# US 12,214,787 B2 · Estimating speed profiles

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

- Number: US 12,214,787 B2
- Document ID: US12214787B2
- Status: Granted
- Kind: B2
- Application number: 18/388,878
- Filed: 2023-11-13
- Published: 2024-03-07
- Granted: 2025-02-04
- Earliest claimed: 2019-05-29
- Assignees: Motional AD LLC
- Applicants: Motional AD LLC
- Inventors: Francesco Seccamonte; Kostyantyn Slutskyy
- Web page: https://patentgrep.com/patent/US12214787B2

## CPC

- B60W 30/143: {Speed control (B60W30/16 takes precedence)}
- G01S 17/931: of land vehicles
- B60W 40/105: Speed
- B60W 50/082: {Selecting or switching between different modes of propelling}
- B60W 50/085: {Changing the parameters of the control units, e.g. changing limit values, working points by control input}
- B60W 60/0011: {involving control alternatives for a single driving scenario, e.g. planning several paths to avoid obstacles}
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- B60W 2520/10: Longitudinal speed
- B60W 2556/20: Data confidence level
- B60W 2720/103: Speed profile
- B60W 2720/24: Direction of travel
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G06N 20/20: Ensemble learning
- 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
- G01S: RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

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.

## Claims

### Claim 1

A computer-implemented method, comprising:
obtaining, by one or more processors in a vehicle, a proposed trajectory for the vehicle in response to a driving scenario;
obtaining, by the one or more processors, a speed profile corresponding to the proposed trajectory for the vehicle from a machine learning network that is trained to generate one or more speed profiles, the machine learning network trained using training vectors corresponding to a plurality of trajectories and a plurality of corresponding speed profiles that are determined based on data received from a plurality of distinct perception pipelines associated with the vehicle;
determining, by the one or more processors, that the obtained speed profile satisfies one or more criteria for driving the vehicle; and
in response to determining that the obtained speed profile satisfies one or more criteria for driving the vehicle, causing operation of the vehicle along the proposed trajectory.

### Claim 2 (depends on 1)

The method of claim 1, wherein obtaining the speed profile from the machine learning network comprises:
providing data corresponding to the driving scenario and the proposed trajectory for the vehicle as an input to the machine learning network; and
obtaining the speed profile of the vehicle and a corresponding confidence score as an output of the machine learning network.

### Claim 3 (depends on 2)

The method of claim 2, wherein determining that the obtained speed profile satisfies one or more criteria for driving the vehicle comprises:
comparing the confidence score corresponding to the speed profile to a specified confidence threshold; and
in response to the comparing, determining that the confidence score is equal to or greater than the specified confidence threshold.

### Claim 4 (depends on 2)

The method of claim 2, wherein the confidence score represents a similarity of the speed profile of the vehicle generated by the machine learning network to an actual speed profile of the vehicle that would be generated by a control circuit of the vehicle for the proposed trajectory for the vehicle.

### Claim 5 (depends on 1)

The method of claim 1, wherein the machine learning network is included in a control circuit of the vehicle that is coupled to the one or more processors, the method further comprising:
receiving, by the control circuit, data corresponding to the proposed trajectory from the one or more processors; and
in response to receiving the data, controlling, by the control circuit, the machine learning network to generate the speed profile using a trained model, wherein the trained model represents a plurality of associations between a plurality of trajectories and a plurality of speed profiles, and wherein each speed profile in the plurality of speed profiles is generated by the control circuit from a respective trajectory in the plurality of trajectories.

### Claim 6 (depends on 5)

The method of claim 5, further comprising:
obtaining, by the control circuit, a set of principal components from a principal component analysis of a plurality of speed profiles; and
controlling, by the control circuit, the machine learning network to generate a confidence score as a variance between the set of principal components and the speed profile.

### Claim 7 (depends on 5)

The method of claim 5, wherein the one or more processors obtain the proposed trajectory using data from a first perception pipeline, and wherein
the trained model is trained on (i) a second plurality of speed profiles generated using other data from the first perception pipeline, and (ii) data from a second perception pipeline that is unavailable to the first perception pipeline.

