Granted Patent B2

US 12,546,615 · App. 17/686,694 · Granted Feb 10, 2026

Inventors: Theobolt N. Leung (San Francisco, CA); Eric Dahl (Newman Lake, WA); Kenneth Jason Sanchez (San Francisco, CA)

Assignee: QUANATA, LLC

G01C21/3469B60W40/09B60W50/10G01C21/3484G01C21/3492G01C21/3617G06Q30/018G06Q30/0201G06Q30/0224G07C5/04B60W2510/0638B60W2530/209B60W2540/30G06Q50/06G06Q50/40

Abstract

Claims (103)

1 . A method for predicting fuel consumption efficiency, the method comprising:

collecting, by a computing device and from one or more sensors, past user driving data for one or more past vehicle trips made by a user, wherein the one or more sensors include at least one of a GPS sensor or an accelerometer, and wherein the past user driving data includes at least one of acceleration data, braking data, or location data;

analyzing, by the computing device, historical trip data including past commute patterns and vehicle route data to:

determine one or more routes for one or more future vehicle trips to be made by the user during a predetermined future period of time; and

modify the one or more routes based on receiving one or more user inputs to update information for the one or more future vehicle trips;

training an artificial neural network (ANN) configured as a convolutional neural network (CNN) or a recurrent neural network (RNN) for analyzing driving features and predicting fuel consumption efficiencies, wherein the training the ANN comprises:

analyzing, by the ANN, training driving data included in one or more sets of training data to determine one or more training driving features associated with past fuel consumption, wherein the analyzing the training driving data includes performing at least one of feature extraction or pattern recognition to identify patterns in the at least one of the acceleration data, the braking data, or the location data, and wherein the patterns include at least one of braking patterns, acceleration patterns, or cornering patterns;

generating, by the ANN, using weight values associated with layers of the ANN, an estimated past efficiency value related to the past fuel consumption based at least in part upon the one or more training driving features determined from the training driving data included in the one or more sets of training data;

comparing, by the ANN, the estimated past efficiency value with an actual past efficiency value to determine an accuracy of the estimated past efficiency value using one or more of a loss function or a cost function; and

adjusting, by the ANN, one or more parameters of the ANN including the weight values associated with the layers of the ANN, based at least in part upon the estimated past efficiency value with the actual past efficiency value, as compared, to reduce one or more magnitudes of one or more outputs of the one or more of the loss function or the cost function until the loss function or the cost function is minimized for the one or more sets of training data;

analyzing, by the computing device using the ANN, as trained, the past user driving data to determine one or more past user driving features, wherein the one or more past user driving features are related to a past fuel consumption efficiency of the user, wherein the past fuel consumption efficiency of the user indicates how fuel was consumed by the user during the one or more past vehicle trips given the one or more past user driving features, and wherein the one or more past user driving features include one or more of braking, accelerating, cornering, speeding, lane changing, tailgating, idling, or timing of gear shifting;

predicting, by the computing device using the ANN, as trained, a future fuel consumption efficiency of the user during the predetermined future period of time based at least in part upon the information for the one or more future vehicle trips and the one or more past user driving features related to the past fuel consumption efficiency of the user during the one or more past vehicle trips given the one or more past user driving features; and

outputting, by the computing device and to a client device for display on the client device, the future fuel consumption efficiency, as predicted, for the one or more routes, as determined and modified, for the one or more future vehicle trips.

2 . The method of claim 1 , wherein the analyzing, by the computing device using the ANN, as trained, the past user driving data to determine the one or more past user driving features includes:

providing the past user driving data;

processing information associated with the past user driving data; and

determining the one or more past user driving features based at least in part upon the past user driving data.

3 . The method of claim 2 , wherein the predicting, by the computing device using the ANN, as trained, the future fuel consumption efficiency of the user during the predetermined future period of time includes:

providing the information for the one or more future vehicle trips;

processing the information for the one or more future vehicle trips and information associated with the one or more past user driving features; and

determining the future fuel consumption efficiency of the user during the predetermined future period of time based at least in part upon the information for the one or more future vehicle trips and the one or more past user driving features.

4 . The method of claim 1 , wherein the one or more past user driving features include:

one or more first past user driving features that increase the past fuel consumption efficiency of the user; and

one or more second past user driving features that decrease the past fuel consumption efficiency of the user;

wherein:

the one or more first past user driving features correspond to one or more first importance levels respectively for increasing the past fuel consumption efficiency of the user; and

the one or more second past user driving features correspond to one or more second importance levels respectively for decreasing the past fuel consumption efficiency of the user, wherein the one or more first importance levels and the one or more second importance levels are determined by the ANN.

5 . The method of claim 1 , wherein the information for the one or more future vehicle trips includes:

vehicle information for the one or more future vehicle trips;

distance information for the one or more future vehicle trips; and

congestion information for the one or more future vehicle trips.

