# US 2013/0253797 A1 · Prediction of driver-specific cruise speed using dynamic modeling

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

- Number: US 2013/0253797 A1
- Document ID: US20130253797A1
- Status: Published application
- Kind: A1
- Application number: 13/424,477
- Filed: 2012-03-20
- Published: 2013-09-26
- Earliest claimed: 2012-03-20
- Assignees: Toyota Motor Engineering & Manufacturing North America (TEMA)
- Inventors: John Michael MCNEW
- Web page: https://patentgrep.com/patent/US20130253797A1

## CPC

- B60W 30/143: {Speed control (B60W30/16 takes precedence)}
- B60W 50/0097: {Predicting future conditions}
- G06N 20/00: Machine learning
- B60K 2310/244: changing target speed or setting a new target speed, e.g. changing algorithms
- B60W 2050/0029: {Mathematical model of the driver}
- B60W 2050/0083: {Setting, resetting, calibration}
- B60W 2540/30: Driving style
- B60W 2552/05: Type of road, e.g. motorways, local streets, paved or unpaved roads
- B60W 2720/10: Longitudinal speed
- B60W 2720/103: Speed profile
- Y02T 10/84: Data processing systems or methods, management, administration
- B60W: CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- 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
- Y02T: CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION

## Abstract

A controller for predicting cruising speeds of a vehicle includes a processor and an extracting unit to extract feature data from segments of a prior trajectory of the vehicle, the feature data including cruising speeds of the vehicle and predictive feature data. The controller also includes a model generator to generate a probabilistic model associating the predictive feature data with the cruising speeds of the vehicle and a predicting unit to predict a cruising speed of the vehicle for a target segment, which is an upcoming cruising segment of the vehicle, by conditioning the probabilistic model on real-time predictive feature data of segments of a current trajectory.

## Claims

### Claim 1

A controller for predicting cruising speeds of a vehicle, the controller including a processor and comprising:
an extracting unit configured to extract feature data from segments of a prior trajectory of the vehicle, the feature data including cruising speeds of the vehicle and predictive feature data;
a model generator configured to generate a probabilistic model associating the predictive feature data with the cruising speeds of the vehicle; and
a predicting unit configured to predict a cruising speed of the vehicle for a target segment, which is an upcoming cruising segment of the vehicle, by conditioning the probabilistic model on real-time predictive feature data of segments of a current trajectory.

### Claim 2 (depends on 1)

The controller according to claim 1, wherein
the predicting unit is configured to condition the probabilistic model on a launch model of an initial segment of the segments of the current trajectory, and
the target segment follows the initial segment.

### Claim 3 (depends on 2)

The controller according to claim 2, further comprising:
a collecting unit configured to collect trajectory data of the vehicle;
a segmenting unit configured to segment the collected trajectory data into segments; and
a selecting unit configured to select one or more predictive features, wherein
the extracting unit is configured to extract feature data corresponding to the selected one or more predictive features from the segments of the current trajectory as the real-time predictive feature data, and
the model generator is configured to generate the probabilistic model based on the feature data corresponding to the selected one or more predictive features.

### Claim 4 (depends on 3)

The controller according to claim 3, further comprising an updating unit configured to update the probabilistic model based on the real-time predictive feature data and an actual cruising speed of the vehicle for the target segment.

### Claim 5 (depends on 4)

The controller according to claim 4, further comprising:
a driver identification unit configured to identify a driver of the vehicle and associate the generated probabilistic model with the identified driver; and
a transmitter configured to transmit a message in response to a determination that the actual cruising speed is different by a predetermined threshold from a predicted cruising speed for the target segment based on the probabilistic model associated with the identified driver.

### Claim 6 (depends on 5)

The controller according to claim 5, wherein the predetermined threshold is a statistically significant difference between the actual cruising speed and the predicted cruising speed based on the probabilistic model associated with the identified driver.

### Claim 7 (depends on 4)

The controller according to claim 4, further comprising:
a vehicle identification unit configured to identify the vehicle and associate the generated probabilistic model with the vehicle; and
a transmitter configured to transmit a message in response to a determination that the actual cruising speed is different by a predetermined threshold from a predicted cruising speed for the target segment based on the probabilistic model associated with the vehicle.

### Claim 8 (depends on 4)

The controller according to claim 4, wherein the updating unit is further configured to identify outlier feature data corresponding to unusual traffic patterns or unusual driving patterns and remove the identified outlier feature data from the real-time predictive feature data used to update the probabilistic model.

### Claim 9 (depends on 2)

The controller according to claim 2, wherein the predicting unit is configured to predict the cruising speed of the vehicle for the target segment by conditioning the probabilistic model on a sequence number of a trajectory including the target segment relative to sequential trajectories forming a navigation path of the vehicle.

### Claim 10 (depends on 2)

The controller according to claim 2, wherein the predicting unit is configured to predict the cruising speed of the vehicle for the target segment by conditioning the probabilistic model on a cruising speed of a cruising segment of the current trajectory, the current trajectory being a trajectory prior to a trajectory including the target segment.

