The use of artificial intelligence (AI) that automatically labels wireless LAN fingerprint connection locations is the key behind the new technology. (image: KAIST)
DAEJEON, Jun. 11 (Korea Bizwire) — A technology that can pinpoint the exact location of a user by collecting wireless signals from smartphones has been developed.
The Korea Advanced Institute of Science and Technology (KAIST) announced on Monday that a research team led by Prof. Han Dong-soo has developed technology that can be used to recognize indoor locations based on crowd sourcing.
The use of artificial intelligence (AI) that automatically labels wireless LAN fingerprint connection locations is the key behind the new technology.
Wireless fingerprint refers to information regarding the strength of a wireless LAN signal in a specific location.
Wireless signals collected through an unspecified number of smartphones are divided by buildings through clustering. The signals are then subdivided by each floor of the building using atmospheric pressure information.
As long as one is within a building with a wireless LAN signal, the new technology can be applied to any building.
Pedestrian Dead Reckoning (PDR) also figured in the development of the new technology. PDR is a method to calculate location by collecting and processing one’s speed, direction, and distance of movement.
Based on wireless signals that are obtained from inertial sensors, the technology undergoes a process of machine learning related to a specific area or region to optimize the location of data collection.
The research team measured the accuracy at an indoor shopping mall with a capacity of 400,000 square meters from the second basement floor to the sixth floor, and showed an accuracy of 3 to 6 meters.
The technology was able to pinpoint the floor a user was on to an accuracy of 95 percent.
The team predicted that using the new technology for signals collected through IT companies, telecommunications and online shopping mall applications will allow it to establish an indoor location infrastructure on a city or national scale.
Kevin Lee (kevinlee@koreabizwire.com)
