# US 2023/0055974 A1 · Method of Selecting a Route for Recording Vehicle

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

- Number: US 2023/0055974 A1
- Document ID: US20230055974A1
- Status: Published application
- Kind: A1
- Application number: 17/820,814
- Filed: 2022-08-18
- Published: 2023-02-23
- Earliest claimed: 2021-08-18
- Assignees: Aptiv Technologies AG
- Applicants: Aptiv Technologies Limited
- Inventors: Michael Arnold; Sönke Behrends; Peet Cremer; Dennis Mueller
- Web page: https://patentgrep.com/patent/US20230055974A1

## CPC

- G01C 21/3461: {Preferred or disfavoured areas, e.g. dangerous zones, toll or emission zones, intersections, manoeuvre types or segments such as motorways, toll roads or ferries}
- G08G 1/0112: {from the vehicle, e.g. floating car data [FCD]}
- G01C 21/3492: {employing speed data or traffic data, e.g. real-time or historical (traffic control systems for road vehicles involving transmission of navigation instructions to the vehicle G08G1/0968)}
- G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- G06N 20/00: Machine learning
- G08G 1/0129: {for creating historical data or processing based on historical data}
- G08G 1/0145: {for active traffic flow control}
- G07C 5/008: {communicating information to a remotely located station (transmission systems for measured values G08C)}
- G01C: MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY (measuring liquid level G01F; radio navigation, determining distance or velocity by use of propagation effects, e.g. Doppler effects, propagation time, of radio waves, analogous arrangements using other waves G01S)
- 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})
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G07C: TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE

## Abstract

Disclosed are aspects of a method, carried out by a vehicle data recording device, that includes downloading, from a host data collecting system, a recording target and determining a plurality of possible routes for the vehicle. The method also includes, for each route, generating a route encoding in numerical values an information on predicted values of the route for metrics, a metric being a function assigning a value representing an amount of progress in achieving an elementary recording target to a piece of data. The method additionally includes providing the route encodings and additional environmental information independent of the routes to a reinforcement learning agent that selects one of the routes to optimize a reward. The method further includes recording data from in-vehicle sources while the vehicle drives along the selected route and uploading at least part of the recording data to the host data collecting system.

## Claims

### Claim 1

A method comprising:
downloading, from a host data collecting system, a recording target (T0) for recording data along a route;
at a current time (t0), determining a plurality of routes that a vehicle can take;
for each route (Ri) of the plurality of routes, generating a route encoding that encodes in numerical values an information on predicted values of the route (Ri) for a plurality of metrics, a metric (vck) of the plurality of metrics being a function assigning a value representing an amount of progress in achieving an elementary recording target to a piece of data;
providing a plurality of route encodings and additional environmental information that is independent of the plurality of routes to a reinforcement learning agent that selects a route of the plurality of routes in order to optimize a reward;
recording data from data sources in the vehicle over time while the vehicle drives along the selected route (Rs); and
uploading at least part of the recording data to the host data collecting system.

### Claim 2 (depends on 1)

The method according to claim 1, further comprising:
in return for the uploading of the at least part of the recording data, receiving a reward from the host data collecting system; and
providing the reward to the reinforcement learning agent.

### Claim 3 (depends on 1)

The method according to claim 1, wherein the generating a route encoding for each route (Ri) comprises:
dividing the route (Ri) into route segments (RSij); and
generating a feature vector (Vij) for each route segment (Rsij),
wherein environmental features related to the route segments (Rsij) and relevant for the plurality of metrics are encoded into numerical values to form a component (Bij) of the feature vector (Vij).

### Claim 4 (depends on 3)

The method according to claim 3, wherein the generating a feature vector (Vij) for each route segment (RSij) comprises:
encoding into numerical values past information on what happened before the current time (t0) to form another component (Aij) of the feature vector (Vij).

### Claim 5 (depends on 3)

The method according to claim 3, further comprising:
for each route segment (RSij),
inputting a corresponding feature vector (Vij) into a machine learning system of value prediction for the plurality of metrics; and
providing, at an output of the machine learning system, a metric value prediction vector including predicted values of the route segment (RSij) for the plurality of metrics.

### Claim 6 (depends on 5)

The method according to claim 5, wherein the metric value prediction vector of each route segment (RSij) includes an uncertainty information.

### Claim 7 (depends on 5)

The method according to claim 5, further comprising:
grouping the metric value prediction vector of each route segment (RSij) of each route (Ri) into a high-dimensional route encoding;
translating the high-dimensional route encoding of each route (Ri) of the plurality of routes into a fixed-length route encoding of lower dimension; and
providing the fixed-length route encoding of each route (Ri) of the plurality of routes as inputs to the reinforcement learning agent.

### Claim 8 (depends on 1)

The method according to claim 1, further comprising:
pre-selecting a short list of promising route encodings; and
providing only the short list of promising route encodings as input to the reinforcement learning agent.

### Claim 9 (depends on 1)

The method according to claim 1, further comprising:
in a preliminary operation,
obtaining historical data related to a plurality of recordings performed when driving along given past routes;
dividing each past route into route segments (RSij);
evaluating values of each route segment (Rsij) for the plurality of metrics based on the historical data; and
training a machine learning system to compute the evaluated values related to the plurality of metrics for corresponding route segments.

