Inferring TV Viewing via Device State Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining if a mobile device user is watching TV and identifying the show being watched are intrusive, require user participation, or are not practical due to the lack of connectivity between mobile devices and TVs, and do not account for privacy concerns.
Innovation Solution
A method that passively monitors mobile device usage patterns and correlates them with TV program schedules to infer if a user is watching live TV and which show, using a user behavior model that updates based on activity streams, device states, and historical data, while reducing dimensionality and noise through normalization and affinity scoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If audio fingerprinting is used to identify TV content on mobile devices, then content identification accuracy is improved, but user privacy is compromised due to continuous microphone monitoring
Solution Approach 1:
The patent extracts only the necessary information (device active/inactive states) from the mobile device data stream, rather than continuously monitoring all device functions or audio content. This selective extraction achieves content identification while minimizing privacy intrusion by capturing only state information, not content or conversations.
Solution Approach 2:
The patent introduces an intermediary correlation analysis mechanism that indirectly infers TV watching behavior through device state patterns rather than directly monitoring TV content or user conversations. This intermediary approach uses device activity as a mediator to deduce viewing behavior without direct access to private user data.
2Measurement precision
If direct connection between mobile device and TV is established, then content information accuracy is improved, but device complexity and compatibility requirements increase
Solution Approach 1:
The patent makes the mobile device the universal monitoring point for all TV watching scenarios, regardless of TV type or connectivity. Instead of requiring TVs or set-top boxes to have specific networking capabilities, the solution universally uses the mobile device's existing sensors and state tracking to infer viewing behavior across all TV configurations.
Solution Approach 2:
The mobile device serves itself by using its own existing state information (active/inactive states from its operating system) to infer TV watching behavior. The device does not need to connect to TV or external monitoring infrastructure; it autonomously generates the data needed for content identification through correlation analysis of its own usage patterns.
3Reliability
If specific applications are launched on mobile devices to connect with TVs, then connection reliability is improved, but ease of operation deteriorates due to user burden
Solution Approach 1:
The system automatically performs correlation analysis between device states and TV program schedules without requiring user action. The mobile device and server system handle all processing autonomously, inferring viewing behavior and identifying content without the user needing to launch applications or configure connections.
Solution Approach 2:
The system pre-processes and stores TV program schedules and device state data in advance, so that when analysis is needed, the correlation can be quickly performed using pre-prepared data. This preliminary preparation eliminates the need for real-time user intervention or complex runtime processing.
4Measurement precision
If continuous microphone monitoring is implemented, then content identification coverage is improved, but energy consumption increases
Solution Approach 1:
The patent extracts only essential state information (active/inactive states) from the mobile device, rather than continuously monitoring high-energy functions like microphone audio processing. This selective data extraction significantly reduces energy consumption while maintaining content identification capability through correlation analysis.
Solution Approach 2:
Instead of continuous monitoring, the system periodically samples device state changes and correlates them with program schedules. This periodic approach, rather than continuous operation, dramatically reduces energy consumption while still capturing sufficient data to identify viewing behavior and content.
Data Source
AI summary
There is disclosed a technique of associating device activity to a broadcast programme, comprising: receiving a model for a broadcast programme identifying portions of content and portions of breaks in the content; monitoring, via a client software module running on users' mobile devices, said device's active or inactive states; receiving an activity stream of a user device; comparing the activity stream to the model of the broadcast programme to identify a correlation between them; inferring a user of the user device as watching the broadcast programme based on a level of the correlation.