System and method for optical localization

A method of detecting a location of a mobile unit comprising:

• using a light sensor, obtaining a plurality of pieces of spectral information of visible light the mobile unit is exposed to, the spectral information being obtained at a limited number of predetermined wavelengths, the mobile unit being in an environment lit by a plurality of light sources;
• generating a location estimate by inputting, as visible light information, the obtained plurality of pieces of spectral information into a trained machine learning model, wherein the mobile unit is configured to determine a signature of a light source independent from a distance of an angle between the light source and the sensor;
• and
• estimating a current location of the mobile unit as one of being at a known installation location of the identified light source or being proximate to the known installation location of the identified light source,
• wherein the signature is calculated by determining at least one of a ratio of powers provided by the light source at two or more of the predetermined wavelengths, or a difference of the powers provided by the light source at two or more of the predetermined wavelengths.

A method of training a machine learning model comprising:

• training a model capable to provide localization information based on radiofrequency measurements further using visible light information detected at a location and known location information of the location at which the visible light information was detected, the radiofrequency measurements being detected on tracked locations of a mobile unit, the visible light information including spectral information of visible light the mobile unit is exposed to, the localization information being derived based on a tracked current location of the mobile unit and from information of a known location of installation of an identified light source,
• wherein the mobile unit is configured to determine a signature of a light source independent from a distance or an angle between the light source and the sensor, compare the determined signature to previously stored signatures of the light sources and identify a light source having a signature that has a minimum error relative to the determined signature in the stored signatures, and estimate a current location of the mobile unit as one of being at a known installation location of the identified light source or being proximate to the known installation location of the identified light source, and
• wherein the signature is calculated by determining at least one of a ratio of powers provided by the light source at two or more of the predetermined wavelengths, or a difference of the powers provided by the light source at two or more of the predetermined wavelengths.

A mobile unit comprising a light sensor, the mobile unit configured to:

• obtain a plurality of pieces of spectral information of visible light the mobile unit is exposed to, the spectral information being obtained at a limited number of predetermined wavelengths, the mobile unit being in an environment lit by a plurality of light sources;
• determine a signature of a light source independent from a distance or an angle between the light source and the sensor;
• compare the determined signature to previously stored signatures of the light sources and identify a light source having a signature that has a minimum error relative to the determined signature in the stored signatures;
• and
• estimate a current location of the mobile unit as one of being at a known installation location of the identified light source or being proximate to the known installation location of the identified light source,
• wherein the signature is calculated by determining at least one of a ratio of powers provided by the light source at two or more of the predetermined wavelengths, or a difference of the powers provided by the light source at two or more of the predetermined wavelengths.

A mobile unit comprising a light sensor and configured to:

• obtain a plurality of pieces of spectral information of visible light the mobile unit is exposed to, the spectral information being obtained at a limited number of predetermined wavelengths, the mobile unit being in an environment lit by a plurality of light sources;
• generate a location estimate by inputting, as visible light information, the obtained plurality of pieces of spectral information into a trained machine learning model;
• determine a signature of a light source independent from a distance or an angle between the light source and the sensor;
• and
• estimate a current location of the mobile unit as one of being at a known installation location of the identified light source or being proximate to the known installation location of the identified light source,
• wherein the signature is calculated by determining at least one of a ratio of powers provided by the light source at two or more of the predetermined wavelengths, or a difference of the powers provided by the light source at two or more of the predetermined wavelengths.