Satellite positioning works best under open sky. Inside buildings, satellite signals are weakened or obstructed, so indoor positioning has to rely on other information — usually on more than one kind of it. KAILOS is a technology stack built around that idea.
What the stack brings together
KAILOS combines the radio signals already present in a building, a description of the building itself, the motion sensors in a device and positioning algorithms that turn those measurements into a location estimate. Published work described KAILOS as an integrated indoor positioning system at IPIN 2014, and later research expanded its radio-map, sensor-fusion and device approach.
- Wireless environment — Wi-Fi and Bluetooth signals observed by a device.
- Building information — registered buildings, floor plans and indoor maps.
- Device sensors — accelerometer, gyroscope and magnetometer measurements that describe movement and heading.
- Positioning algorithms — hybrid positioning with signal filtering, Kalman filtering and map matching.
Wi-Fi fingerprinting in plain language
Every Wi-Fi access point broadcasts a signal. At any spot in a building, a device observes a particular combination of access points and signal strengths. That combination behaves like a fingerprint of the location.
Fingerprint-based positioning works in two phases. First, measurements are collected at known locations and stored as a radio map. Later, a device compares what it currently observes with the radio map and estimates where it is most likely to be.
Building and maintaining the radio map
The radio map is the foundation of fingerprint positioning, and it has to exist before positioning can begin. KAILOS includes tools for this work: a web-based map workflow for registering buildings and floor plans and planning surveys, and walking-survey workflows that capture Wi-Fi and sensor measurements as a surveyor moves through the space.
Radio environments change as buildings are renovated and networks are reconfigured. The research behind KAILOS therefore also explored crowdsourced approaches, in which measurements contributed during normal use help construct and maintain radio maps.
Why hybrid positioning
No single signal is dependable everywhere indoors. Wi-Fi coverage varies, Bluetooth depends on the devices deployed in a space, and motion sensors accumulate error over time. KAILOS combines these sources and applies filtering — including Kalman filtering and map matching against the indoor map — to produce a steadier estimate than any one source provides alone.
From a position to an application
A location estimate is only useful when an application can act on it. KAILOS was designed as a location-service platform rather than a standalone algorithm, with interfaces for indoor map information, routing and other location-aware functions. That platform layer is what allows indoor positioning to support navigation, safety and tracking applications.
See how the platform is deployed on the KAILOS Wi-Fi Positioning page, and follow the step-by-step workflow in the KAILOS Manual.