US 11,669,770 B2
Activity recognition method with automatic training based on inertial sensors
Alberto Zancanato, Milan (IT); and Stefano Paolo Rivolta, Desio (IT)
Assigned to STMICROELECTRONICS S.r.l., Agrate Brianza (IT)
Filed by STMicroelectronics S.r.l., Agrate Brianza (IT)
Filed on Nov. 28, 2018, as Appl. No. 16/203,275.
Prior Publication US 2020/0167559 A1, May 28, 2020
Int. Cl. G06N 20/00 (2019.01); G01C 19/00 (2013.01); G01P 15/00 (2006.01); G06F 3/0346 (2013.01)
CPC G06N 20/00 (2019.01) [G01C 19/00 (2013.01); G01P 15/00 (2013.01); G06F 3/0346 (2013.01); G06F 2218/02 (2023.01); G06F 2218/08 (2023.01); G06F 2218/12 (2023.01)] 25 Claims
OG exemplary drawing
 
1. An integrated circuit, comprising:
an inertial sensor, which, in operation, generates inertial sensor data associated with the integrated circuit; and
processing circuitry communicatively coupled to the inertial sensor, wherein the processing circuitry of the integrated circuit, in operation:
determines a new feature array from the inertial sensor data;
determines if the new feature array is within an existing class within a state space associated with the inertial sensor data;
in response to the new feature array being included in the existing class, adds the new feature array to the existing class and updates a representation of the existing class in the state space based on the new feature array and an existing representation of the existing class;
in response to the new feature array not being included in the existing class, creates a new class based on the new feature array;
determines if the new feature array is within the state space; and
in response to the new feature array not being included in the state space:
renormalizes the state space based on the new feature array and modifies the representation of the existing class based on the renormalized state space;
renormalizes the state space based on the new feature array and rescales an area of the existing class based on the renormalized state space; or
renormalizes the state space based on the new feature array, modifies the representation of the existing class based on the renormalized state space, and rescales the area of the existing class based on the renormalized state space.