| CPC G05B 13/027 (2013.01) [G05B 2219/25257 (2013.01)] | 20 Claims |
|
1. A training arrangement for training a machine learning model for state determination in control, comprising a microcontroller and a data processing system external to the microcontroller, wherein the microcontroller:
is configured to receive operational information about a plant and feed the operational information to a machine learning model implemented on the microcontroller and configured to determine plant state information about the plant from the operational information; and
is configured to control the plant using the plant state information; and
wherein the microcontroller comprises a hardware trace system and a corresponding filter configured to capture microcontroller trace information corresponding to input and output data between a core of the microcontroller and the machine learning model associated with the operational information and the plant state information and to transmit the captured microcontroller trace information to the data processing system without interfering with normal operation of the of the core; and
wherein the data processing system is configured to receive the operational information, the microcontroller trace information, and the plant state information, to compare the plant state information with ground truth sensor information about the state of the plant, to determine whether an update for the machine learning model is required depending on a comparison of the microcontroller trace information and the ground truth sensor information, if it has determined that an update for the machine learning model is required, to determine a learning model update for the machine learning model and transmit the learning model update to the microcontroller; and
wherein the microcontroller is configured to update the machine learning model in response to reception of the learning model update from the data processing system.
|