# US 10,769,866 B2 · Generating estimates of failure risk for a vehicular component

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

- Number: US 10,769,866 B2
- Document ID: US10769866B2
- Status: Granted
- Kind: B2
- Application number: 14/498,703
- Filed: 2014-09-26
- Published: 2016-03-31
- Granted: 2020-09-08
- Earliest claimed: 2014-09-26
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Hyung-il Ahn; Matthew Denesuk; Axel Hochstein; Ying Tat Leung
- Web page: https://patentgrep.com/patent/US10769866B2

## CPC

- G07C 5/006: {Indicating maintenance}
- B60K 35/90: Calibration of instruments, e.g. setting initial or reference parameters; Testing of instruments, e.g. detecting malfunction
- G05B 23/024: {Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks}
- G07C 5/008: {communicating information to a remotely located station (transmission systems for measured values G08C)}
- G07C 5/0808: {Diagnosing performance data (testing of vehicles G01M17/00; testing of electrical installation on vehicles G01R31/005)}
- B60K 35/22: Display screens
- 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
- B60K: ARRANGEMENT OR MOUNTING OF PROPULSION UNITS OR OF TRANSMISSIONS IN VEHICLES; ARRANGEMENT OR MOUNTING OF PLURAL DIVERSE PRIME-MOVERS IN VEHICLES; AUXILIARY DRIVES FOR VEHICLES; INSTRUMENTATION OR DASHBOARDS FOR VEHICLES; ARRANGEMENTS IN CONNECTION WITH COOLING, AIR INTAKE, GAS EXHAUST OR FUEL SUPPLY OF PROPULSION UNITS IN VEHICLES
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)

## Abstract

Methods, systems, and computer program products for generating estimates of failure risk for a vehicular component are provided herein. A method includes splitting an input time series pertaining to a vehicular component across a fleet of multiple vehicles into multiple sub-time series, wherein each sub-time series comprises multiple data points of the input time series that correspond to measurements derived from the vehicular component; determining a weight applied to each of the sub-time series based on a pre-determined weight associated with the input time series; applying a failure or non-failure classification label to each of the sub-time series and the input time series; calculating a performance measure for the input time series; determining an updated weight associated with the input time series; and generating an estimate of failure risk for the vehicular component based on the classification label applied to each input time series and the updated weight.

## Claims

### Claim 1

A method comprising:
splitting an input time series pertaining to a vehicular component across a fleet of multiple vehicles into multiple sub-time series, wherein each of the multiple sub-time series comprises multiple data points of the input time series that correspond to one of multiple measurements derived from the vehicular component;
determining a weight applied to each of the multiple sub-time series based on a pre-determined weight associated with the input time series, wherein said determining comprises associating the pre-determined weight associated with the input time series with each of the multiple sub-time series;
applying one of two classification labels to each of the multiple sub-time series based on a relationship of the data points of the given sub-time series and a selected threshold value, wherein the classification labels comprise (i) associated with a failure of the vehicular component and (ii) not associated with a failure of the vehicular component;
applying one of the two classification labels to the input time series to correspond to the label applied to a majority of the multiple sub-time series;
calculating a performance measure for the input time series based on (i) the weight applied to each of the multiple sub-time series, and (ii) a determination as to whether the correct classification label has been applied to each of the multiple sub-time series;
determining an updated weight associated with the input time series based on (i) the pre-determined weight associated with the input time series, (ii) the weight applied to each of the multiple sub-time series, and/or (iii) the performance measure, wherein said determining the updated weight associated with the input time series comprises calculating w*exp(p), if any of the one or more sub-time series is classified incorrectly, and reusing the pre-determined weight otherwise, wherein (w) represents the pre-determined weight and (p) represents the performance measure;
generating an estimate of failure risk for the vehicular component based on (i) the classification label applied to each input time series and (ii) the updated weight associated with the input time series; and
replacing the vehicular component in each of the multiple vehicles based at least in part on the estimate of failure risk for the vehicular component;
wherein the method is carried out by at least one computing device.

### Claim 2 (depends on 1)

The method of claim 1, comprising:
obtaining the input time series pertaining to the vehicular component across the fleet of multiple vehicles, wherein the input time series comprises multiple data points, and wherein the multiple data points comprise a collection of multiple measurements derived from the vehicular component across the fleet of multiple vehicles.

