Estimating speed profiles
Among other things, we describe techniques for estimating a speed profile for a proposed trajectory for a vehicle and operating the vehicle along the proposed trajectory according to the speed profile, including a method for: obtaining, by a planning circuit on a vehicle, a proposed trajectory for the vehicle in response to a driving scenario; obtaining, by the planning circuit, an estimated speed profile, and a confidence score, wherein the confidence score represents a similarity of the estimated speed profile to an actual speed profile that would be generated by a control circuit for the proposed trajectory; determining whether the confidence score meets a confidence threshold; and in accordance with a determination that the confidence score exceeds the confidence threshold, operating, by a control circuit on the vehicle, the vehicle along the proposed trajectory.
1 . A method comprising:
configuring a plurality of weights for a machine learning model to estimate a speed profile for a vehicle in a driving environment, wherein at least one weight of the plurality of weights is based on a past trajectory of the vehicle and a speed profile corresponding to the past trajectory;
providing a proposed trajectory for the vehicle in the driving environment as an input to the configured machine learning model;
in response to providing the proposed trajectory as the input to the machine learning model, obtaining an output of the machine learning model comprising an estimated speed profile corresponding to the proposed trajectory; and
determining a driving behavior of the vehicle in the driving environment using the estimated speed profile.
2 . The method of claim 1 , wherein the weights of the machine learning model are further based on one or more of:
comfort metrics of passengers of the vehicle,
information about objects proximate to the vehicle,
features corresponding to physical characteristics of the vehicle, or
motion characteristics of the vehicle.
3 . The method of claim 1 , further comprising:
determining the plurality of weights or a corresponding plurality of speed profiles based at least on past driving behavior of the vehicle in different driving scenarios.
4 . The method of claim 1 , wherein the output of the machine learning model further comprises a confidence score corresponding to the estimated speed profile, the confidence score indicating a similarity of the estimated speed profile to an actual speed profile for the proposed trajectory, the method further comprising:
determining the driving behavior of the vehicle in the driving environment using the estimated speed profile and the confidence score.
5 . The method of claim 4 , wherein determining the driving behavior of the vehicle in the driving environment using the estimated speed profile and the confidence score further comprises:
determining whether the confidence score meets a confidence threshold; and
in accordance with a determination that the confidence score meets or exceeds the confidence threshold, operating the vehicle along the proposed trajectory.
6 . The method of claim 4 , wherein determining the driving behavior of the vehicle in the driving environment using the estimated speed profile and the confidence score further comprises:
determining whether the confidence score meets a confidence threshold; and
in accordance with a determination that the confidence score does not meet the confidence threshold:
obtaining a second trajectory based on a predetermined speed profile heuristic, and
operating the vehicle along the second trajectory.
7 . The method of claim 4 , further comprising:
determining a variance between the estimated speed profile and the actual speed profile;
upon determining that the variance is high, assigning a low value to the confidence score; and
upon determining that the variance is low, assigning a high value to the confidence score.
8 . The method of claim 4 , wherein the estimated speed profile for the proposed trajectory is based on:
comparing the driving environment corresponding to the proposed trajectory to a plurality of driving scenarios in memory;
in response to the comparing, identifying a particular driving scenario of the plurality of driving scenarios that is similar to the driving environment within a threshold value; and
upon identifying the particular driving scenario, assigning a speed profile associated with the particular driving scenario as the estimated speed profile for the proposed trajectory.
9 . An apparatus comprising:
one or more processors; and
machine-readable memory storing instructions that, when executed, are configured to cause the one or more processors to perform operations comprising:
configuring a plurality of weights for a machine learning model to estimate a speed profile for a vehicle in a driving environment, wherein at least one weight of the plurality of weights is based on a past trajectory of the vehicle and a speed profile corresponding to the past trajectory;
providing a proposed trajectory for the vehicle in the driving environment as an input to the configured machine learning model;
in response to providing the proposed trajectory as the input to the machine learning model, obtaining an output of the machine learning model comprising an estimated speed profile corresponding to the proposed trajectory; and
determining a driving behavior of the vehicle in the driving environment using the estimated speed profile.
10 . The apparatus of claim 9 , wherein the weights of the machine learning model are further based on one or more of:
comfort metrics of passengers of the vehicle,
information about objects proximate to the vehicle,
features corresponding to physical characteristics of the vehicle, or
motion characteristics of the vehicle.
