Wi-Fi Localization Using Multi-Parameter Fingerprinting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wi-Fi localization systems face inaccuracies due to signal fluctuations and multipath issues indoors, where GPS and GLONASS are inadequate, leading to errors in user positioning.
Innovation Solution
A computer program executed by processors that generates an electromagnetic signal map by measuring location-dependent measurements from access points, using multi-objective optimization and clustering to determine the device's position, incorporating algorithms like LASSO, GLMNET, and GS for improved accuracy and outlier detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If signal intensity measurement and fingerprinting are used for Wi-Fi localization, then positioning capability is provided in indoor environments where GPS is inadequate, but measurement precision deteriorates due to signal fluctuations and multipath issues
Solution Approach 1:
The system performs preliminary fingerprinting during an offline phase, storing signal characteristics (RSSI, angle of arrival, time of arrival) at known locations before actual positioning is needed. This pre-collected data serves as a reference database that compensates for signal fluctuations during online positioning, improving accuracy without requiring real-time recalibration
Solution Approach 2:
The system transitions from using single parameter (signal intensity) to multi-parameter fingerprinting including RSSI, angle of arrival, and time of arrival. This parameter diversification makes the localization system more robust against signal fluctuations and multipath effects, as multiple parameters provide redundant information that can be cross-validated
2Measurement precision
If the number of positions in the database is increased to improve localization accuracy, then measurement precision improves, but device complexity and computational resources increase
Solution Approach 1:
The system extracts and stores only the most distinctive and informative signal characteristics (fingerprint features) at each location, rather than storing complete raw signal data. This selective extraction reduces database size while maintaining localization accuracy by keeping only the essential features needed for position determination
Solution Approach 2:
The localization space is divided into discrete positions or regions, each with its own fingerprint profile. This segmentation allows the system to build a manageable database of representative positions rather than continuous space, reducing complexity while maintaining accuracy through the discrete position model
3Ease of manufacture
If traditional fingerprinting methods are used without optimization, then implementation is simple, but localization accuracy deteriorates due to errors from signal fluctuations
Solution Approach 1:
The system implements feedback mechanisms where the measured signal characteristics are compared against the stored fingerprint database, and the position estimate is refined iteratively. This feedback loop allows the system to correct for signal fluctuations by continuously comparing actual measurements with expected fingerprints from the database, improving accuracy while maintaining relatively simple implementation
Data Source
AI summary
Disclosed are various embodiments that enable the identification of the location of a computing device based on radio data. A radio map can be identified for an area. The computing device can measure signal strengths to reference points. The signal strengths can be compared to the radio map. The computing device can determine its location based on the comparison of the signal strengths to the radio map.