Verfahren und System zur Erkennung eines Bewegungszustands einer Elektronischen Vorrichtung, insbesondere Ob ein Benutzer die Elektronische Vorrichtung in einem Beutel Trägt

Details

Veröffentlichungs-Nr.

EP4783559

Anmeldetag

9. Januar 2026

Veröffentlichung

29. Juli 2026

Rechtsraum

EP

Abstract

An in-bag detection method computes a plurality of features from sensors data and, through an AI model, performs a classification of the state of motion of an electronic device based on the plurality of features. The method then performs a meta-classification over an output of the AI model by using counters for each of the values that the output can assume to check for consistency of the output. Once the counters of the meta-classifier of the output of the AI model have reached consistency, an inbuilt stability mechanism is triggered. The inbuilt stability mechanism does not require allocation of timers or other external resources. Therefore, instead of calculating the entire plurality of features needed by the AI models, only a subset of features related to motion is calculated.

Anmelder

Firma

STMicroelectronics International N.V.

Land

🇨🇭 Schweiz