6 . A computing device for predicting fuel consumption efficiency, the computing device comprising:

one or more processors; and

a memory storing instructions that, when executed by the one or more processors, cause the one or more processors to:

collect, from one or more sensors, past user driving data for one or more past vehicle trips made by a user, wherein the one or more sensors include at least one of a GPS sensor or an accelerometer, and wherein the past user driving data includes at least one of acceleration data, braking data, or location data;

analyze historical trip data including past commute patterns and vehicle route data to:

determine one or more routes for one or more future vehicle trips to be made by a user during a predetermined future period of time; and

modify the one or more routes based on receiving one or more user inputs to update information for the one or more future vehicle trips;

train an artificial neural network (ANN) configured as a convolutional neural network (CNN) or a recurrent neural network (RNN) for analyzing driving features and predicting fuel consumption efficiencies, comprising:

analyze, by the ANN, training driving data included in one or more sets of training data to determine one or more training driving features associated with past fuel consumption, wherein analyzing the training driving data includes performing at least one of feature extraction or pattern recognition to identify patterns in the at least one of the acceleration data, the braking data, or the location data, and wherein the patterns include at least one of braking patterns, acceleration patterns, or cornering patterns;

generate, by the ANN, using weight values associated with layers of the ANN, an estimated past efficiency value related to the past fuel consumption based at least in part upon the one or more training driving features determined from the training driving data included in the one or more sets of training data;

compare, by the ANN, the estimated past efficiency value with an actual past efficiency value to determine an accuracy of the estimated past efficiency value using one or more of a loss function or a cost function; and

adjust, by the ANN, one or more parameters of the ANN including the weight values associated with the layers of the ANN, based at least in part upon the estimated past efficiency value with the actual past efficiency value, as compared, to reduce one or more magnitudes of one or more outputs of the one or more of the loss function or the cost function until the loss function or the cost function is minimized for the one or more sets of training data;

analyze, using the ANN, as trained, the past user driving data to determine one or more past user driving features, wherein the one or more past user driving features are related to a past fuel consumption efficiency of the user, wherein the past fuel consumption efficiency of the user indicates how fuel was consumed by the user during the one or more past vehicle trips given the one or more past user driving features, and wherein the one or more past user driving features include one or more of braking, accelerating, comering, speeding, lane changing, tailgating, idling, or timing of gear shifting;

predict, using the ANN, as trained, a future fuel consumption efficiency of the user during the predetermined future period of time based at least in part upon the information for the one or more future vehicle trips and the one or more past user driving features related to the past fuel consumption efficiency of the user during the one or more past vehicle trips given the one or more past user driving features; and

output to a client device for display on the client device, the future fuel consumption efficiency, as predicted, for the one or more routes, as determined and modified, for the one or more future vehicle trips.

7 . The computing device of claim 6 , wherein the instructions that cause the one or more processors to analyze, using the ANN, as trained, the past user driving data to determine the one or more past user driving features further comprise instructions that cause the one or more processors to:

provide the past user driving data;

process information associated with the past user driving data; and

determine the one or more past user driving features based at least in part upon the past user driving data.

8 . The computing device of claim 7 , wherein the instructions that cause the one or more processors to predict, using the ANN, as trained, the future fuel consumption efficiency of the user during the predetermined future period of time further comprise instructions that cause the one or more processors to:

provide the information for the one or more future vehicle trips;

process the information for the one or more future vehicle trips and information associated with the one or more past user driving features; and

determine the future fuel consumption efficiency of the user during the predetermined future period of time based at least in part upon the information for the one or more future vehicle trips and the one or more past user driving features.

9 . The computing device of claim 6 , wherein the one or more past user driving features include:

one or more first past user driving features that increase the past fuel consumption efficiency of the user; and

one or more second past user driving features that decrease the past fuel consumption efficiency of the user;

wherein:

the one or more first past user driving features correspond to one or more first importance levels respectively for increasing the past fuel consumption efficiency of the user; and

the one or more second past user driving features correspond to one or more second importance levels respectively for decreasing the past fuel consumption efficiency of the user, wherein the one or more first importance levels and the one or more second importance levels are determined by the ANN.

10 . The computing device of claim 6 , wherein the information for the one or more future vehicle trips includes:

vehicle information for the one or more future vehicle trips;

distance information for the one or more future vehicle trips; and

congestion information for the one or more future vehicle trips.

