There is no English translation for this web page.
Ultrasonic Indoor Location Identification using a Smartphone
LOCATION: TEL, Auditorium 1 (20th floor), Ernst-Reuter-Platz 7, 10587 Berlin
Date/Time: 03.07.2017, 15:00-15:45
SPEAKER: Fabian Knappe
The presented thesis deals with ultrasonic indoor location identification using a smartphone. Indoor localization systems often need external equipment and/or are itself integrated in special hardware devices. For that reason a completely autarkic indoor localization system at room level accuracy, which can be integrated in an off-the-shelf smartphone, is realized. The system is based on active acoustic fingerprints in the ultrasonic frequency range.
An Android application is developed to record acoustic room impulse responses in two indoor environments, consisting of the two floors of a house. Active acoustic fingerprints of each room are created by extracting RT60 and MFCC features from the impulse responses. One-class SVMs are used to estimate the current location. Additional sensor data of the smartphone are used to raise the robustness of the localization system.
Different random motion paths through both indoor environments are created and the room estimation is performed for each test sample. An over-all accuracy of the indoor localization system of 80% correct estimated rooms is reached.
In the attachment you find a image representing the work.