NRL Home
 
Search
Members | Research | Lectures | Publication | Contact | Lab Wiki | KAIST KSE
Recent Research Projects
 Mobile Health with Data Science and Wearables
  • Our lab maintains the KAIST DrM health platform for collecting and visualizing mobile health data. We distributed various wearable devices such as Apple Watch, Samsung Gear S2, and Fitbit Charge and have been collecting activity tracker data (e.g., step counts, heart rates, and GPS traces) from a large number of people in campus. We have been analyzing this longitudinal activity tracking dataset to understand wearing behaviors and user experiences of wearable devices. We recently extended our platform to collect detailed user interaction data from smart devices. This will help us to uncover behavioral markers of mental and physical well-being metrics (e.g., depression, stress) and to design novel mobile intervention software. Furthermore, we have been exploring how novel wearable technologies (e.g., smartphone apps and VR headsets) can be used to innovate diagnosis and treatment of various health problems by supporting in-situ measurement and data processing.
 Problematic Smartphone Use: Data Analytics and SW-based Intervention Services
  • Concomitant with the explosive increase in the popularity of smartphones in recent years, negative aspects of their usage, such as social conflicts, sleep deprivation, and attention deficit, have emerged. Our research has been focused on dealing with problematic usage behaviors by applying computational techniques and developing computer-assisted intervention methods. Our major research contributions include 1) identification of usage patterns related to problematic smartphone usage, and development of automatic problematic usage classification systems (CHI 2014); 2) development and validation of several computer-assisted smartphone overuse intervention services that leverage social support. For example, NUGU (CSCW 2015) and FamiLync (Ubicomp 2015) were designed to foster social awareness and improve self-regulation of smartphone use by considering key social contexts (e.g., among friends or family members). Lock n’ LoL (CHI 2016, Best Paper Award) was designed to mitigate social exclusion due to smartphone distraction during group activities by supporting synchronous awareness to limit usage.
 Designing Serious Games for Well-being and Education
  • Our lab has been working on designing and evaluating various serious games whose primary goals are to promote physical well-being using computer games and gamification techniques. Designing interactive technologies that can improve physical wellbeing has been an active area of research in the field of HCI and health sciences. Our lab has investigated how to gamify solitary exercises such as stationary cycling and running on treadmills (Mobisys 2012; CSCW 2013; CHI 2014). We have also examined how interactive technologies can enrich group fitness exercises such as group swimming (Sensys 2014; CHI 2016). In addition, we explore the usability and user experiences of integrating fitness equipment into a workstation environment (Ubicomp 2016).
 Smart Factory: Machine Condition Monitoring
  • Monitoring machine condition is critical for maintaining the assets in any manufacturing process. Breakdown can cause serious consequences such as production loss or costly repairs. The goal of this project is to diagnose the health of manufacturing equipment and to predict the remaining useful life (RUL) of machine parts by analyzing vibration sensor data.
 Novel Mobile Service Design and Evaluation
  • Our lab has been exploring how novel sensing and collaboration technologies can innovate conventional services, ranging from sharing color reviews of products purchased online (MobileHCI 2015) to collaborative photographing (ACM CHI 2017) and community policing (ACM CHI 2016, 2017). Here, we explain our recent studies on community policing which is defined as the police's efforts to partner with community members and civic organizations to enhance a wide range of neighborhood safety issues (e.g., crimes, norms), by letting them to participate in various prevention, problem solving, and law enforcement activities. In this project, we perform an exploratory study of designing and evaluating new forms of community policing by leveraging pervasive recording technologies. In particular, we study privacy concerns and motives behind video sharing (ACM CHI 2016). Furthermore, we designed and evaluated a mobile app that helps citizen record traffic violation with their smartphones and report the recorded videos to the police (ACM CHI 2017). We have been incorporating advanced computer vision techniques into Mobile Roadwatch for automatically capturing traffic violations and safety risks, including potholes and obstacles.