### Claim 8 (depends on 1)

The method of claim 1, further comprising:
obtaining a new proposed trajectory for the vehicle in response to a new driving scenario;
obtaining a new speed profile corresponding to the new proposed trajectory from the machine learning network;
determining that the new speed profile does not satisfy the one or more criteria for driving the vehicle;
in response to determining that the obtained speed profile does not satisfy the one or more criteria for driving the vehicle, obtaining a second trajectory based on a predetermined speed profile heuristic; and
causing operation of the vehicle along the second trajectory.

### Claim 9 (depends on 1)

The method of claim 1, wherein the machine learning network is further trained using another plurality of trajectories and corresponding speed profiles that are determined based on past driving behavior of the vehicle in different driving scenarios.

### Claim 10

A system 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:
obtaining a proposed trajectory for a vehicle in response to a driving scenario;
obtaining a speed profile corresponding to the proposed trajectory for the vehicle from a machine learning network that is trained to generate one or more speed profiles, the machine learning network trained using training vectors corresponding to a plurality of trajectories and a plurality of corresponding speed profiles that are determined based on data received from a plurality of distinct perception pipelines associated with the vehicle;
determining that the obtained speed profile satisfies one or more criteria for driving the vehicle; and
in response to determining that the obtained speed profile satisfies one or more criteria for driving the vehicle, causing operation of the vehicle along the proposed trajectory.

### Claim 11 (depends on 10)

The system of claim 10, wherein obtaining the speed profile from the machine learning network comprises:
providing data corresponding to the driving scenario and the proposed trajectory for the vehicle as an input to the machine learning network; and
obtaining the speed profile of the vehicle and a corresponding confidence score as an output of the machine learning network.

### Claim 12 (depends on 11)

The system of claim 11, wherein determining that the obtained speed profile satisfies one or more criteria for driving the vehicle comprises:
comparing the confidence score corresponding to the speed profile to a specified confidence threshold; and
in response to the comparing, determining that the confidence score is equal to or greater than the specified confidence threshold.

### Claim 13 (depends on 11)

The system of claim 11, wherein the confidence score represents a similarity of the speed profile of the vehicle generated by the machine learning network to an actual speed profile of the vehicle that would be generated by a control circuit of the vehicle for the proposed trajectory for the vehicle.

### Claim 14 (depends on 10)

The system of claim 10, further comprising a control circuit of the vehicle that includes the machine learning network and that is coupled to the one or more processors, the operations further comprising:
receiving, by the control circuit, data corresponding to the proposed trajectory from the one or more processors; and
in response to receiving the data, controlling, by the control circuit, the machine learning network to generate the speed profile using a trained model, wherein the trained model represents a plurality of associations between a plurality of trajectories and a plurality of speed profiles, and wherein each speed profile in the plurality of speed profiles is generated by the control circuit from a respective trajectory in the plurality of trajectories.

### Claim 15 (depends on 14)

The system of claim 14, wherein the operations further comprise:
obtaining, by the control circuit, a set of principal components from a principal component analysis of a plurality of speed profiles; and
controlling, by the control circuit, the machine learning network to generate a confidence score as a variance between the set of principal components and the speed profile.

### Claim 16 (depends on 14)

The system of claim 14, wherein the one or more processors obtain the proposed trajectory using data from a first perception pipeline, and wherein
the trained model is trained on (i) a second plurality of speed profiles generated using other data from the first perception pipeline, and (ii) data from a second perception pipeline that is unavailable to the first perception pipeline.

### Claim 17 (depends on 10)

The system of claim 10, wherein the operations further comprise:
obtaining a new proposed trajectory for the vehicle in response to a new driving scenario;
obtaining a new speed profile corresponding to the new proposed trajectory from the machine learning network;
determining that the new speed profile does not satisfy the one or more criteria for driving the vehicle;
in response to determining that the obtained speed profile does not satisfy the one or more criteria for driving the vehicle, obtaining a second trajectory based on a predetermined speed profile heuristic; and
causing operation of the vehicle along the second trajectory.