### Claim 3 (depends on 1)

The method of claim 1, comprising:
repeating all of said steps for a given number of additional input time series pertaining to the vehicular component.

### Claim 4 (depends on 1)

The method of claim 1, wherein said pre-determined weight comprises a uniform weight for all input time series pertaining to the vehicular component.

### Claim 5 (depends on 1)

A computer program product, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing device to cause the computing device to carry out the steps of claim 1.

### Claim 6 (depends on 1)

A system comprising a memory, and at least one processor coupled to the memory and configured for carrying out the steps of claim 1.

### Claim 7

A method comprising:
splitting an input time series pertaining to a vehicular component across a fleet of multiple vehicles into multiple sub-time series, wherein each of the multiple sub-time series comprises multiple data points of the input time series that correspond to one of multiple measurements derived from the vehicular component;
determining a weight applied to each of the multiple sub-time series based on a pre-determined weight associated with the input time series, wherein said determining comprises evenly distributing the pre-determined weight associated with the input time series across the multiple sub-time series;
applying one of two classification labels to each of the multiple sub-time series based on a relationship of the data points of the given sub-time series and a selected threshold value, wherein the classification labels comprise (i) associated with a failure of the vehicular component and (ii) not associated with a failure of the vehicular component;
applying one of the two classification labels to the input time series to correspond to the label applied to a majority of the multiple sub-time series;
calculating a performance measure for the input time series based on (i) the weight applied to each of the multiple sub-time series, and (ii) a determination as to whether the correct classification label has been applied to each of the multiple sub-time series;
determining an updated weight associated with the input time series based on (i) the pre-determined weight associated with the input time series, (ii) the weight applied to each of the multiple sub-time series, and/or (iii) the performance measure, wherein said determining the updated weight (w) associated with the input time series comprises calculating:
w
=
∑
all
⁢
⁢
sub
⁢
-
⁢
time
⁢
⁢
series
⁢
{
assigned
⁢
⁢
weight
⁢
⁢
to
⁢
⁢
a
⁢
⁢
given
sub
⁢
-
⁢
time
⁢
⁢
series
*
exp
⁡
(
p
)
,
if
⁢
⁢
the
⁢
⁢
given
⁢
⁢
sub
⁢
-
⁢
time
⁢
⁢
series
is
⁢
⁢
classified
⁢
⁢
wrongly
and
w
=
∑
all
⁢
⁢
sub
⁢
-
⁢
time
⁢
⁢
series
⁢
{
assigned
⁢
⁢
weights
⁢
⁢
to
⁢
⁢
the
⁢
⁢
given
⁢
⁢
sub
⁢
-
⁢
time
⁢
⁢
series
,
otherwise
,
wherein (p) represents the performance measure;
generating an estimate of failure risk for the vehicular component based on (i) the classification label applied to each input time series and (ii) the updated weight associated with the input time series; and
replacing the vehicular component in each of the multiple vehicles based at least in part on the estimate of failure risk for the vehicular component;
wherein the method is carried out by at least one computing device.

### Claim 8 (depends on 7)

The method of claim 7, wherein said pre-determined weight comprises a uniform weight for all input time series pertaining to the vehicular component.

### Claim 9 (depends on 7)

A computer program product, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing device to cause the computing device to carry out the steps of claim 7.

### Claim 10 (depends on 7)

A system comprising a memory, and at least one processor coupled to the memory and configured for carrying out the steps of claim 7.

