US 11,868,197 B2
Learning method for the detection of anomalies implemented on a microcontroller and associated method for detecting anomalies
Francois De Rochebouet, Toulon (FR); and He Huang, Toulon (FR)
Assigned to STMicroelectronics International N.V., Geneva (CH)
Filed by Cartesiam, Toulon (FR)
Filed on Apr. 25, 2022, as Appl. No. 17/660,548.
Application 17/660,548 is a continuation of application No. 16/880,133, filed on May 21, 2020, granted, now 11,341,016.
Claims priority of application No. 1905293 (FR), filed on May 21, 2019.
Prior Publication US 2022/0245046 A1, Aug. 4, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 11/00 (2006.01); G06F 11/07 (2006.01); G06F 9/50 (2006.01); G06F 11/30 (2006.01); G06F 11/32 (2006.01); G06F 18/22 (2023.01)
CPC G06F 11/0751 (2013.01) [G06F 9/5016 (2013.01); G06F 11/3013 (2013.01); G06F 11/3072 (2013.01); G06F 11/3089 (2013.01); G06F 11/327 (2013.01); G06F 18/22 (2023.01)] 20 Claims
OG exemplary drawing
 
1. A method for detecting anomalies implemented on a microcontroller having a memory, the method comprising:
receiving, by the microcontroller, data sets from a sensor; and
in response to determining that a rate of progress is less than a threshold, for each data set of a set of the data sets:
storing the data set in a memory of the microcontroller,
calculating a signature of the data set, and
in response to determining that a category is stored in the memory:
calculating a similarity measurement between the signature of the data set and the signature of each category, each category comprising a signature and an occurrence,
selecting a maximum similarity measurement from calculated similarity measurements, and
updating a corresponding category from the data set in response to the maximum similarity measurement meeting a condition of updating.