### Claim 18 (depends on 10)

The system of claim 10, wherein the machine learning network is further trained using another plurality of trajectories and corresponding speed profiles that are determined based on past driving behavior of the vehicle in different driving scenarios.

### Claim 19

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:
obtaining a proposed trajectory for a vehicle in response to a driving scenario;
obtaining a speed profile corresponding to the proposed trajectory for the vehicle from a machine learning network that is trained to generate one or more speed profiles, the machine learning network trained using training vectors corresponding to a plurality of trajectories and a plurality of corresponding speed profiles that are determined based on data received from a plurality of distinct perception pipelines associated with the vehicle;
determining that the obtained speed profile satisfies one or more criteria for driving the vehicle; and
in response to determining that the obtained speed profile satisfies one or more criteria for driving the vehicle, causing operation of the vehicle along the proposed trajectory.

### Claim 20 (depends on 19)

The one or more non-transitory machine-readable media of claim 19, wherein obtaining the speed profile from the machine learning network comprises:
providing data corresponding to the driving scenario and the proposed trajectory for the vehicle as an input to the machine learning network; and
obtaining the speed profile of the vehicle and a corresponding confidence score as an output of the machine learning network.

### Claim 21 (depends on 20)

The one or more non-transitory machine-readable media of claim 20, wherein determining that the obtained speed profile satisfies one or more criteria for driving the vehicle comprises:
comparing the confidence score corresponding to the speed profile to a specified confidence threshold; and
in response to the comparing, determining that the confidence score is equal to or greater than the specified confidence threshold.

### Claim 22 (depends on 20)

The one or more non-transitory machine-readable media of claim 20, wherein the confidence score represents a similarity of the speed profile of the vehicle generated by the machine learning network to an actual speed profile of the vehicle that would be generated by a control circuit of the vehicle for the proposed trajectory for the vehicle.

### Claim 23 (depends on 19)

The one or more non-transitory machine-readable media of claim 19, wherein the machine learning network is included in a control circuit of the vehicle that is coupled to the one or more processors, the operations further comprising:
receiving, by the control circuit, data corresponding to the proposed trajectory from the one or more processors; and
in response to receiving the data, controlling, by the control circuit, the machine learning network to generate the speed profile using a trained model, wherein the trained model represents a plurality of associations between a plurality of trajectories and a plurality of speed profiles, and wherein each speed profile in the plurality of speed profiles is generated by the control circuit from a respective trajectory in the plurality of trajectories.

### Claim 24 (depends on 23)

The one or more non-transitory machine-readable media of claim 23, wherein the operations further comprise:
obtaining, by the control circuit, a set of principal components from a principal component analysis of a plurality of speed profiles; and
controlling, by the control circuit, the machine learning network to generate a confidence score as a variance between the set of principal components and the speed profile.

### Claim 25 (depends on 23)

The one or more non-transitory machine-readable media of claim 23, wherein the one or more processors obtain the proposed trajectory using data from a first perception pipeline, and wherein
the trained model is trained on (i) a second plurality of speed profiles generated using other data from the first perception pipeline, and (ii) data from a second perception pipeline that is unavailable to the first perception pipeline.

### Claim 26 (depends on 19)

The one or more non-transitory machine-readable media of claim 19, wherein the operations further comprise:
obtaining a new proposed trajectory for the vehicle in response to a new driving scenario;
obtaining a new speed profile corresponding to the new proposed trajectory from the machine learning network;
determining that the new speed profile does not satisfy the one or more criteria for driving the vehicle;
in response to determining that the obtained speed profile does not satisfy the one or more criteria for driving the vehicle, obtaining a second trajectory based on a predetermined speed profile heuristic; and
causing operation of the vehicle along the second trajectory.

### Claim 27 (depends on 19)

The one or more non-transitory machine-readable media of claim 19, wherein the machine learning network is further trained using another plurality of trajectories and corresponding speed profiles that are determined based on past driving behavior of the vehicle in different driving scenarios.