Displaying mobile advertising based on determining user's physical activity from mobile device sensor data

• US 10,127,565 B2
• Filed: 08/14/2012
• Issued: 11/13/2018
• Est. Priority Date: 12/09/2011
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Phan, Thomas
• Primary Examiner(s): Stoltenberg, David J
• Assistant Examiner(s): Beheshti, Brendon
• Application Number:
  US13/585,740
• Publication Number:
  US 20130151343A1
• Time in Patent Office: 2,282 Days
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: H04W 4/00 Services specially adapted for wireless communication networks; Facilities therefor
• CPC Class Codes: H04W 4/02 Services making use of location information
• CPC Class Codes: H04W 4/21 for social networking applications

Abstract

A mobile handset device determines a physical activity performed by a user based on local sensor data. The local sensor data can include sensors within the mobile handset device and local sensors in communication with the mobile handset device. The mobile handset device provides an indicator of the physical activity to a publisher. The publisher targets advertisements based on the current activity or activity pattern of the user.

Claims

1. A method of providing mobile advertising to a mobile handset device having a processor and a memory, comprising:

processing a model encoded on the mobile handset device, the model including a decision tree comprising activity labels and frequency domain features extracted from sensor data;

capturing, at the mobile handset device, new sensor data from a combination of sensors connected to the mobile handset device;

classifying, at the mobile handset device, the new sensor data as a perceived activity from a physical activity of a user of the mobile handset device, the classifying comprising inputting the new sensor data to the model, and outputting one of the activity labels indicating the perceived activity;

generating one or more keywords as an indicator of the perceived activity based on said classifying;

transmitting the one or more keywords to a server; and

receiving at least one advertisement selected based on at least one advertising keyword associated with the indicator.

2. The method of claim 1, further comprising displaying the at least one advertisement on a display of the mobile handset device.

3. The method of claim 1, wherein the sensors reside within the mobile handset device.

6. The method of claim 1, wherein the model is built by:

partitioning training sensor data samples using at least one window;

determining per window magnitude frequency response for the partitioned training sensor data samples; and

extracting the frequency domain features from the per window magnitude frequency response for the partitioned training sensor data samples.

7. The method of claim 6, wherein the one of the activity labels output from the model comprises a JavaScript event that is output as an activity predicate on the mobile handset device.

4. A non-transitory processor-readable medium that includes a program that when executed by a processor performs a method comprising:

processing a model encoded on a mobile handset device, the model including a decision tree comprising activity labels and frequency domain features extracted from sensor data;

capturing, at the mobile handset device, new sensor data from a combination of sensors connected to the mobile handset device;

classifying, at the mobile handset device, the new sensor data as a perceived activity from a physical activity of a user of the mobile handset, the classifying comprising inputting the new sensor data to the model, and outputting one of the activity labels indicating the perceived activity;

generating one or more keywords as an indicator of the perceived activity based on said classifying;

transmitting the one or more keywords to a server; and

receiving at least one advertisement selected based on at least one advertising keyword associated with the indicator.

5. The non-transitory processor-readable medium of claim 4, wherein the indicator indicates a pattern for the perceived activity based on sensor data collected over time for a particular week day.

8. The non-transitory processor-readable medium of claim 4, wherein the model is built by:

partitioning training sensor data samples using at least one window;

determining per window magnitude frequency response for the partitioned training sensor data samples; and

extracting the frequency domain features from the per window magnitude frequency response for the partitioned training sensor data samples.

9. The non-transitory processor-readable medium of claim 8, wherein the one of the activity labels output from the model comprises a JavaScript event that is output as an activity predicate on the mobile handset device.