250148381A1
- Status: Published application
- Kind: A1
- Application number: 18/668,769
- Filed: 2024-05-20
- Published: 2025-05-08
- Earliest claimed: 2023-11-08
- Assignees: Hyundai Motor Company; Kia Corporation
- Applicants: Hyundai Motor Company; Kia Corporation
- Inventors: Nam Hyuk Kim
- Web page: https://patentgrep.com/patent/US20250148381A1

## CPC

- G06Q 10/047: Optimisation of routes or paths, e.g. travelling salesman problem
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

An apparatus for estimating a traffic volume based on demand for route search, includes a processor and a storage medium configured to record one or more programs configured to be executable by the processor. The processor is configured to collect a plurality of pieces of route search data, generate route search demand data based on the collected plurality of pieces of route search data, correct the route search demand data based on an overcrowded road exceeding a marginal traffic volume, and estimate an actual traffic volume for each road by applying the corrected route search demand data to a pre-trained learning model.

## Claims

### Claim 1

An apparatus for estimating a traffic volume based on demand for route search, the apparatus comprising:
a processor; and
a storage medium configured to record one or more programs configured to be executable by the processor;
wherein the processor is configured to:
collect a plurality of pieces of route search data;
generate route search demand data based on the collected plurality of pieces of route search data;
correct the route search demand data based on an overcrowded road exceeding a marginal traffic volume; and
estimate an actual traffic volume for a plurality of roads by applying the corrected route search demand data to a pre-trained learning model.

### Claim 2 (depends on 1)

The apparatus of claim 1, wherein:
each of the plurality of pieces of route search data includes a road ID of each of the plurality of roads, included in a route searched by a vehicle, and a predicted entry time for each road of the plurality of roads;
the route search demand data includes a traffic volume for each road of the plurality of roads estimated based on the road ID and the predicted entry time, the traffic volume representing a number of vehicles; and
the marginal traffic volume is a maximum number of vehicles set for each road of the plurality of roads.

### Claim 3 (depends on 2)

The apparatus of claim 2, wherein the processor is configured to disperse, with respect to the overcrowded road, an excess demand traffic volume exceeding the marginal traffic volume to an upstream road connected to the overcrowded road, the upstream road being a road connected to the overcrowded road in a reverse direction of a direction of travel.

### Claim 4 (depends on 3)

The apparatus of claim 3, wherein the processor is configured to:
repeatedly calculate a process of dispersing the excess demand traffic volume to the upstream road according to a preset percentage; and
end the process of dispersing the excess demand traffic volume to the upstream road when a total sum of excess demand traffic volumes for respective roads, after dispersion is performed, is less than or equal to a certain percentage of a total sum of marginal traffic volumes for respective roads.

### Claim 5 (depends on 3)

The apparatus of claim 3, wherein the route search demand data and the marginal traffic volume are represented by a matrix.

### Claim 6 (depends on 3)

The apparatus of claim 3, wherein the processor is configured to perform dispersion according to:
D
over
,
t
prop
=
P
prop
,
t
*
D
over
,
t
,
where Dover,tprop is a matrix representing a dispersion volume for an excess demand traffic volume for each road at a specific point in time t, Pprop,t is a matrix representing a dispersion percentage at the specific point in time t, and Dover,t is a matrix representing an excess demand traffic volume for each road at the specific point in time t.

### Claim 7 (depends on 6)

The apparatus of claim 6, wherein Pprop,t is obtained by:
P
prop
,
t
=
α
*
diag
⁢
(
U
over
,
t
)
*
I
+
(
1
-
α
)
*
A
,
where α is a constant, Uover,t is a unit matrix of a matrix representing an excess demand traffic volume for each road at the specific point in time t, diag( ) is a function turning Uover,t into a diagonal matrix, I is a unit matrix, and A is a matrix representing a connection relationship between roads according to the direction of travel.

### Claim 8 (depends on 6)

The apparatus of claim 6, wherein the plurality of specific points in time have a predetermined time interval.

### Claim 9 (depends on 1)

The apparatus of claim 1, wherein the learning model includes a generative adversarial network (GAN) including a generator and a discriminator.

### Claim 10 (depends on 9)

The apparatus of claim 9, wherein the processor is configured to train the discriminator using estimated traffic volume data, the estimated traffic volume data generated by the generator based on corrected route search demand data and actual traffic volume data, and then to train the generator in a direction of deceiving the trained discriminator.

### Claim 11

A method for estimating a traffic volume based on demand for route search, the method comprising:
collecting, by a processor, a plurality of pieces of route search data;
generating route search demand data based on the collected plurality of pieces of route search data;
correcting the route search demand data based on an overcrowded road exceeding a marginal traffic volume; and
estimating an actual traffic volume for a plurality of roads by applying the corrected route search demand data to a pre-trained learning model.

### Claim 12 (depends on 11)

The method of claim 11, wherein:
each of the plurality of pieces of route search data includes a road ID of each of the plurality of roads, included in a route searched by a vehicle, and a predicted entry time for each road of the plurality of roads;
the route search demand data includes a traffic volume for each road of the plurality of roads estimated based on the road ID and the predicted entry time, the traffic volume representing a number of vehicles; and
the marginal traffic volume is a maximum number of vehicles set for each road of the plurality of roads.

### Claim 13 (depends on 12)

The method of claim 12, wherein the correcting includes dispersing, with respect to the overcrowded road, an excess demand traffic volume exceeding the marginal traffic volume to an upstream road connected to the overcrowded road, the upstream road being a road connected to the overcrowded road in a reverse direction of a direction of travel.

### Claim 14 (depends on 13)

The method of claim 13, wherein the correcting further includes:
repeatedly calculating a process of dispersing the excess demand traffic volume to the upstream road according to a preset percentage; and
ending the process of dispersing the excess demand traffic volume to the upstream road when a total sum of excess demand traffic volumes for respective roads, after dispersion is performed, is less than or equal to a certain percentage of a total sum of marginal traffic volumes for respective roads.

### Claim 15 (depends on 13)

The method of claim 13, wherein the route search demand data and the marginal traffic volume are represented by a matrix.

### Claim 16 (depends on 13)

The method of claim 13, wherein the dispersing includes performing dispersion according to:
D
over
,
t
prop
=
P
prop
,
t
*
D
over
,
t
,
where Dover,tprop is a matrix representing a dispersion volume for an excess demand traffic volume at a specific point in time t, Pprop,t is a matrix representing a dispersion percentage at the specific point in time t, and Dover,t is a matrix representing an excess demand traffic volume at the specific point in time t.

### Claim 17 (depends on 16)

The method of claim 16, wherein Pprop,t is obtained according to:
P
prop
,
t
=
α
*
diag
⁢
(
U
over
,
t
)
*
I
+
(
1
-
α
)
*
A
,
where α is a constant, Uover,t is a unit matrix of a matrix representing an excess demand traffic volume at the specific point in time t, diag( ) is a function turning Uover,t into a square matrix, and A is a matrix representing a connection relationship between roads according to the direction of travel.

### Claim 18 (depends on 16)

The method of claim 16, wherein the plurality of specific points in time have a predetermined time interval.

### Claim 19 (depends on 11)

The method of claim 11, wherein the learning model includes a generative adversarial network (GAN) including a generator and a discriminator.

### Claim 20 (depends on 19)

The method of claim 19, further comprising:
training the discriminator using estimated traffic volume data, the estimated traffic volume data generated by the generator based on corrected route search demand data, and actual traffic volume data, and then training the generator in a direction of deceiving the trained discriminator.
