Browse Prior Art Database

Method for Indoor Localization using Crowd-created Signal attenuation Patterns Disclosure Number: IPCOM000240408D
Publication Date: 2015-Jan-29
Document File: 3 page(s) / 69K

Publishing Venue

The Prior Art Database


The following document proposes a methodology which aims to improve resolution of WiFi based indoor localizations using crowd sensed WiFi attenuation signatures

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 3

Method for Indoor Localization using Crowd -created Signal attenuation Patterns

Disclosed is a method to improve accuracy of indoor localization using two sets of WiFi fingerprints - 'steady state fingerprint' and 'crowd-state fingerprint'. The method relies on finding unique attenuation patterns for micro grids based on crowd states in nearby grids.

Indoor positioning technologies present a thriving business in the coming years. For e.g. Hospitals: provide navigation aid, Retail: better marketing to customers, Tourism: provide information via audio for tours, offer video or augmented reality experiences. Given an indoor infrastructure scenario e.g. malls, airports etc., the aim is to devise a methodology to predict the location of a user within the premises with tolerable accuracy. A tolerable accuracy is subject to the context of deployment. For a shopping mall accuracy may be comparable to size of shops while other use cases may need higher accuracy.

Indoor localization solutions divide a physical space into discrete grids and predict the most probable grid the user can be present in using WiFi signature of the space. A WiFi signature/fingerprint is a vector of Signal strengths from all visible AP's at any particular location. Multiple factors influence the resolution of these grids - very fine grids would increase computational cost and may not show significant difference in WiFi signatures, coarse grids may not hold utility for the targeted use case. Given a grid let us associate two parameters with it - Uncertainty - this is the resolution of the grid and Recall - this is the probability of accurate prediction for that grid. Large grids will tend to have higher recall because they capture a large variation of RSSI but will have a higher uncertainty whereas smaller grids will be susceptible to dynamic attenuation in WiFi signals and hence will have lower recall.

Two important definitions will help formalize the proposed solution -

      Micro/fine grid - The grid which has a tolerable uncertainty with respect to the use case

      Macro/coarse grid - The grid encompassing multiple micro grids which has a higher recall than most of the micro grids

The objective is to minimize the uncertainty of prediction by localizing users to micro grids while having a high recall. A two fold approach is henceforth proposed (refer Fig. 1) -

       Initial prediction - A vanilla algorithm predicts a macro grid with an associated uncertainty and a good recall

       Refinement - Given this macro decrease the uncertainty in the predicted location while not compromising with the associated recall i.e. accurately predict a micro location within the macro location

The challenge is to distinguish between the micro grids inside a macro grid. Experimental results show that human bodies are good attenuators o...