Scheduling state transitions in an autonomous vehicle

1. A method, comprising:

detecting a transition signal for transitioning from a first state associated with a first machine learning model to a second state associated with a second machine learning model, wherein an autonomous vehicle is configured to use the first machine learning model instead of the second machine learning model when in the first state, and wherein the autonomous vehicle is configured to use the second machine learning model instead of the first machine learning model when in the second state;

detecting, subsequent to detecting the transition signal, that a precondition for generating output by the second machine learning model is unsatisfied, wherein the precondition comprises a predefined number of frames of image data input to the second machine learning model; and

delaying, until the precondition for generating output by the second machine learning model is satisfied, use of the second machine learning model instead of the first machine learning model by delaying a transition from the first state to the second state.

2. The method of claim 1 , wherein delaying the transition from the first state to the second state comprises transitioning from the first state to an intermediate state between the first state and the second state.

3. The method of claim 2 , further comprising transitioning, when the precondition is satisfied, from the intermediate state to the second state.

4. The method of claim 1 , wherein the first state is included in a first state space associated with the first machine learning model and the second state is included in a second state space associated with the second machine learning model.

5. The method of claim 1 , wherein the first machine learning model and the second machine learning model are configured to generate one or more control actions for the autonomous vehicle.

6. The method of claim 1 , wherein the precondition comprises a predefined number of outputs of a third machine learning model having been provided to the second machine learning model.

7. The method of claim 1 , wherein the transition signal comprises one or more of: an operational parameter of the autonomous vehicle crossing a threshold, an output of a machine learning model, a user input, or a detected error.

8. An apparatus configured to perform steps, comprising:

detecting a transition signal for transitioning from a first state associated with a first machine learning model to a second state associated with a second machine learning model, wherein an autonomous vehicle is configured to use the first machine learning model instead of the second machine learning model when in the first state, and wherein the autonomous vehicle is configured to use the second machine learning model instead of the first machine learning model when in the second state;

detecting, subsequent to detecting the transition signal, that a precondition for generating output by the second machine learning model is unsatisfied, wherein the precondition comprises a predefined number of frames of image data input to the second machine learning model; and

delaying, until the precondition for generating output by the second machine learning model is satisfied, use of the second machine learning model instead of the first machine learning model by delaying a transition from the first state to the second state.

9. The apparatus of claim 8 , wherein delaying the transition from the first state to the second state comprises transitioning from the first state to an intermediate state between the first state and the second state.

10. The apparatus of claim 9 , wherein the steps further comprise transitioning, when the precondition is satisfied, from the intermediate state to the second state.

11. The apparatus of claim 8 , wherein the first state is included in a first state space associated with the first machine learning model and the second state is included in a second state space associated with the second machine learning model.

12. The apparatus of claim 8 , wherein the first machine learning model and the second machine learning model are configured to generate one or more control actions for the autonomous vehicle.

13. The apparatus of claim 8 , wherein the precondition comprises a predefined number of outputs of a third machine learning model having been provided to the second machine learning model.

14. The apparatus of claim 8 , wherein the transition signal comprises one or more of: an operational parameter of the autonomous vehicle crossing a threshold, an output of a machine learning model, a user input, or a detected error.

15. An autonomous vehicle, comprising:

an apparatus configured to perform steps comprising:

detecting a transition signal for transitioning from a first state associated with a first machine learning model to a second state associated with a second machine learning model, wherein an autonomous vehicle is configured to use the first machine learning model instead of the second machine learning model when in the first state, and wherein the autonomous vehicle is configured to use the second machine learning model instead of the first machine learning model when in the second state;

detecting, subsequent to detecting the transition signal, that a precondition for generating output by the second machine learning model is unsatisfied, wherein the precondition comprises a predefined number of frames of image data input to the second machine learning model; and

delaying, until the precondition for generating output by the second machine learning model is satisfied, use of the second machine learning model instead of the first machine learning model by delaying a transition from the first state to the second state.

16. The autonomous vehicle of claim 15 , wherein delaying the transition from the first state to the second state comprises transitioning from the first state to an intermediate state between the first state and the second state.

17. The autonomous vehicle of claim 16 , wherein the steps further comprise transitioning, when the precondition is satisfied, from the intermediate state to the second state.

18. The autonomous vehicle of claim 15 , wherein the first state is included in a first state space associated with the first machine learning model and the second state is included in a second state space associated with the second machine learning model.

19. The autonomous vehicle of claim 15 , wherein the first machine learning model and the second machine learning model are configured to generate one or more control actions for the autonomous vehicle.

20. A computer program product disposed upon a non-transitory computer-readable medium, the computer program product comprising computer program instructions for scheduling state transitions in an autonomous vehicle that, when executed, cause a computer system of the autonomous vehicle to carry out the steps of:

detecting a transition signal for transitioning from a first state associated with a first machine learning model to a second state associated with a second machine learning model, wherein an autonomous vehicle is configured to use the first machine learning model instead of the second machine learning model when in the first state, and wherein the autonomous vehicle is configured to use the second machine learning model instead of the first machine learning model when in the second state;

detecting, subsequent to detecting the transition signal, that a precondition for generating output by the second machine learning model is unsatisfied, wherein the precondition comprises a predefined number of frames of image data input to the second machine learning model; and

delaying, until the precondition for generating output by the second machine learning model is satisfied, use of the second machine learning model instead of the first machine learning model by delaying a transition from the first state to the second state.