Output device, method for generating a machine learning model, and computer program

1. An output device, comprising:

an acquisition unit that acquires input information related to traveling of a human-powered vehicle; and

a processor configured to train a machine learning model in accordance with an index value indicating electric power consumption efficiency in a power supply device for supplying electric power to a component of the human-powered vehicle, and to output output information related to controlling the component when the input information is inputted, wherein

the processor trains the machine learning model by reinforcement learning using a reward based on the index value, the reward being further based on a value indicating comfortability based on an evaluation for a cadence and a torque of a crank of the human-powered vehicle that is traveling.

2. The output device according to claim 1 , wherein the index value is a value related to a travel distance obtained before electric power in the power supply device is equal to or less than a preset value.

3. The output device according to claim 1 , wherein the index value is a value related to an amount of electric power outputted from the power supply device.

4. The output device according to claim 1 , wherein the output information is information related to at least one of an assist ratio in an assist mechanism configured to assist propulsion of the human-powered vehicle by power supplied from the power supply device and an upper limit value of an assist force by the assist mechanism.

5. The output device according to claim 1 , wherein the output information is information related to a gear stage or a gear ratio of a transmission of the human-powered vehicle.

6. The output device according to claim 1 , wherein the output information is information related to a number of rotations of an electric motor driven by electric power supplied from the power supply device.

7. The output device according to claim 1 , wherein the input information includes at least one of information on the cadence of the crank in the human-powered vehicle, information on the torque of the crank, information on power obtained by calculation using the information on the cadence, and the information on the torque.

8. The output device according to claim 1 , wherein the component includes an electric motor driven by electric power supplied from the power supply device, and

the input information is information related to at least one of a torque of the electric motor, an amount of current of the electric motor, a voltage value of the electric motor, and a temperature of the electric motor.

9. The output device according to claim 1 , wherein the input information includes information related to an amount of electric power remaining in the power supply device.

10. The output device according to claim 1 , wherein the input information includes detection data related to an attitude of the human-powered vehicle.

11. The output device according to claim 1 , wherein the input information includes detection data related to a traveling environment of the human-powered vehicle.

12. The output device according to claim 1 , wherein

output information that allows comfortability of a user for operation of the component of the human-powered vehicle that is traveling to fall within a preset range is outputted in accordance with the output information outputted from the machine learning model.

13. An output device, comprising:

an acquisition unit that acquires input information related to traveling of a human-powered vehicle;

a processor configured to train a machine learning model in accordance with an index value indicating electric power consumption efficiency in a power supply device for supplying electric power to a component of the human-powered vehicle, and to output output information related to controlling the component when the input information is inputted; and

an operating unit that is configured to input operation information for operating the component, wherein

the processor trains the machine learning model by reinforcement learning using the index value as a reward, and

the processor is configured to provide output information with the reward based on collation of the output information outputted from the machine learning model in accordance with input information acquired by the acquisition unit with operation information inputted by the operating unit.

14. An output device, comprising:

an acquisition unit that acquires input information related to traveling of a human-powered vehicle;

a processor configured to train a machine learning model in accordance with an index value indicating electric power consumption efficiency in a power supply device for supplying electric power to a component of the human-powered vehicle, and to output output information related to controlling the component when the input information is inputted;

an output unit that is configured to output the output information related to controlling the component of the human-powered vehicle that is outputted by the machine learning model; and

an evaluation input unit that is configured to input an evaluation indicating comfortability of a user after the component is operated in accordance with the output information, wherein

the machine learning model is configured to be updated in accordance with training data including a detail of the evaluation inputted, the output information outputted from the machine learning model at a timing when the evaluation is inputted, and the input information inputted in accordance with the output information.

15. The output device according to claim 14 , wherein the output unit is configured to output the output information indicating a result of controlling the component.

16. A method for generating a machine learning model, comprising:

configuring a neural network to output output information related to controlling a component of a human-powered vehicle when input information related to traveling of the human-powered vehicle is inputted;

acquiring input information related to traveling of the human-powered vehicle;

identifying output information outputted by providing the neural network with the input information acquired;

training a weight in an intermediate layer of the neural network such that an index value indicating electric power consumption efficiency in a power supply device configured to supply electric power to the component is improved when the component is controlled in accordance with the output information identified; and

training the neural network by reinforcement learning using a reward based on the index value, the reward being further based on a value indicating comfortability based on an evaluation for a cadence and a torque of a crank of the human-powered vehicle that is traveling.

17. A computer-readable storage medium comprising instructions configured to be executed by a processor of a computer, to cause the processor to:

configure a neural network to output output information related to controlling a component of a human-powered vehicle when input information related to traveling of the human-powered vehicle is inputted;

acquire input information related to traveling of the human-powered vehicle;

identify output information outputted by providing the neural network with the input information acquired; and

train a weight in an intermediate layer of the neural network such that an index value indicating electric power consumption efficiency in a power supply device configured to supply electric power to the component is improved when the component is controlled in accordance with the output information identified, wherein

the processor trains the neural network by reinforcement learning using a reward based on the index value, the reward being further based on a value indicating comfortability based on an evaluation for a cadence and a torque of a crank of the human-powered vehicle that is traveling.