Browse Prior Art Database

Automatic detection of building access

IP.com Disclosure Number: IPCOM000249343D
Publication Date: 2017-Feb-17
Document File: 3 page(s) / 88K

Publishing Venue

The IP.com Prior Art Database

Abstract

Determining whether somebody entered a building can be useful for several reasons such as targeted ads, display of welcome messages with information about the building, security concerns. This idea describes a system that uses existing personal and company’s technical infrastructures such as a phone and WiFi, to detect whether a person entered a building through application of dynamical machine learning to signals available on a person's phone.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Automatic detection of building access

Determining whether somebody entered a building can be useful for several reasons such as targeted ads, display of welcome messages with information about the building, security concerns. This idea describes a system that uses existing personal and company’s technical infrastructures such as a phone and WiFi, to detect whether a person entered a building through application of dynamical machine learning to signals available on a person's phone.

The novel aspect of this idea is that it provides a procedure to determine whether a person entered a building in a way that only relies on them having a phone and does not require any additional infrastructure within said building or on the person. Furthermore, this idea describes a flexible cognitive system that leverages all the available information, but is not bound to the availability of all the information.

The value of this idea is in the fact that it would allow to send targeted notifications to people who are more likely to be interested in receiving them (people who enter a building as opposed to a random passer-by), and could be of interest for any company with stores accessible to the wide public. Other uses include monitoring flow of people through a building, and presenting them with meaningful information as they enter it. The idea works as follows:

Upon measuring, e.g. through GPS, that someone is in the proximity of the entrance of a building, activate a procedure on a person's phone which starts gathering meaningful measures to determine if and when they enter said building. Examples include:

- phone signal: upon entering a building it is likely to drop slightly - WiFi signal: upon entering a building, likely to go up significantly if WiFi is

available within the building itself - microphone signal: can detect sounds that are typical of the building

entrance such as supermarket sliding doors or music - camera signal: if held in hand (outside of pocket) could detect landmarks - light sensor: measure change in illumination levels (again if phone is

outside pocket) - barometers: measure fluctuations in pressure related to changes in

temperature - magnetometers: measure disturbances in magnetic field and relates them

to surrounding structures. Combine the measurements above (the ones that are available) and apply

dynamical machine learning to them. Each building entrance will have a distinctive pattern in terms of how those signals vary upon entering a building. Thus machine learning will be carried out on a building-specific (or better entrance-specific) basis: if the signals available for a person evolve in a way that is similar to what is expected for th...