Rank-Based Fingerprinting for Wireless Device Location
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
GPS-based positioning techniques are ineffective in indoor or dense urban environments due to satellite signal attenuation, and fingerprinting methods face challenges with positioning errors caused by hardware differences between wireless devices used during training and location phases.
Innovation Solution
Implementing rank-based signal value difference and corresponding difference deviation techniques to compensate for hardware differences by preprocessing and post-processing signal values during fingerprint matching, using techniques like Gaussian similarity calculations and weighted centroid location determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If GPS-based positioning techniques are used, then positioning can be performed in open environments, but positioning becomes ineffective in indoor or dense urban environments due to satellite signal attenuation
Solution Approach 1:
The patent introduces wireless access points and their signal measurements as an intermediary system to enable positioning when GPS satellites are blocked. Instead of directly receiving satellite signals, the device uses intermediate wireless signals from access points to determine location through fingerprinting comparison.
2Reliability
If fingerprinting methods are used for indoor positioning, then positioning can be achieved in indoor environments, but positioning errors occur due to hardware differences between wireless devices used during training and location phases
Solution Approach 1:
The patent transforms the raw signal strength measurements into rank-based ordinal parameters. By converting continuous signal strength values into ranked positions (1st strongest, 2nd strongest, etc.), the system changes the parameter representation to be invariant across different hardware devices, thereby eliminating hardware-specific measurement variations.
Solution Approach 2:
The patent creates a hardware-agnostic copy of the positioning data by using rank-based representations instead of device-specific signal measurements. This copied representation can be universally applied across different wireless device types without requiring device-specific calibration or training data.
Data Source
AI summary
In some examples, a query including signal values providing measures of signals between a first wireless device and corresponding access points is received. Access points are ranked for the query based on sorting the signal values. Similarity values are computed, using the ranking, between the query and respective fingerprints of signal values collected at a plurality of locations in an area, each signal value of the signal values in the fingerprints providing a measure of a signal between at least one wireless device at a location of the plurality of locations and a respective access point of the access points. A location of the first wireless device is determined based on the computed similarity values.