Study on WiFi-based Indoor Positioning Prediction using Machine Learning Techniques

Kyounghyun Park, Giyoung Lee, Yangkoo Lee, Daesub Yoon, Seonghun Seo, JaeJun Yoo, Min Jung Kim

Abstract: With the proliferation of smartphones and advancements in artificial intelligence, WiFi-based indoor positioning technology continues to evolve. The fingerprinting approach generates a fingerprint map using RSSI (Received Signal Strength Indicator) from WiFi Access Points (APs) for indoor positioning. However, WiFi signals are susceptible to environment, leading to the challenge of rebuilding the fingerprint map whenever the environment changes. Machine learning techniques can overcome the drawback of the fingerprint method and therefore enhance indoor positioning accuracy. Ultimately, machine learning techniques can improve the accuracy, cost-effectiveness, and scalability of the indoor positioning system compared to traditional statistical methods. In this paper, we examine representative machine learning algorithms applicable to indoor positioning system and discuss the performance of the algorithms.