Stmicroelectronics Gmbh

Stmicroelectronics International N.v.

Ghanapriya Singh

Arun Kumar

Mahesh Chowdhary

Rajendar Bahl

Kashif R J Meer

Indra Narayan Kar

A distributed computing system for artificial intelligence in autonomously appreciating a circumstance context of a smart device. Raw context data is detected by sensors associated with the smart device. The raw context data is pre-processed by the smart device and then provided to a cloud based server for further processing. At the cloud based server, various sets of feature data are obtained from the pre-processed context data. The various sets of feature data are compared with corresponding classification parameters to determine a classification of a continuous event and/or a classification of transient event, if any, which occur in the context. The determined classification of the continuous event and the transient event will be used to autonomously configure the smart device or another related smart device to fit the context.

Context awareness of a smart device through sensing transient and continuous events

A method and apparatus for classifying a spatial environment as open or enclosed are provided. In the method and apparatus, one or more microphones detect ambient sound in a spatial environment and output an audio signal representative of the ambient sound. A processor determines a spatial environment impulse response (SEIR) for the audio signal and extracts one or more features of the SEIR. The processor classifies the spatial environment as open or enclosed based on the one or more features of the SEIR.

Open vs enclosed spatial environment classification for a mobile or wearable device using microphone and deep learning method

Motion activity data is collected from at least one sensor. An initial motion activity classifier function is applied to the motion activity data to produce an initial motion activity posteriorgram. Pre-processing and segmenting the motion activity data into windows produces segmented motion activity data from which sensor specific features are extracted. An updated motion activity classifier function is generated from the extracted sensor specific features. Subsequent motion activity data is also collected from the at least one sensor, and the updated motion activity classifier function is applied to the subsequent motion activity data to produce an updated motion activity posteriorgram.

Method for generating a personalized classifier for human motion activities of a mobile or wearable device user with unsupervised learning

Disclosed herein is a method of operating an electronic device. The method includes activating a first sensing device, and determining a first probabilistic context of the electronic device relative to its surroundings. The method includes outputting the first probabilistic context, and determining a confidence measure of the first probabilistic context. Where the confidence measure of the first probabilistic context is below a threshold, the method includes activating a second sensing device, determining a second probabilistic context of the electronic device relative to its surroundings. outputting the second probabilistic context, and determining a confidence measure of the second probabilistic context. Where the confidence measure of the second probabilistic context is above the threshold, the second sensing device is deactivated and the method returns to determining the first probabilistic context.

Method and apparatus for energy efficient probabilistic context awareness of a mobile or wearable device user by switching between a single sensor and multiple sensors

Described herein is a method of operating an electronic device that includes collecting initial motion activity data from at least one sensor of the electronic device, and generating a initial probabilistic context of the electronic device relative to its surroundings from the initial collected motion activity data using a motion activity classifier function. The collected motion activity data is stored in a training data set, and the motion activity classifier function is updated using the training data set. The method also includes collecting subsequent motion activity data from the at least one sensor of the electronic device, and generating a subsequent probabilistic context of the electronic device relative to its surroundings from the subsequently collected motion activity data using the updated motion activity classifier function.

An electronic device described herein includes a sensing unit having at least one sensor to acquire sensing data. An associated computing device extracts sensor specific features from the sensing data, and generates a motion activity vector, a voice activity vector, and a spatial environment vector as a function of the sensor specific features. The motion activity vector, voice activity vector, and spatial environment vector are processed to determine a base level context of the electronic device relative to its surroundings, with the base level context having aspects each based on the motion activity vector, voice activity vector, and spatial environment vector. Meta level context of the electronic device relative to its surroundings is determined as a function of the base level context, with the meta level context being at least one inference made from at least two aspects of the plurality of aspects of the base level context.

Growing community of inventors

New Delhi, India

Ghanapriya Singh