11 . A non-transitory computer-readable medium storing instructions for predicting fuel consumption efficiency, the instructions when executed by one or more processors of a computing device cause the computing device to:

collect, from one or more sensors, past user driving data for one or more past vehicle trips made by a user, wherein the one or more sensors include at least one of a GPS sensor or an accelerometer, and wherein the past user driving data includes at least one of acceleration data, braking data, or location data;

analyze historical trip data including past commute patterns and vehicle route data to:

determine one or more routes for one or more future vehicle trips to be made by a user during a predetermined future period of time; and

modify the one or more routes based on receiving one or more user inputs to update information for the one or more future vehicle trips;

use an artificial neural network (ANN) configured as a convolutional neural network (CNN) or a recurrent neural network (RNN), as trained, for analyzing driving features and predicting fuel consumption efficiencies, wherein training the ANN comprises steps to:

analyze, by the ANN, training driving data included in one or more sets of training data to determine one or more training driving features associated with past fuel consumption, wherein analyzing the training driving data includes performing at least one of feature extraction or pattern recognition to identify patterns in the at least one of the acceleration data, the braking data, or the location data, and wherein the patterns include at least one of braking patterns, acceleration patterns, or cornering patterns;

generate, by the ANN, using weight values associated with layers of the ANN, an estimated past efficiency value related to the past fuel consumption based at least in part upon the one or more training driving features determined from the training driving data included in the one or more sets of training data;

compare, by the ANN, the estimated past efficiency value with an actual past efficiency value to determine an accuracy of the estimated past efficiency value using one or more of a loss function or a cost function; and

adjust, by the ANN, one or more parameters of the ANN including the weight values associated with the layers of the ANN, based at least in part upon the estimated past efficiency value with the actual past efficiency value, as compared, to reduce one or more magnitudes of one or more outputs of the one or more of the loss function or the cost function until the loss function or the cost function is minimized for the one or more sets of training data;

analyze, using the ANN, as trained, the past user driving data to determine one or more past user driving features, wherein the one or more past user driving features are related to a past fuel consumption efficiency of the user, wherein the past fuel consumption efficiency of the user indicates how fuel was consumed by the user during the one or more past vehicle trips given the one or more past user driving features, and wherein the one or more past user driving features include one or more of braking, accelerating, comering, speeding, lane changing, tailgating, idling, or timing of gear shifting;

predict, using the ANN, as trained, a future fuel consumption efficiency of the user during the predetermined future period of time based at least in part upon the information for the one or more future vehicle trips and the one or more past user driving features related to the past fuel consumption efficiency of the user during the one or more past vehicle trips given the one or more past user driving features; and

output, to a client device for display on the client device, the future fuel consumption efficiency, as predicted, for the one or more routes, as determined and modified, for the one or more future vehicle trips.

12 . The non-transitory computer-readable medium storing the instructions for predicting the fuel consumption efficiency of claim 11 , wherein the instructions when executed by the one or more processors that cause the computing device to analyze, using the ANN, as trained, the past user driving data to determine the one or more past user driving features further cause the computing device to:

provide the past user driving data;

process information associated with the past user driving data; and

determine the one or more past user driving features based at least in part upon the past user driving data.

13 . The non-transitory computer-readable medium storing the instructions for predicting the fuel consumption efficiency of claim 12 , wherein the instructions when executed by the one or more processors that cause the computing device to predict, using the ANN, as trained, the future fuel consumption efficiency of the user during the predetermined future period of time further cause the computing device to:

provide the information for the one or more future vehicle trips;

process the information for the one or more future vehicle trips and information associated with the one or more past user driving features; and

determine the future fuel consumption efficiency of the user during the predetermined future period of time based at least in part upon the information for the one or more future vehicle trips and the one or more past user driving features.

14 . The non-transitory computer-readable medium of claim 11 , wherein the one or more sensors further comprise one or more of a gyroscope, a magnetometer, a speedometer, a steering-angle sensor, a brake sensor, a yaw-rate sensor, or a proximity detector.

15 . The non-transitory computer-readable medium of claim 11 , wherein collecting the past user driving data is performed in response to a triggering event comprising that each sensor of the one or more sensors acquires a threshold amount of sensor measurements.

16 . The non-transitory computer-readable medium of claim 11 , wherein updating the information for the one or more future vehicle trips further comprises obtaining congestion information derived from analysis of historical traffic data provided by a third party.

17 . The non-transitory computer-readable medium of claim 11 , wherein the computing device comprises the client device embedded in or connected to a vehicle and includes a display unit and the one or more sensors.

18 . The non-transitory computer-readable medium of claim 11 , wherein the analyzing, using the ANN, as trained, the past user driving data to determine the one or more past user driving features includes:

providing the past user driving data;

processing information associated with the past user driving data; and

determining the one or more past user driving features based at least in part upon the past user driving data.

19 . The non-transitory computer-readable medium of claim 11 , wherein the one or more past user driving features include:

one or more first past user driving features that increase the past fuel consumption efficiency of the user; and

one or more second past user driving features that decrease the past fuel consumption efficiency of the user;

wherein:

the one or more first past user driving features correspond to one or more first importance levels respectively for increasing the past fuel consumption efficiency of the user; and

the one or more second past user driving features correspond to one or more second importance levels respectively for decreasing the past fuel consumption efficiency of the user, wherein the one or more first importance levels and the one or more second importance levels are determined by the ANN.

20 . The non-transitory computer-readable medium of claim 11 , wherein the information for the one or more future vehicle trips includes:

vehicle information for the one or more future vehicle trips;

distance information for the one or more future vehicle trips; and

congestion information for the one or more future vehicle trips.

Assignments (2)

Continuity (3)

References Cited (62)