LocaRDS: A Localization Reference Data Set.

Area of Science:

• Computer Science
• Electrical Engineering
• Geospatial Science

Background:

• Wireless signal localization is crucial for applications like radar and satellite navigation.
• Existing research lacks standardized methods for real-world performance assessment and comparison of localization techniques.
• There is a need for a robust scientific methodology to evaluate and compare diverse localization solutions.

Purpose of the Study:

• To establish a solid scientific foundation for localization research.
• To introduce LocaRDS, an open reference dataset for crowdsourced flight data.
• To enable rigorous testing, analysis, and direct comparison of localization methods.

Main Methods:

• Developed LocaRDS, an open dataset comprising over 222 million real-world crowdsourced flight measurements.
• Collected data from over 50 million transmissions recorded by 323 sensors.
• Demonstrated methods for verifying data quality within LocaRDS.

Main Results:

• LocaRDS provides a verified dataset suitable for scientific evaluation of localization techniques.
• The dataset facilitates direct comparison of different localization algorithms.
• An example implementation for aircraft localization and relevant metrics were provided.

Conclusions:

• LocaRDS significantly advances the methodology for localization research by providing a standardized, real-world dataset.
• The dataset enables objective performance assessment and comparison of wireless localization methods.
• Future investigations in localization can leverage LocaRDS for enhanced scientific rigor.