Detecting driving and modifying access to a user device
Systems and methods are included for detecting driving based on user-specific models for driving detection, and restricting access to an application of the user device while a user is driving. A management agent installed on the user device can collect data from sensors in a user device and provide the data to a management server, which can build a user-specific model for driving detection for that user. The management agent can then use that user-specific model for detecting when the user is driving. When the agent determines that the user is driving, it can enforce a driving policy that limits access to applications and delay or modify notifications generated by applications.
1. A method for detecting driving based on user-specific models for driving detection, comprising:
sending a management agent from a management server to a first user device for installation, wherein the management agent collects first sensor data from the first user device while a first user is logged into the first user device, and wherein the management agent collects first user-interaction data from the first user device based, at least in part, on the first user's interaction with a graphical user interface (GUI) of the first user device;
training a first model at the management server based on receiving the first sensor data and first user-interaction data from the management agent, wherein training comprises modifying the first model based on whether the first model correctly predicts that the first user is driving;
storing the first model at the management server in association with the first user, the first model being different than a second model trained for a second user; and
sending the first model from the management server to the management agent installed on the first user device, for use when the first user is logged in, wherein the management agent detects that the first user is driving by using the first model to analyze data collected from the first user device.
2. The method of claim 1 , wherein the management agent determines a processed feature and provides information associated with the processed feature as an input to the first model.
3. The method of claim 1 , wherein the management agent determines a calibrated orientation of the first user device, based on data received from an accelerometer and gyroscope of the first user device, and sends information associated with the calibrated orientation to the management server.
4. The method of claim 1 , wherein determining whether the first model correctly predicts that the first user is driving comprises prompting the first user with a GUI element that requires the first user's interaction.
5. The method of claim 1 , wherein the first model trained by the management server distinguishes between when the first user is driving and when the first user is a passenger in a moving vehicle.
6. The method of claim 1 , wherein the management agent modifies the first model based on whether the first model correctly predicts that the first user is driving.
7. The method of claim 1 , further comprising enforcing, by the management agent, a driving mode at the first user device based on detecting driving.
8. A non-transitory, computer-readable medium containing instructions that, when executed by a processor of a management server, performs stages for detecting driving at a user device based on user-specific models for driving detection, the stages comprising:
sending a management agent from the management server to a first user device for installation, wherein the management agent collects first sensor data from the first user device while a first user is logged into the first user device, and wherein the management agent collects first user-interaction data from the first user device based, at least in part, on the first user's interaction with a graphical user interface (GUI) of the first user device;
training a first model at the management server based on receiving the first sensor data and first user-interaction data from the management agent, wherein training comprises modifying the first model based on whether the first model correctly predicts that the first user is driving;
storing the first model at the management server in association with the first user, the first model being different than a second model trained for a second user; and
sending the first model from the management server to the management agent installed on the first user device, for use when the first user is logged in, wherein the management agent detects that the first user is driving by using the first model to analyze data collected from the first user device.
9. The non-transitory, computer-readable medium of claim 8 , wherein the management agent determines a processed feature and provides information associated with the processed feature as an input to the first model.
10. The non-transitory, computer-readable medium of claim 8 , wherein the management agent determines a calibrated orientation of the first user device, based on data received from an accelerometer and gyroscope of the first user device, and sends information associated with the calibrated orientation to the management server.
11. The non-transitory, computer-readable medium of claim 8 , wherein determining whether the first model correctly predicts that the first user is driving comprises prompting the first user with a GUI element that requires the first user's interaction.
12. The non-transitory, computer-readable medium of claim 8 , wherein the first model trained by the management server distinguishes between when the first user is driving and when the first user is a passenger in a moving vehicle.
13. The non-transitory, computer-readable medium of claim 8 , wherein the management agent modifies the first model based on whether the first model correctly predicts that the first user is driving.
14. The non-transitory, computer-readable medium of claim 8 , wherein the management agent enforces a driving mode at the first user device based on detecting driving.
15. A system for detecting driving based on user-specific models for driving detection, comprising:
a non-transitory, computer-readable medium that contains instructions;
a management server comprising a processor that executes the instructions to perform stages including:
sending a management agent from the management server to a first user device for installation, wherein the management agent collects first sensor data from the first user device while a first user is logged into the first user device, and wherein the management agent collects first user-interaction data from the first user device based, at least in part, on the first user's interaction with a graphical user interface (GUI) of the first user device;
training a first model at the management server based on receiving the first sensor data and first user-interaction data from the management agent, wherein training comprises modifying the first model based on whether the first model correctly predicts that the first user is driving;
storing the first model at the management server in association with the first user, the first model being different than a second model trained for a second user; and
sending the first model from the management server to the management agent installed on the first user device, for use when the first user is logged in, wherein the management agent detects that the first user is driving by using the first model to analyze data collected from the first user device.
16. The system of claim 15 , wherein the management agent determines a processed feature and provides information associated with the processed feature as an input to the first model.
17. The system of claim 15 , wherein the management agent determines a calibrated orientation of the first user device, based on data received from an accelerometer and gyroscope of the first user device, and sends information associated with the calibrated orientation to the management server.
18. The system of claim 15 , wherein determining whether the first model correctly predicts that the first user is driving comprises prompting the first user with a GUI element that requires the first user's interaction.
19. The system of claim 15 , wherein the first model trained by the management server distinguishes between when the first user is driving and when the first user is a passenger in a moving vehicle.
20. The system of claim 15 , wherein the management agent modifies the first model based on whether the first model correctly predicts that the first user is driving.