Page 63 - Undergraduate Research Opportunities Program Proceedings 2017-18

School of Engineering
Department of Computer Science & Engineering

potentials. This report focuses on indoor tracking technique that solely depends on data from user’s smartphone for the
entire course of map construction and user tracking. These data include Wi-Fi fingerprints as well as data from built-in
physical sensors in the smartphone. Accompanying a smartphone app that collects and sends the data behind the user’s
back at-the-spot, our project aims an efficient and accurate map construction via crowdsourcing. With the complete
floor plan set as a reliable constraint, the next step for our project would be to utilize Conditional Random Field model
to provide reliable, real-time location information.

Supervisor: CHAN Gary Shueng Han / CSE
Student: HUANG Zijian / COSC Course: UROP1100, Spring
ZHANG Lin / COSC UROP1100, Spring

Indoor localization has been a hot topic in recent years, leading many researchers to develop different kinds of
localization methods using different sensors. This report will mainly discuss our implement of indoor localization using
beacons and Android devices. We will illustrate how we choose our hardware devices and software platforms, and the
structure of our system, as well as our mathematical foundations. The system we implemented is not so complex, we
only implemented the skeleton structure of the system, but the function is completed and can be used for further
developing. The structure basically is a server-client system structure and will track the users and save the users’
location into our database. Further improvement will include using more advanced localization algorithm, adding more
functions to our front-end page, etc.

Supervisor: CHAN Gary Shueng Han / CSE
Student: MENG Zihan / COMP Course: UROP1100, Fall

Smartphones have numerous sensors implemented and are now ubiquitously used. These sensors will have the
capability to provide more information about the user with proper use. Especially for step counting use, if combined
with indoor location technology, we will be able to improve the accuracy of the outcomes. We have studied some
algorithms for step counting taking advantage of the accelerometers in modern smartphones. For most of the papers
currently available about step counting methods, they evaluate step counting algorithms using independent sensors and
have these sensors fixed to only some body part. Our experiment will use a common smartphone to test the algorithm,
this algorithm will be implemented as a mobile application and installed on a smartphone. Also, the smartphone will be
used naturally instead of being attached to some body part.

Supervisor: CHAN Gary Shueng Han / CSE
Student: MU Yifan / COMP Course: UROP1100, Fall
UROP2100, Spring

With the development of mobile devices and other location-based services, there are a large number of datasets of
individual location information is available nowadays, which provides us an opportunity to infer people’s social ties
through analysis on these datasets. Several models or algorithms to infer social strength by analyzing co-occurrence
information and trajectory similarity have already been introduced in some pervious papers. In this project, we are
trying to improve current algorithms to measure trajectory similarity and then figure out social ties according to
trajectory similarity first, and combine co-occurrence information with our algorithm.

53