Past Research Projects
 Social Q&A and Beyond
  • Question and answering (Q&A) sites such as Yahoo! Answers and Naver KiN facilitate knowledge sharing among users by leveraging the wisdom of crowds. Q&A sites complement existing Web search engines by supporting flexible query formulation and customized answer delivery. Further, with the rising popularity of mobile phones, mobile versions of Q&A sites have recently been introduced to the market, including Naver Mobile Q&A, ChaCha, Jisiklog, and AQA. Our goal is to understand usage behavior and social dynamics of such social Q&A systems, particularly in mobile contexts [CHI12] [CHI13]. Another research direction is to build a personalized social Q&A system that can automate the answering process based on the knowledge on the web data and Q&A data.
 Mobile (Social) Computing Systems
  • Recent smartphones are equipped with various sensors such as GPS, camera, audio, and video and support various communications means such as 2/3G, WiFi, and Bluetooth. Bluetooth enables us to connect other external sensors via a wireless data acquisition board (e.g., BlueSentry); and 3G/LTE connections will provide always-on Internet connection, making data access and retrieval amenable. By leveraging such resources, we have been investigating how to implement the enabling techniques for location-aware services, namely indoor navigation [PerCom10], localization [PerCom13], location-aware data offloading [INFOCOM13], and semantic location recognition [Work-in-progress]. Further, we have been examining how mobile phones can mediate social interactions in everyday lives [MobiSys13].
 Networked Collaboration Platforms
  • Our work on networked collaboration covers two directions: (1) facilitating research collaboration in a large university (by supporting topic-based people search and lab info browsing to spur online research group activities) [Work-in-progress], and (2) facilitating end-user and developer interactions in app stores (by enabling new communication channels and supporting supplementary tools, e.g., review summarization) [CHI13].
 Persuasive Computing
  • We have been working on persuasive computing platforms that aim at changing human behavior for environment sustainability and psycho-physical well-being, by leveraging ubiquitous sensing and networked collaboration. For instance, we have been working on building pervasive exercise game platforms that can gamify existing exercises as fun games to improve exercise motivation [MobiSys 2012][CSCW 2013].

 User Behavior Study: Web Search and P2P File Sharing
  • We investigated the characterization of users searching the Web with the purpose of improving the quality of the search engine. We identified two types of behaviors, namely user-click behavior and anchor-link distribution, and showed that they accurately identify a user's goal [WWW05]. Any P2P content distribution system (fixed or mobile) is vulnerable to attacks. For P2P file sharing in the Internet we studied the problem of pollution caused by large numbers of decoy files injected in the system by the attackers. The results of a human subject study (validated also by a mathematical model) show that pollution dynamics are closely related to user behavior [IPTPS06].
 Vehicular Sensor Networking
  • One can envision that vehicles equipped with various sensors (from toxic detectors to still/video cameras) will inter-work in the future to provide a mobile urban sensor fabric on which an unlimited number of urban monitoring applications can be built. Unlike traditional sensor networks, vehicular sensor networks (VSNs) are not subject to major memory, processing, storage, and energy limitations. However, the typical scale of a VSN over wide geographic areas (e.g., millions of nodes), the volume of generated data (e.g., streaming data), and mobility of vehicles make it infeasible to adopt traditional sensor network solutions where sensed data tends to be systematically delivered to sinks using data-centric protocols. Our focus has been developing efficient data gathering and searching protocols [WIRMAG06, TVT09, ADHOC09].
  Content Dissemination
  • Together with navigation safety and urban sensing, content distribution is one of the most promising vehicular ad hoc network (VANET) applications due to the growing demand from passengers and drivers (the latter through carefully designed interfaces) of location and situation aware content ranging from road condition data to upcoming attractions, car-to-car games and vehicle diagnostic software. Given that road side APs are sporadically installed, we used peer-to-peer (P2P) content distribution techniques and also incorporated network coding in the system design to efficiently handle network dynamics [MOBISHARE06, SECON08, JCN08].
  • We studied content distribution to mobile Bluetooth users, enabling them to download multimedia files from the infrastructure (e.g., Digital Billboards) and through direct Peer-to-Peer (P2P) exchanges [PERCOM07, PMC07]. Since the protocol design is mainly dependent on Bluetooth, we separately investigated key Bluetooth functions such as peer discovery and data downloading and measured the performance in dynamic environments with mobility and wireless interference [TVT10].
 Vehicular Networking: Routing, Sensing, Apps


  • The sharp increase of vehicles in the recent years has made driving more challenging and dangerous. For safe driving, leading car manufacturers have been jointly working with national government agencies to develop solutions to help drivers anticipate hazardous events and avoid traffic jams. One of the recent outcomes is a novel wireless architecture called Wireless Access for Vehicular Environment (WAVE) that provides short-range inter-vehicular communications to enable fast dissemination of emergency related messages. While the major objective has clearly been to improve the overall safety of vehicular traffic, industry labs and academia have been exploring novel vehicular applications such as traffic management and on-board entertainment. Our group has been working on various vehicular routing protocols [WONS09, COMMAG10, AUTONET07] which are essential to support wide-are coverage and developing novel vehicular applications ranging from virtual market places [V2VCOM06, TVT10] to emergency video streaming [VANET06, IEICE08].
 Drifting Sensor Networks
  • Moreover, we worked on wireless mobile sensor drifters that navigate and monitor the Wastewater Collection System (WCS). Note that monitoring the aging WCS infrastructure has become a top priority due to the potential danger to public health, yet the cost and complications associated with the current inspection techniques are prohibiting. Our proposed sensor SewerSnort drifts sewer pipelines and measures toxic gases such as hydrogen sulfide caused by sewer-bed sediments [SECON09].
©2010 Interactive Computing Lab. at KAIST KSE, Powered by UseModWiki