### Claim 11

A method comprising:
splitting an input time series pertaining to a vehicular component across a fleet of multiple vehicles into multiple sub-time series, wherein each of the multiple sub-time series comprises multiple data points of the input time series that correspond to one of multiple measurements derived from the vehicular component;
determining a weight applied to each of the multiple sub-time series based on a pre-determined weight associated with the input time series, wherein said determining comprises distributing the pre-determined weight associated with the input time series to each of the multiple sub-time series inversely proportional to the length of each of the multiple sub-time series;
applying one of two classification labels to each of the multiple sub-time series based on a relationship of the data points of the given sub-time series and a selected threshold value, wherein the classification labels comprise (i) associated with a failure of the vehicular component and (ii) not associated with a failure of the vehicular component;
applying one of the two classification labels to the input time series to correspond to the label applied to a majority of the multiple sub-time series;
calculating a performance measure for the input time series based on (i) the weight applied to each of the multiple sub-time series, and (ii) a determination as to whether the correct classification label has been applied to each of the multiple sub-time series;
determining an updated weight associated with the input time series based on (i) the pre-determined weight associated with the input time series, (ii) the weight applied to each of the multiple sub-time series, and/or (iii) the performance measure, wherein said determining the updated weight (w) associated with the input time series comprises calculating:
w
=
∑
all
⁢
⁢
sub
⁢
-
⁢
time
⁢
⁢
series
⁢
{
assigned
⁢
⁢
weight
⁢
⁢
to
⁢
⁢
a
⁢
⁢
given
sub
⁢
-
⁢
time
⁢
⁢
series
*
exp
⁡
(
p
)
,
if
⁢
⁢
the
⁢
⁢
given
⁢
⁢
sub
⁢
-
⁢
time
⁢
⁢
series
is
⁢
⁢
classified
⁢
⁢
wrongly
and
w
=
∑
all
⁢
⁢
sub
⁢
-
⁢
time
⁢
⁢
series
⁢
{
assigned
⁢
⁢
weights
⁢
⁢
to
⁢
⁢
the
⁢
⁢
given
⁢
⁢
sub
⁢
-
⁢
time
⁢
⁢
series
,
otherwise
,
wherein (p) represents the performance measure;
generating an estimate of failure risk for the vehicular component based on (i) the classification label applied to each input time series and (ii) the updated weight associated with the input time series; and
replacing the vehicular component in each of the multiple vehicles based at least in part on the estimate of failure risk for the vehicular component;
wherein the method is carried out by at least one computing device.

### Claim 12 (depends on 11)

The method of claim 11, wherein said pre-determined weight comprises a uniform weight for all input time series pertaining to the vehicular component.

### Claim 13 (depends on 11)

A computer program product, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing device to cause the computing device to carry out the steps of claim 11.

### Claim 14 (depends on 11)

A system comprising a memory, and at least one processor coupled to the memory and configured for carrying out the steps of claim 11.

### Claim 15

A method comprising:
splitting an input time series pertaining to a vehicular component across a fleet of multiple vehicles into multiple sub-time series, wherein each of the multiple sub-time series comprises multiple data points of the input time series that correspond to one of multiple measurements derived from the vehicular component;
determining a weight applied to each of the multiple sub-time series based on a pre-determined weight associated with the input time series, wherein said determining comprises distributing the pre-determined weight associated with the input time series exclusively to the temporally last of the multiple sub-time series;
applying one of two classification labels to each of the multiple sub-time series based on a relationship of the data points of the given sub-time series and a selected threshold value, wherein the classification labels comprise (i) associated with a failure of the vehicular component and (ii) not associated with a failure of the vehicular component;
applying one of the two classification labels to the input time series to correspond to the label applied to a majority of the multiple sub-time series;
calculating a performance measure for the input time series based on (i) the weight applied to each of the multiple sub-time series, and (ii) a determination as to whether the correct classification label has been applied to each of the multiple sub-time series;
determining an updated weight associated with the input time series based on (i) the pre-determined weight associated with the input time series, (ii) the weight applied to each of the multiple sub-time series, and/or (iii) the performance measure, wherein said determining the updated weight associated with the input time series comprises: calculating w*exp(p), if the last of the multiple sub-time series is classified incorrectly and reusing the pre-determined weight otherwise, wherein (w) represents the pre-determined weight and (p) represents the performance measure;
generating an estimate of failure risk for the vehicular component based on (i) the classification label applied to each input time series and (ii) the updated weight associated with the input time series; and
replacing the vehicular component in each of the multiple vehicles based at least in part on the estimate of failure risk for the vehicular component;
wherein the method is carried out by at least one computing device.

### Claim 16 (depends on 15)

The method of claim 15, wherein said pre-determined weight comprises a uniform weight for all input time series pertaining to the vehicular component.

### Claim 17 (depends on 15)

A computer program product, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing device to cause the computing device to carry out the steps of claim 15.

### Claim 18 (depends on 15)

A system comprising a memory, and at least one processor coupled to the memory and configured for carrying out the steps of claim 15.