SIG Sensing

SIG Chair: Shao-Yi Chien 

-Professor, Dept. of Electrical Engineering, National Taiwan University

Four sub-projects pertain to SIG Sensing:

 (1)Distributed Intelligence 

-PI: Shao-Yi Chien, CoPI: Yu Tsao

The main objective of this project is to adopt state-of-the-art deep learning algorithms to improve the performance and overcome the obstacles of the internet of things (IoT) and augmented collective beings (ACB) framework. We select object tracking across a camera network and audio signal processing as two application examples for our developed distributed intelligence techniques, including CNN model optimization and CNN hardware architecture design.

(2)Next-Generation CamCom 

-PI: Michael Hsin-Mu Tsai, CoPI: Kate Ching-ju Lin

Camera communications (CamCom) has the capability to separate signals from transmitters at different locations due to its high spatial resolution, yet the low frame rate of commodity cameras limit the system data rate and hence the applications it can support. In this meeting, we will report a solution that possess the spatial filtering capability of the camera while improving the throughput. The solution centered around a dynamic field-of-view (FoV) VLC receiver, where the size and location of the FoV of the receiver can be dynamically adjusted in real time, locking to the location of the transmitter and blocking interference from other location.

(3)Enhancing Security and Privacy in Augmented Collective Beings 

-PI: Hsu-Chun Hsiao

This year(2018) our research focuses on enhancing user privacy when Internet-of-Thing (IoT) devices communicate with untrusted clouds. The proposed privacy-enhancing mechanisms can not only protect IoT users, but also help IoT service providers comply with increasingly rigorous privacy regulations such as the EU General Data Protection Regulation (GDPR). We summarize our proposed solutions that protect users’ privacy against cloud-based IoT automation services and anomaly detection services.


-PI: Daniel Chih-Sheng Shih, CoPI: Kwei-Jay Lin

One issue for self-navigating moving robot is the self-localization. Robots need to know where they are so that they can conduct location-dependent services such as navigate to a specific location to access some resource or deliver a service. Although many technologies have been proposed to provide indoor localization, factors such as cost, simplicity, accuracy and adaptability make them unsuitable for moving robot applications in smart factory, smart warehouse and smart office.