### Claim 11 (depends on 2)

The controller according to claim 2, wherein the predicting unit is configured to predict the cruising speed of the vehicle for the target segment by conditioning the probabilistic model on a deceleration model of a last segment of the current trajectory, the current trajectory being a trajectory prior to a trajectory including the target segment.

### Claim 12 (depends on 2)

The controller according to claim 2, wherein, when an actual cruising speed for the target segment is reached by the vehicle, the controller designates an upcoming cruising segment as a next target segment for cruising speed prediction.

### Claim 13 (depends on 10)

The controller according to claim 10, wherein the deceleration model is a time-based polynomial model, a state-space-based linear model, or a state-space-based non-linear model.

### Claim 14 (depends on 2)

The controller according to claim 2, wherein the launch model is a time-based polynomial model, a state-space-based linear model, or a state-space-based non-linear model.

### Claim 15 (depends on 1)

The controller according to claim 1, wherein
the vehicle includes a Global Positioning System (GPS) and a map database,
the feature data includes additional features, including environmental features, road geometry features, road condition features, and a road speed limit, and
the predicted cruising speed is based on a speed limit of the target segment.

### Claim 16 (depends on 15)

The controller according to claim 15, wherein
the controller is configured to receive image data from a camera mounted to the vehicle, and
the controller extracts the speed limit from speed limit signs captured by the camera.

### Claim 17 (depends on 15)

The controller according to claim 15, wherein the additional features include one or more of weather, road surface conditions, time of day, light level, atmospheric temperature, road inclination, radius of road curvature, and road composition.

### Claim 18 (depends on 1)

The controller according to claim 1, wherein
the vehicle includes a Global Positioning System (GPS), and
the model generator generates the probabilistic model for each new road segment driven by the vehicle and associates the probabilistic model with the respective road segment.

### Claim 19 (depends on 1)

The controller according to claim 1, wherein the predicted cruising speed is a probability density function.

### Claim 20 (depends on 18)

The controller according to claim 18, further comprising a warning unit configured to integrate the probability density function to determine a cumulative distribution function, determine a cruising speed value at a user-selected certainty threshold of the cumulative distribution function, and issue a warning in response to the actual cruising speed exceeding the determined cruising speed value.

### Claim 21 (depends on 1)

The controller according to claim 1, wherein the predicted cruising speed is a value.

### Claim 22 (depends on 1)

The controller according to claim 1, further comprising an eco-acceleration cruise controller that receives the predicted cruising speed and determines an acceleration sequence to achieve the predicted cruising speed and maximize fuel efficiency.

### Claim 23 (depends on 1)

The controller according to claim 1, wherein the model generator performs kernel density estimation, least absolute shrinkage and selection operator (LASSO) linear regression, or boosted tree regression to generate the probabilistic model.

### Claim 24 (depends on 1)

The controller according to claim 1, wherein the predicting unit is further configured to predict the cruising speed for the target segment as a mean of actual cruising speeds of prior segments when no feature data has been extracted in the current trajectory.

### Claim 25 (depends on 1)

The controller according to claim 1, wherein the controller outputs the predictive cruising speed to a client application.

### Claim 26 (depends on 1)

A vehicle including the controller according to claim 1.

### Claim 27

A method of predicting cruising speeds of a vehicle, the method comprising:
extracting feature data from segments of a prior trajectory of the vehicle, the feature data including cruising speeds of the vehicle and predictive feature data;
generating a probabilistic model associating the predictive feature data with the cruising speeds of the vehicle; and
predicting a cruising speed of the vehicle for a target segment, which is an upcoming cruising segment of the vehicle, by conditioning the probabilistic model on real-time predictive feature data of segments of a current trajectory.

### Claim 28

A computer readable storage medium storing computer readable instructions thereon which, when executed by a computer, cause the computer to perform a method of predicting cruising speeds of a vehicle, the method comprising:
extracting feature data from segments of a prior trajectory of the vehicle, the feature data including cruising speeds of the vehicle and predictive feature data;
generating a probabilistic model associating the predictive feature data with the cruising speeds of the vehicle; and
predicting a cruising speed of the vehicle for a target segment, which is an upcoming cruising segment of the vehicle, by conditioning the probabilistic model on real-time predictive feature data of segments of a current trajectory.

### Claim 29

A cruising speed prediction system comprising:
a controller aboard a vehicle, the controller being configured to collect trajectory data of the vehicle; and
a remote processing server configured to predict cruising speeds of the vehicle based on the trajectory data of the vehicle, the remote processing server including:
an extracting unit configured to extract feature data from segments of a prior trajectory of the vehicle, the feature data including cruising speeds of the vehicle and predictive feature data;
a model generator configured to generate a probabilistic model associating the predictive feature data with the cruising speeds of the vehicle; and
a predicting unit configured to predict a cruising speed of the vehicle for a target segment, which is an upcoming cruising segment of the vehicle, by conditioning the probabilistic model on real-time predictive feature data of segments of a current trajectory, wherein
the controller transmits the collected trajectory data to the remote processing server and receives from the remote processing server the predicted cruising speed of the vehicle for the target segment.
