RSS and inertial navigation based indoor localization

• Main Author: Liu, Bowen
• Other Authors: Tan Soon Yim
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Electrical and electronic engineering::Wireless communication systems
• Online Access: https://hdl.handle.net/10356/171915
• Institution: Nanyang Technological University

With the rapid evolution of mobile Internet and mobile terminal equipment, the demand for location-based services is becoming increasingly robust. In addition, indoor activities have taken up most of people's time in the day, the importance of indoor positioning is constantly gaining attention. Since satellite signal is not available indoors, indoor positioning is also known as the last mile of positioning and navigation. This dissertation improves the WiFi/PDR-based indoor localization technique, specifically, by predicting the location by INS data and generating the corresponding predicted Received Signal Strength (RSS), which is then weighted with the observation to obtain the robust RSS and use it as a basis for target location determination based on optimization methods. Furthermore, sliding window filtering algorithm averaging the adjacent data is leveraged to mitigate the INS noise, it can increase the perception ability of IMU data to reflect pedestrian movement. We find that when the window value is 50, it can effectively denoise INS data, and the smoothed data can better perceive pedestrian walking cycles. Furthermore, the experimental results in this dissertation illustrate that the INS-based positioning method is able to provide more accurate position estimation results with higher weights in the early stage of pedestrian movement, while the WiFi-based position estimation results increase in weights as time increases.