Networking and Mobile Systems Laboratory (NMSL) performs research on every networking aspect that provides reliable, secure, seamless and high capacity connectivity to mobile users. We cover end-user computing, infrastructure networks, and edge networking. To improve the quality of life of mobile network users, we (i) identify challenging real-world problems, (ii) design novel network architecture, protocols, systems and applications, and (iii) build our solutions in working systems for practical validation.

Research interests

Our current interests include mobile computing, wireless networking (802.11ac/ax, visible light communication, inaudible sound communication), smart sensing systems, IoT, mobile security, mobile HCI, cloud networking, and wireless big data analytics. We are also interested in interdisciplinary, high impact research, and seek collaboration with other academic research groups, industry, and government worldwide.

Wireless Networking

- Wi-Fi, Bluetooth, Visible light communication
- Drone/smartphone networking
- Wireless RF analytics
- Cloud-based Wi-Fi management

Mobile Computing

- Novel mobile applications
- Mobile sensing systems
- Energy efficiency
- Context awareness

Mobile HCI

- Social problems caused by mobile devices
- Quality of life of mobile users
- Family interaction applications


- Protocol co-existence
- Cross protocol communication
- IoT gateway design

Mobile Security

- Drone security
- IoT security
- Wireless jamming mitigation

Cloud Networking

- Cloud computing for mobile services
- Cloud architecture for WLAN control



Text messaging is still the most popular app for smartphones. Beyond simple text, emojis and image can be used to deliver emotions, nuances, and detail information. However, selecting an appropriate emoji from 1,500-plus emojis could be time consuming. Similarly, importing an image from the web requires switching applications. Hence we develop MilliCat, a real-time emoji and image suggestion service for mobile chat applications. We use deep learning and natural language processing techniques to understand the user's chat context. MilliCat provides an emoji set sorted based on chat context. MilliCat also suggests images for words that would be better expressed with visuals.

Secure IoT Gateway

Internet of Things (IoT) technology has the potential to be applied on industrial environments beyond households. While IoT enables new applications and services in distributed systems, it is crucial to make IoT devices secure. In this project, we aim to build a virtual gateway to centralize the network policies using SDN. The IoT specialized SDN gateway can inspect packets based on network fingerprinting, apply security policies with ease, and control traffic congestion.


We aim to improve the throughput experienced by user devices to Gb/s in highly dense Wi-Fi networks. Previous research has primarily focused on increasing the capacity of wireless APs (Access Points), not the throughput of the users. We instead, take a network systematic approach to analyze the gap between the theoretic capacity and practical throughput, and design algorithms that bridge this gap. We devise and implement algorithms that effectively and efficiently select channel width, modulation, transmission rate and mode, and evaluate them on our testbed in various network scenarios.

Context-aware Smartphone Notification System

A push notification on smartphones (not just text messages, but notifications from various apps) is sometimes seen as an interruption rather than a useful piece of information. In particular, interactions in offline social gatherings are often interfered by unregulated smartphone usage. To alleviate this social interruption, we are building a system that intelligently pushes notifications on smartphones so that people can better focus on real life social interactions. Our system detects a situation in which the user interacts with others in a social gathering. Push notifications are then delivered in a batch at social breakpoints, which we define as a time when the user is not engaged in a conversation.

Versatile Network System

The number of mobile devices on the network is growing every day, and we are witnessing new applications and services emerge with various QoS requirements. The future Internet must address multi-dimensional diversity; diversity in end devices, communication protocols (e.g., cellular, Wi-Fi, Bluetooth, ZigBee, visible lights, inaudible sounds, etc), and application requirements, in addition to the increase in the number of network devices, thanks to IoT. To this end, we devise a network architecture where we bring the intelligence to the edge. Through edge computing, we enable effective resource pooling and matching based on application requirements and network resource availability. In order to provide better networking experience, we are also working on network protocol coexistence and cross technology communication.

Virtual Earplugs: Noise Control With Smartphones

While a certain sound serves a purpose to someone, for others, the same sound could be noise. Specifically, an alarm sound would be necessary for someone to wake up and start the day, while it would be an unwanted sound for people sharing the same room, who need not wake up as early. Noise cancellation is useful in this scenario, but most existing techniques require costly equipments or devices that are uncomfortable to wear during sleep. Our Virtual Earplugs system controls noise with only smartphones. Virtual Earplugs detect the alarm sound and generate the opposite sound (same amplitude but inverted phase) of the alarm by volume control and phase control processes, to cancel the alarm sound.

Drone Networking & Security

Drones are gaining popularity in various applications, including military, aerial surveillance, mapping, inspection, delivery, photography/video, agriculture and even Internet provision. Drones with networking capabilities open up a new research area for networking; at the same time, their security vulnerabilities draw attentions from security researchers. In this project, we devise methods to increase the networking security of drones by addressing issues such as GPS spoofing and jamming. We are also interested in building wireless mesh networks with drones.

Cloud-based WLAN Controller Using Machine Learning

Unplanned and dense installation of Wi-Fi APs generates a highly interfering Wi-Fi network, causing users to experience a severe degradation of Wi-Fi throughput and QoS. In this research, we aim to improve Wi-Fi performance in a highly dense and interfering environment. Specifically, we propose to design and implement (i) a cloud-based large scale WLAN controller, (ii) a reinforcement learning-based administrative domain independent AP control algorithm, and (iii) network MIMO using a large scale WLAN controller.


Our research has been generously supported by: