Abstract: 

Magnetic information is frequently used in combination with other inertial data for orientation purposes in navigation systems and vehicle localization. The present invention introduces an outdoor localization system based on magnetic landmarks. These landmarks are magnetic anomalies that distort the expected geomagnetic field due to strong local magnetic fields oftentimes originated by fixed man-made structures. We propose to leverage them as signatures for localization purposes

Background: 

The most common systems for outdoor positioning use GNSS (e.g. GPS) and / or inertial data. Satellite-based solutions have strong limitations (exhibit large positioning errors or outages) in multipath environments and non-line-of-sight conditions to the satellites, such as urban canyons, tunnels, and underground parking. They are often combined with other sources on information to complement them, such as inertial sensors.

Inertial data is frequently collected for navigation purposes . This system applies dead reckoning and presents significant cumulative errors. The quality of the sensors has a high impact on the accuracy of the estimation, but also influences their price.

Benefits: 

The present invention uses a mobile device equipped with intertial sensors and magnetometer to detect and match the magnetic anomalies to references, providing location data. The proposed method is projected to technically support current localization systems, by allowing to improve their efficiency. Particularly, dead reckoning, which greatly suffers from the impact in orientation estimation of geomagnetic field anomalies created by local magnetic fields, and satellite-based systems, which have strong limitations in dense infrastructure areas due to multipath and lack of visibility to satellites.

The main contributions of this work are:
 
1) A method for detecting the magnetic anomalies in this scenario;
2) A system for collecting and saving classified instances for reference;
3) An algorithm that matches input magnetic anomalies to the known stored references, both for complete and partial time series.

Potential comercial use/applications: 

This techonology may result in a localization system in which the GPS is not available, not accurate enough, or malfunctions. It may also be turned into a modulo to improve the robustness, availability and accuracy of a localization system, improving its resilience to disturbances in GPS or cumulative errors.