UROP Proceedings 2022-23

School of Engineering Department of Computer Science and Engineering 92 Indoor Localization and Mobile Computing Supervisor: CHAN, Gary Shueng Han / CSE Student: DING, Yiyi / COMP Course: UROP3100, Fall UROP4100, Spring With the development of modern technology, both indoor and outdoor localization techniques now have low computational time. This makes pervasive positioning—the seamless positioning of an object anywhere on a national scale—feasible. However, compared to the outdoor localization techniques which can be supported by the Global Navigation Satellite System (GNSS) only, indoor localization has different techniques which require different hardware, protocols, and data to provide the results to the user. Currently, there are proximity-based indoor localization which requires iBeacon data, and fingerprint-based indoor localization which requires Wifi fingerprint data. Hence, to ensure seamless property during the transitioning of positioning techniques, a standard that serves as a bridge between places that deploys different localization methods should be developed. Indoor Localization and Mobile Computing Supervisor: CHAN, Gary Shueng Han / CSE Student: JU, Jongho / COSC Course: UROP1100, Fall Pervasive positioning locates an object anywhere on a country scale by providing application requests location services from the correct parties using the correct formats. To bridge every party together, a software development kit (SDK) provides an interface to allow the application and the server to communicate when detecting the location of the object. The SDK provides the facility of handshaking, downloading site signals, and 4 modes of the protection level from mode 0 to mode 3. Among these modes, operation mode 2 is developed to provide platform-supported edge localization in case of the site owner shares signals but application users do not. To implement in mode 2, SDK will request site signal from the server and compute the location locally with the aid of the application interface. Indoor Localization and Mobile Computing Supervisor: CHAN, Gary Shueng Han / CSE Student: TAN, Kee Meng / SENG Course: UROP1000, Summer This project explores how one can preprocess data for Espresso, a novel approach to associate Wi-Fi probe requests under MAC address randomization while protecting user privacy. Espresso works by comparing several variables, namely timestamp, MAC address, information element, sequence number and received signal strength, to estimate the correlation between frames. These data have to be extracted from the packets collected from access points, and we implemented several ways to clean and preprocess the raw packet capture data. After filtering irrelevant packets, merging packets from multiple sensors and sorting based on the timestamp, we can then convert these packets into a CSV file, which then can act as the input for Espresso.