11 . The apparatus of claim 9 , the operations further comprising:
determining the plurality of weights or a corresponding plurality of speed profiles based at least on past driving behavior of the vehicle in different driving scenarios.
12 . The apparatus of claim 9 , wherein the output of the machine learning model further comprises a confidence score corresponding to the estimated speed profile, the confidence score indicating a similarity of the estimated speed profile to an actual speed profile for the proposed trajectory, the operations further comprising:
determining the driving behavior of the vehicle in the driving environment using the estimated speed profile and the confidence score.
13 . The apparatus of claim 12 , wherein determining the driving behavior of the vehicle in the driving environment using the estimated speed profile and the confidence score further comprises:
determining whether the confidence score meets a confidence threshold; and
in accordance with a determination that the confidence score meets or exceeds the confidence threshold, operating the vehicle along the proposed trajectory.
14 . The apparatus of claim 12 , wherein determining the driving behavior of the vehicle in the driving environment using the estimated speed profile and the confidence score further comprises:
determining whether the confidence score meets a confidence threshold; and
in accordance with a determination that the confidence score does not meet the confidence threshold:
obtaining a second trajectory based on a predetermined speed profile heuristic, and
operating the vehicle along the second trajectory.
15 . The apparatus of claim 12 , the operations further comprising:
determining a variance between the estimated speed profile and the actual speed profile;
upon determining that the variance is high, assigning a low value to the confidence score; and
upon determining that the variance is low, assigning a high value to the confidence score.
16 . The apparatus of claim 12 , wherein the estimated speed profile for the proposed trajectory is based on:
comparing the driving environment corresponding to the proposed trajectory to a plurality of driving scenarios in memory;
in response to the comparing, identifying a particular driving scenario of the plurality of driving scenarios that is similar to the driving environment within a threshold value; and
upon identifying the particular driving scenario, assigning a speed profile associated with the particular driving scenario as the estimated speed profile for the proposed trajectory.
17 . One or more non-transitory machine-readable media storing instructions that, when executed, are configured to cause one or more processors to perform operations comprising:
configuring a plurality of weights for a machine learning model to estimate a speed profile for a vehicle in a driving environment, wherein at least one weight of the plurality of weights is based on a past trajectory of the vehicle and a speed profile corresponding to the past trajectory;
providing a proposed trajectory for the vehicle in the driving environment as an input to the configured machine learning model;
in response to providing the proposed trajectory as the input to the machine learning model, obtaining an output of the machine learning model comprising an estimated speed profile corresponding to the proposed trajectory; and
determining a driving behavior of the vehicle in the driving environment using the estimated speed profile.
18 . The one or more non-transitory machine-readable media of claim 17 , wherein the weights of the machine learning model are further based on one or more of:
comfort metrics of passengers of the vehicle,
information about objects proximate to the vehicle,
features corresponding to physical characteristics of the vehicle, or
motion characteristics of the vehicle.
19 . The one or more non-transitory machine-readable media of claim 17 , wherein the output of the machine learning model further comprises a confidence score corresponding to the estimated speed profile, the confidence score indicating a similarity of the estimated speed profile to an actual speed profile for the proposed trajectory, the operations further comprising:
determining whether the confidence score meets a confidence threshold; and
in accordance with a determination that the confidence score meets or exceeds the confidence threshold, operating the vehicle along the proposed trajectory; and
in accordance with a determination that the confidence score does not meet the confidence threshold:
obtaining a second trajectory based on a predetermined speed profile heuristic, and
operating the vehicle along the second trajectory.
20 . The one or more non-transitory machine-readable media of claim 19 , the operations further comprising:
determining a variance between the estimated speed profile and the actual speed profile;
upon determining that the variance is high, assigning a low value to the confidence score; and
upon determining that the variance is low, assigning a high value to the confidence score.
21 . The one or more non-transitory machine-readable media of claim 19 , wherein the estimated speed profile for the proposed trajectory is based on:
comparing the driving environment corresponding to the proposed trajectory to a plurality of driving scenarios in memory;
in response to the comparing, identifying a particular driving scenario of the plurality of driving scenarios that is similar to the driving environment within a threshold value; and
upon identifying the particular driving scenario, assigning a speed profile associated with the particular driving scenario as the estimated speed profile for the proposed trajectory.