System and method for machine learning (ML)-based IoT device provisioning

Patent metadata

Field	Value
Publication number	US-12524901-B2
Application number	US-202217748541-A
Country	US
Kind code	B2
Filing date	May 19, 2022
Priority date	May 19, 2022
Publication date	Jan 13, 2026
Grant date	Jan 13, 2026

Abstract

A system and method are described for identifying an IoT device using object recognition techniques. For example, one embodiment of a system comprises: an Internet of Things (IoT) service to provide back-end data processing for a plurality of IoT devices, the IoT service comprising: interface logic to couple the IoT service to an IoT app executed on a mobile device of a user, an IoT device recognition engine coupled to the interface logic, the IoT device recognition engine to identify a model of a new IoT device captured in an image by the IoT app, the IoT device recognition engine to transmit an indication of the IoT device model to the interface logic, wherein the IoT app is to use the indication of the IoT device model during setup of the new IoT device.

First claim

What is claimed is:

1 . A system comprising:
an Internet of Things (IoT) service to provide back-end data processing for a plurality of IoT devices, the IoT service comprising:
device provisioning logic to provision IoT devices to operate with the IoT service, the device provisioning logic operable as an interface to couple the IoT service to an IoT app executed on a mobile device of a user, wherein the IoT app is to step the user through an installation process to register a new IoT device with the IoT service, wherein at a particular stage in the installation process, the IoT app is to prompt the user to capture an image of the new IoT device, or a portion thereof, the IoT app to transmit image data corresponding to the captured image to the IoT service via the device provisioning logic;
image pre-processing logic operable on the IoT service to normalize the image data to produce normalized image data, wherein to normalize the image data, the image pre-processing logic is to perform one or more of: scaling the image data in accordance with a particular resolution and translating the image data into a data format usable for further processing; and
an IoT device recognition engine comprising a machine learning (ML)-based object recognition model pre-trained with images of IoT device models, the IoT device recognition engine coupled to the device provisioning logic, the IoT device recognition engine to process the normalized image data with the ML-based object recognition model to identify an IoT device model of the new IoT device captured in the image by the IoT app, the IoT device recognition engine to transmit information indicating a model number of the new IoT device to the IoT app via the interface logic, wherein the IoT app is to use the information indicating the model number of the new IoT device during the installation process.

2 . The system of claim 1 , further comprising:
training logic to train the ML-based object recognition engine using a plurality of images of a corresponding plurality of IoT devices, the ML-based object recognition model to associate object characteristics extracted from the plurality of images with the corresponding plurality of IoT devices.

3 . The system of claim 1 , wherein the IoT app is to prompt the user to capture the image of the new IoT device upon a user indication that a machine-readable optical code is not available for scanning.

4 . The system of claim 3 , wherein the IoT app is to use the information indicating the model number of the IoT device to instruct the user as to how to configure the IoT device to enter into a setup mode.

5 . The system of claim 4 , wherein, while in the setup mode, the IoT app is to configure the IoT device to communicate over a wireless network of the user.

6 . The system of claim 5 , wherein the IoT device is to establish a secure communication channel with the IoT service via the wireless network, and wherein the IoT service is to associate the IoT device with an account of the user.

7 . A method comprising:
initiating a setup process on an IoT app to register a new IoT device with an account of a user on an IoT service; detecting, via input from the user on the IoT app, that no machine-readable optical code is available during the setup process for the new IoT device; prompting, by the IoT app, the user to capture an image of the IoT device, or a portion thereof; transmitting image data corresponding to the image of the IoT device, or portion thereof, to the IoT service; normalizing the image data by image pre-processing logic operable on the IoT service to produce normalized image data, wherein to normalize the image data, the image pre-processing logic is to perform one or more of: scaling the image data in accordance with a particular resolution and translating the image data into a data format usable for further processing; identifying, by an IoT device recognition engine of the IoT service, a model of the new IoT device based on the image, or portion thereof, the IoT device recognition engine comprising a machine learning (ML)-based object recognition model pre-trained with images of IoT device models, the IoT device recognition engine coupled to the device provisioning logic, the IoT device recognition engine to process the normalized image data with the ML-based object recognition model to identify an IoT device model of the IoT device captured in the image by the IoT app; transmitting, by the IoT service, information indicating a model number of the new IoT device to the IoT app, wherein the IoT app is to use the information indicating the model number during the setup process; and completing the setup process on the IoT app in accordance with the information indicating the model number of the IoT device.

8 . The method of claim 7 , further comprising:
training the ML-based object recognition engine using a plurality of images of a corresponding plurality of IoT devices, the ML-based object recognition engine to associate object characteristics extracted from the plurality of images with the corresponding plurality of IoT devices.

9 . The method of claim 7 , wherein the IoT app is to prompt the user to capture the image of the new IoT device upon a user indication that a machine-readable optical code is not available for scanning.

10 . The method of claim 9 , wherein the IoT app is to use the information indicating the model number of the IoT device to instruct the user as to how to configure the IoT device to enter into a setup mode.

11 . The method of claim 10 , further comprising:
configuring, by the IoT app in the setup mode, the IoT device to communicate over a wireless network of the user.

12 . The method of claim 11 , further comprising:
establishing, by the IoT device, a secure communication channel with the IoT service via the wireless network; and associating IoT device with the account of the user.

13 . A non-transitory machine-readable medium having program code stored thereon which, when executed by one or more machines, is to cause the one or more machines to perform the operation of:
initiating a setup process on an IoT app to register a new IoT device with an account of a user on an IoT service; detecting, via input from the user on the IoT app, that no machine-readable optical code is available during the setup process for the new IoT device; prompting, by the IoT app, the user to capture an image of the IoT device, or a portion thereof; transmitting image data corresponding to the image of the IoT device, or portion thereof, to the IoT service; normalizing the image data by image pre-processing logic operable on the IoT service to produce normalized image data, wherein to normalize the image data, the image pre-processing logic is to perform one or more of: scaling the image data in accordance with a particular resolution and translating the image data into a data format usable for further processing; identifying, by an IoT device recognition engine of the IoT service, a model of the new IoT device based on the image, or portion thereof, the IoT device recognition engine comprising a machine learning (ML)-based object recognition model pre-trained with images of IoT device models, the IoT device recognition engine coupled to the device provisioning logic, the IoT device recognition engine to process the normalized image data with the ML-based object recognition model to identify an IoT device model of the IoT device captured in the image by the IoT app; transmitting, by the IoT service, information indicating a model number of the new IoT device to the IoT app, wherein the IoT app is to use the information

Assignees

• Afero Inc

Inventors

• Britt Joe
• Quinn Kerry

Classifications

• H04W76/10 Primary

  Connection setup · CPC title

• G06V10/32

  Normalisation of the pattern dimensions · CPC title

• G06V10/774

  Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• G06V2201/07

  Target detection · CPC title

• H04W12/033

  of the user plane, e.g. user's traffic · CPC title