Ultrasound analytics for actionable information

1. A monitoring system comprising:

one or more sensors located at a property and configured to generate first sensor data;

a monitor control unit configured to:

receive the first sensor data;

based on the first sensor data, generate an alarm event for the property; and

based on generating the alarm event for the property, dispatch an autonomous drone;

the autonomous drone configured to:

generate, using an onboard sensor, second sensor data;

based on the second sensor data, determine a location within the property where a person is likely located;

navigate to the location within the property where the person is likely located;

based on navigating to the location within the property where the person is likely located, generate a set of ultrasound scans of the person; and

provide, for output, data indicating the set of ultrasound scans to a remote processing engine; and

the remote processing engine configured to:

provide the set of ultrasound scans to a model trained to predict health conditions of users based on data represented by input ultrasound scans;

obtain, as output of the model, a predicted health condition of the person that was predicted by the model based on the set of ultrasound scans; and

provide, for output, data indicating the predicted health condition of the person.

2. The monitoring system of claim 1 , wherein the autonomous drone is configured to:

generate a map of the property; and

provide, for output, the data indicating the set of ultrasound scans to the remote processing engine by providing, for output, the map of the property with the location where the person is likely located to the remote processing engine.

3. The monitoring system of claim 1 , wherein the predicted health condition of the person indicates that the person is likely injured.

4. The monitoring system of claim 1 , wherein the autonomous drone is configured to:

provide the second sensor data as an input to a second model trained to identify locations of people; and

determine the location within the property where a person is likely located based on an output of the second model trained to identify locations of people based on the second sensor data.

5. The monitoring system of claim 4 , wherein the autonomous drone is configured to:

receive labeled training data that includes first labeled sensor data that corresponds locations with people and second labeled sensor data that corresponds to locations without people;

train, using machine learning, the first labeled sensor data, and the second labeled sensor data, the second model to identify locations of people based on the second sensor data.

6. The monitoring system of claim 1 , wherein the autonomous drone is configured to:

based on the second sensor data, determine that the person is likely alive; and

provide, for output, data indicating that the person is likely alive.

7. The monitoring system of claim 6 , wherein:

the autonomous drone comprises a microphone and the second sensor data is audio data collected by the microphone, and

the autonomous drone is configured to:

provide the audio data as an input to a third model trained to identify human sounds; and

determine that the person is likely alive based on an output of the third model trained to identify human sounds.

8. The monitoring system of claim 1 , wherein the autonomous drone is configured to:

based on determining the location within the property where a person is likely located, activate a communication channel between a device outside the property and the autonomous drone.

9. A computer-implemented method, comprising:

generating, by one or more sensors of a monitoring system that is configured to monitor a property, first sensor data;

based on the first sensor data, generating, by the monitoring system, an alarm event for the property;

based on generating the alarm event for the property, dispatching, by the monitoring system, an autonomous drone;

generating, by the autonomous drone, second sensor data;

based on the second sensor data, determining, by the monitoring system, a location within the property where a person is likely located;

navigating, by the autonomous drone, to the location within the property where the person is likely located;

based on navigating to the location within the property where the person is likely located, generating, by the autonomous drone, a set of ultrasound scans of the person;

providing, for output by the autonomous drone and to a remote processing engine of the monitoring system, data indicating the set of ultrasound scans;

providing, by the remote processing engine, the set of ultrasound scans to a model trained to predict health conditions of users based on data represented by input ultrasound scans;

obtaining, by the remote processing engine and as output of the model, a predicted health condition of the person that was predicted by the model based on the set of ultrasound scans; and

providing, for output by the monitoring system, data indicating the predicted health condition of the person.

10. The computer-implemented method of claim 9 , comprising:

generating, by the monitoring system, a map of the property; and

providing, by the monitoring system, for output, data indicating the set of ultrasound scans to the remote processing engine by providing, for output, the map of the property with the location where the person is likely located.

11. The computer-implemented method of claim 9 , the predicted health condition of the person indicates that the person is likely injured.

12. The computer-implemented method of claim 9 , comprising:

providing, by the autonomous drone of the monitoring system, second sensor data as an input to a second model trained to identify locations of people; and

determining, by the monitoring system, the location within the property where a person is likely located based on an output of the second model trained to identify locations of people based on the second sensor data.

13. The computer-implemented method of claim 12 , comprising:

receiving, by the monitoring system, labeled training data that includes first labeled sensor data that corresponds locations with people and second labeled sensor data that corresponds to locations without people; and

training, by the monitoring system, using machine learning, the first labeled sensor data, and the second labeled sensor data, the second model to identify locations of people based on the second sensor data.

14. The computer-implemented method of claim 9 , comprising:

based on the second sensor data, determining, by the monitoring system, that the person is likely alive; and

providing, for output by the monitoring system, data indicating that the person is likely alive.

15. The computer-implemented method of claim 14 , wherein:

the autonomous drone comprises a microphone and the second sensor data is audio data collected by the microphone, and the method comprises:

providing, by the monitoring system, the audio data as an input to a third model trained to identify human sounds; and

determining, by the monitoring system, that the person is likely alive based on an output of the third model trained to identify human sounds.

16. The computer-implemented method of claim 9 , comprising:

based on determining the location within the property where a person is likely located, activating, by the monitoring system, a communication channel between a device outside the property and the autonomous drone.

17. The monitoring system of claim 1 , wherein the remote processing engine is configured to:

based on the predicted health condition of the person, determine (i) that the person is likely injured and (ii) a severity of the injury to the person; and

provide, for output, data indicating the predicted health condition of the person by providing, for output, data indicating (i) that the person is likely injured and (ii) the severity of the injury to the person.