Page 66 - Undergraduate Research Opportunities Program Proceedings 2017-18

School of Engineering
Department of Computer Science & Engineering

Supervisor: CHAN Gary Shueng Han / CSE
Student: WONG Tsz Nok / SENG Course: UROP1100, Summer

Indoor localisation service has been using different signals to overcome low precision locating by the Global Positioning
System. Wherami is a project targeting at providing dependable techniques for mobile localisation. This UROP research
project is to assist the iOS development with the Research & Development team. The major tasks of it are to create a
map view framework for the project and to debug iOS existing applications. Influencing technical and personal
development, valuable learnings and skills are earned from both tasks. It is concluded that this UROP project is a
precious opportunity to participate in a technology-based project and gain incisive viewpoints.

Supervisor: CHAN Gary Shueng Han / CSE
Student: ZHANG Luoshu / COMP Course: UROP1000, Summer

Indoor localization is a technology where numerous sensors and communication technology is utilized to detect the
locations of targets in indoor environment, which is widely applied in various aspects of our daily life, such as logistic
tracking, shopping mall map design, etc. However, the prevailing localization method using GPS cannot be accessed
effectively in the indoor environment. To remedy the situation, tracking the user’s location from the wireless signal is
utilized. In this research project, the data from the Wi-Fi signal shall be exploited to detect the user’s location. To be
more specific, a convenient and responsive method that extracts data from a set of routers and compute the location
through a delicate algorithm is proposed to improve the situation.

Supervisor: CHAN Gary Shueng Han / CSE
Student: ZHAO Ziqi / COSC Course: UROP2100, Summer

Indoor crowdsourcing-based localization solutions have been under development for a few years, but most of them use
the prior knowledge of the building (e.g. floor plan or Wi-Fi AP locations) to build the map. Therefore, we propose a new
way to construct without knowing building information. This method is based on IMU (Inertial measurement unit)
sensors and Wi-Fi scanning on the smartphone. The phone first collects accelerometer, gyroscope, gravity,
magnetometer and Wi-Fi signals, and send them back to the server. Then backend program will recover the path based
on sensors, segment path using turnings, and use signal similarity to concatenate different segments. Results shows that
this method could generate a satisfying pathway map, but related work is still needed to reduce the complexity and
improve adaptability.

Supervisor: CHAN Gary Shueng Han / CSE
Student: ZHU Xinyu / COSC Course: UROP4100, Fall

In this semester, I worked on the mobile sensor-based tracking system (Mosent) with two Ph.D. students and I
contributed to this project by doing literature survey, implementing some localization algorithms with Python,
conducting experiments and writing the paper. Mosent is a novel and potent tracking system based a series of advanced
algorithms such as Forward Dependent Particle Filter (FDPF) and Propagation Loss Matrix Learning (SCAL). It can track
multiple mobile targets in a large area with only a few of sensors and can handle the environmental influence on signal
propagation and influence from sensor location error. Our experimental results show that the Mosent can achieve high
accuracy with fewer sensors compared with the state-of-the-art.

56