### Claim 10 (depends on 1)

The method according to claim 1, further comprising:
determining the plurality of routes using a tree expansion algorithm.

### Claim 11

A vehicle data recording device comprising:
at least one interface configured to receive recording data from data sources in a vehicle and to transmit the recording data to upload to a host data collecting system;
a memory configured to store the recording data; and
a processor coupled to the memory, the processor configured to cause the vehicle data recording device to:
download, from the host data collecting system, a recording target (T0) for recording data along a route;
at a current time (t0), determine a plurality of routes that a vehicle can take;
for each route (Ri) of the plurality of routes, generate a route encoding that encodes in numerical values an information on predicted values of the route (Ri) for a plurality of metrics, a metric (vck) of the plurality of metrics being a function assigning a value representing an amount of progress in achieving an elementary recording target to a piece of data;
provide a plurality of route encodings and additional environmental information that is independent of the plurality of routes to a reinforcement learning agent that selects a route (Rs) of the plurality of routes in order to optimize a reward;
record data from data sources in the vehicle over time while the vehicle drives along the selected route (Rs) to produce the recording data; and
upload at least part of the recording data to the host data collecting system.

### Claim 12 (depends on 11)

The vehicle data recording device according to claim 11, wherein the processor is further configured to cause the vehicle data recording device to:
receive a reward from the host data collecting system in return for the upload of the at least part of the recording data; and
provide the reward to the reinforcement learning agent.

### Claim 13 (depends on 11)

The vehicle data recording device according to claim 11, wherein to generate a route encoding for each route (Ri), the processor is further configured to cause the vehicle data recording device to:
divide the route (Ri) into route segments (RSij); and
generate a feature vector (Vij) for each route segment (Rsij),
wherein environmental features related to the route segments (Rsij) and relevant for the plurality of metrics are encoded into numerical values to form a component (Bij) of the feature vector (Vij).

### Claim 14 (depends on 13)

The vehicle data recording device according to claim 13, wherein to generate a feature vector (Vij) for each route segment (RSij), the processor is further configured to cause the vehicle data recording device to:
encode into numerical values past information on what happened before the current time (t0) to form another component (Aij) of the feature vector (Vij).

### Claim 15 (depends on 13)

The vehicle data recording device according to claim 13, wherein the processor is further configured to cause the vehicle data recording device to:
for each route segment (RSij),
input a corresponding feature vector (Vij) into a machine learning system of value prediction for the plurality of metrics; and
provide, at an output of the machine learning system, a metric value prediction vector including predicted values of the route segment (RSij) for the plurality of metrics.

### Claim 16 (depends on 15)

The vehicle data recording device according to claim 15, wherein the metric value prediction vector of each route segment (RSij) includes an uncertainty information.

### Claim 17 (depends on 15)

The vehicle data recording device according to claim 15, wherein the processor is further configured to cause the vehicle data recording device to:
group the metric value prediction vector of each route segment (RSij) of each route (Ri) into a high-dimensional route encoding;
translate the high-dimensional route encoding of each route (Ri) of the plurality of routes into a fixed-length route encoding of lower dimension; and
provide the fixed-length route encoding of each route (Ri) of the plurality of routes as inputs to the reinforcement learning agent.

### Claim 18 (depends on 11)

The vehicle data recording device according to claim 11, wherein the processor is further configured to cause the vehicle data recording device to:
pre-select a short list of promising route encodings; and
provide only the short list of promising route encodings as input to the reinforcement learning agent.

### Claim 19 (depends on 11)

The vehicle data recording device according to claim 11, wherein the processor is further configured to cause the vehicle data recording device to:
in a preliminary operation,
obtain historical data related to a plurality of recordings performed when driving along given past routes;
divide each past route into route segments (RSij);
evaluate values of each route segment (Rsij) for the plurality of metrics based on the historical data; and
train a machine learning system to compute the evaluated values related to the plurality of metrics for corresponding route segments.

### Claim 20

A vehicle comprising:
a vehicle data recording device comprising:
at least one interface configured to receive recording data from data sources in the vehicle and to transmit the recording data to upload to a host data collecting system;
a memory configured to store the recording data; and
a processor coupled to the memory, the processor configured to cause the vehicle data recording device to:
download, from the host data collecting system, a recording target (T0) for recording data along a route;
at a current time (t0), determine a plurality of routes that a vehicle can take;
for each route (Ri) of the plurality of routes, generate a route encoding that encodes in numerical values an information on predicted values of the route (Ri) for a plurality of metrics, a metric (vck) of the plurality of metrics being a function assigning a value representing an amount of progress in achieving an elementary recording target to a piece of data;
provide a plurality of route encodings and additional environmental information that is independent of the plurality of routes to a reinforcement learning agent that selects a route (Rs) of the plurality of routes in order to optimize a reward;
record data from data sources in the vehicle over time while the vehicle drives along the selected route (Rs) to produce the recording data; and
upload at least part of the recording data to the host data collecting system.
