Entity Vector Generation for Wireless Sensor Data Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing approaches to processing crowdsourced wireless device sensor data lack the ability to generate a universal model of the physical world that leverages vast amounts of sensor data combined with user interactions, such as search engine searches and web browsing history, to create entity vectors that reflect user-independent and user-dependent properties of physical entities.
Innovation Solution
A method and system that efficiently collect and analyze wireless device sensor data and user interaction history to generate universal feature vectors, or entity vectors, associated with physical entities, which can be used for various analytical purposes like classification, rating, and real-time geo-fencing, enabling customized map displays and user-specific content targeting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If crowdsourced sensor data from multiple wireless devices is collected and processed, then the quantity and diversity of data increases, but the complexity of data processing and analysis increases
Solution Approach 1:
The patent segments the crowdsourced sensor data into multiple feature vectors, each representing different aspects or dimensions of the physical entity. This segmentation allows the system to process and analyze different types of sensor data (acceleration, gyroscope, magnetometer, barometer) separately and then integrate them, reducing the overall processing complexity while maintaining data quantity and diversity.
Solution Approach 2:
The patent introduces feature vectors as intermediary representations between raw sensor data and the final entity model. These feature vectors serve as a standardized intermediate format that simplifies the integration and analysis of heterogeneous sensor data from multiple sources, reducing the complexity of direct processing.
2Loss of information
If user interaction history data is integrated with sensor data, then the comprehensiveness of the entity model increases, but the difficulty of data analysis increases
Solution Approach 1:
The patent merges sensor data and user interaction history data into a unified feature vector representation. By combining these different data types into a single integrated structure, the system maintains comprehensive entity information while simplifying the analysis process through unified processing rather than separate analysis of multiple data sources.
Solution Approach 2:
The patent transforms heterogeneous data from different sources (sensor readings and user interactions) into a standardized parameter format through feature vectors. This parameter transformation allows diverse data types to be analyzed using consistent methods, reducing the difficulty of data analysis while preserving information completeness.
3Measurement precision
If multiple analysis models are applied to generate entity vectors, then the accuracy of physical entity classification increases, but the computational processing time increases
Solution Approach 1:
The patent applies multiple analysis models in a sequential pipeline where earlier models perform preliminary classification and filtering. This preliminary action reduces the complexity and processing time required by subsequent models, as they only need to refine rather than perform complete analysis from scratch, thereby reducing overall computational time while maintaining accuracy.
Solution Approach 2:
The patent implements a dynamic multi-model analysis approach where the selection and application of analysis models can be adjusted based on the specific characteristics of the data and the classification task. This dynamic approach allows the system to use simpler models when sufficient and more complex models only when necessary, optimizing the balance between accuracy and processing time.
Data Source
AI summary
A method for generating an entity-vector associated with a geographical location is disclosed. The entity-vector is generated based on data from sensors of user-wireless-devices and user profile associated with user-wireless-devices. The method comprises, for a given location: receiving from user-wireless-devices being temporally-associated with the location, first and second sensed parameters from first and second sensors, the sensed parameters having been captured during the user-wireless-device being temporally-associated with the location; for each user-wireless-devices, retrieving an associated user profile and analyzing sensed parameters and user profile to generate, respectively, a given-user first and second parameter vectors and a given-user profile vector; aggregating, each user-wireless devices' (i) given-user first parameter vector, (ii) given-user second parameter vector, and (iii) the given-user profile vector into a pool; applying at least two analysis models to the pool, each analysis model generating its respective pooled-data vector; based on pooled-data vectors, generating an entity-vector associated with the location.