Telemetry reporting in vehicle super resolution systems
In one embodiment, a processor of a vehicle detects a difference between a physical characteristic of the vehicle predicted by a first machine learning-based model and a physical characteristic of the vehicle indicated by telemetry data generated by a sub-system of the vehicle. The processor forms a packet payload of an update packet indicative of the detected difference, based in part on a relevancy of the physical characteristic to the first machine learning-based model. The processor applies a synchronization strategy to the update packet, to synchronize the update packet with a second machine learning-based model executed by a receiver. The processor sends the update packet to the receiver via a network, to update the second machine learning-based model.
1. A method comprising:
receiving, by a gateway node local to a vehicle, first telemetry data generated by a first subsystem of the vehicle;
generating, by the gateway node, a first predicted state of the first subsystem of the vehicle, based in part on data generated by a machine learning model using the first telemetry data;
subsequently receiving, by the gateway node, second telemetry data generated by the first subsystem of the vehicle;
determining, by the gateway node, an indicated state of the first subsystem of the vehicle, based in part on the second telemetry data, wherein the indicated state differs from the first predicted state; and
sending, by the gateway node, information about a difference between the first predicted state and the indicated state to a remote service.
2. The method of claim 1 , wherein the remote service updates a remote machine learning model based upon the first predicted state of the first subsystem.
3. The method of claim 1 , further comprising:
receiving the first telemetry data and the second telemetry data via a Controller Area Network (CAN) bus.
4. The method of claim 1 , wherein the machine learning model is configured to predict a physical characteristic of the vehicle for input to an application.
5. The method of claim 1 , wherein the indicated state of the first subsystem of the vehicle differs from the first predicted state according to at least one defined threshold.
6. The method of claim 1 , wherein sending, by the gateway node, information about the difference between the first predicted state and the indicated state to the remote service includes forming an update packet.
7. A tangible, non-transitory, computer-readable medium storing instructions that cause a gateway node local to a vehicle to execute a process comprising:
receiving first telemetry data generated by a first subsystem of the vehicle;
generating a first predicted state of the first subsystem of the vehicle, based in part on data generated by a machine learning model using the first telemetry data;
subsequently receiving second telemetry data generated by the first subsystem of the vehicle;
determining an indicated state of the first subsystem of the vehicle, based in part on the second telemetry data, wherein the indicated state differs from the first predicted state;
sending information about a difference between the first predicted state and the indicated state to a remote service.
8. The computer-readable medium of claim 7 , wherein the first telemetry data and the second telemetry data are received via a Controller Area Network (CAN) bus.
9. The computer-readable medium of claim 7 , wherein the machine learning model is configured to predict a physical characteristic of the vehicle for input to an application.
10. The computer-readable medium of claim 7 , wherein the indicated state of the first subsystem of the vehicle differs from the first predicted state according to at least one defined threshold.
11. The computer-readable medium of claim 7 , wherein sending information about the difference between the first predicted state and the indicated state to the remote service includes forming an update packet.
12. A gateway node local to a vehicle comprising:
a processor, a memory, at least one connection to a sensor network, and at least one connection to a communications network;
wherein the gateway node is configured to:
receive, via the sensor network, first telemetry data generated by a first subsystem of the vehicle;
generate a first predicted state of the first subsystem of the vehicle, based in part on data generated by a machine learning model using the first telemetry data;
subsequently receive second telemetry data generated by the first subsystem of the vehicle;
determine an indicated state of the first subsystem of the vehicle, based in part on the second telemetry data, wherein the indicated state differs from the first predicted state; and
send information about a difference between the first predicted state and the indicated state to a remote service via the communications network.
13. The gateway node of claim 12 , wherein the at least one connection to the sensor network is via a Controller Area Network (CAN) bus.
14. The gateway node of claim 12 , wherein the machine learning model is configured to predict a physical characteristic of the vehicle for input to an application.
15. The gateway node of claim 12 , wherein the indicated state of the first subsystem of the vehicle differs from the first predicted state according to at least one defined threshold.
16. The gateway node of claim 12 , wherein the gateway node is configured to send information about the difference between the first predicted state and the indicated state to the remote service by forming an update packet.