Sugang Li

Crowd++: unsupervised speaker count with smartphones

Smartphones are excellent mobile sensing platforms, with the microphone in particular being exercised in several audio inference applications. We take smartphone audio inference a step further and demonstrate for the first time that its possible to accurately estimate the number of people talking in a certain place -- with an average error distance of 1.5 speakers -- through unsupervised machine learning analysis on audio segments captured by the smartphones. Inference occurs transparently to the user and no human intervention is needed to derive the classification model. Our results are based on the design, implementation, and evaluation of a system called Crowd++, involving 120 participants in 10 very different environments. We show that no dedicated external hardware or cumbersome supervised learning approaches are needed but only off-the-shelf smartphones used in a transparent manner. We believe our findings have profound implications in many research fields, including social sensing and personal wellbeing assessment.

Whose move is it anyway? Authenticating smart wearable devices using unique head movement patterns

In this paper, we present the design, implementation and evaluation of a user authentication system, Headbanger, for smart head-worn devices, through monitoring the users unique head-movement patterns in response to an external audio stimulus. Compared to todays solutions, which primarily rely on indirect authentication mechanisms via the users smartphone, thus cumbersome and susceptible to adversary intrusions, the proposed head-movement based authentication provides an accurate, robust, light-weight and convenient solution. Through extensive experimental evaluation with 95 participants, we show that our mechanism can accurately authenticate users with an average true acceptance rate of 95.57% while keeping the average false acceptance rate of 4.43%. We also show that even simple head-movement patterns are robust against imitation attacks. Finally, we demonstrate our authentication algorithm is rather light-weight: the overall processing latency on Google Glass is around 1.9 seconds.

Crowdsensing the speaker count in the wild: implications and applications

The mobile crowdsensing (MCS) paradigm enables large-scale sensing opportunities at lower deployment costs than dedicated infrastructures by utilizing today¿s large number of mobile devices. In the context of MCS, end users with sensing and computing devices can share and extract information of common interest. In this article, we examine Crowd++, an MCS application that accurately estimates the number of people talking in a certain place through unsupervised machine learning analysis on audio segments captured by mobile devices. Such a technique can find application in many domains, such as crowd estimation, social sensing, and personal well being assessment. In this article, we demonstrate the utility of this technique in the context of conference room usage estimation, social diaries, and social engagement in a power-efficient manner followed by a discussion on privacy and possible optimizations to Crowd++ software.

A comparative study of MobilityFirst and NDN based ICN-IoT architectures

To develop unified IoT platforms where objects can be made accessible to applications across organizations and domains, popular solutions are based on client-server overlays on todays Internet. These solutions, however, inherit the inefficiencies of the current Internet - especially in terms of mobility, scalability, and communication reliability. To address this problem, we propose to build the unified IoT platform leveraging the salient feats of Information-Centric Network (ICN) architectures, which we call ICN-IoT. Specifically, we explore two ICN architectures - MobilityFirst and NDN - to support IoT, and refer to them as MF-IoT and NDN-IoT, respectively. Through detailed simulations, we find that though these two architectures fare comparably, MF-IoT incurs lower control overheads.

A Mobile Phone Based WSN Infrastructure for IoT over Future Internet Architecture

Large-scale wireless sensor network (WSN) deployment is a major challenge for Internet of Things (IoT) to connect physical world through sensors or tags at the scale of billions of objects. Todays WSNs normally use dedicated gateways to bridge sensors and IP networks. The installation and maintenance of such a WSN infrastructure are expensive and non-scalable. As the number of smart phone users grows exponentially every year, a powerful mobile computing platform could be used as an alternative WSN infrastructure, which is ubiquitous and scalable. This paper studies the major challenges of using mobile phones as spontaneous gateways of WSNs in IoT systems. The challenges include (1) matching the throughput of mobile phone gateways and sensor data rate at hotspot locations, (2) securing sensor data access and (3) providing accounting records of gateway the service offered by mobile phones. In this paper, we show that using a name-based future Internet architecture (FIA), such as Mobility First, these challenges can be overcome effectively though its native networking layer functions. A proof-of-concept prototype system demonstrates the feasibility of proposed solution. The system delivers a temperature sensor data from an Android phone directly to multiple applications via in-network multicast over an Mobility First network test bed.

IoT Middleware Architecture over Information-Centric Network

Many approaches have been proposed to build a unified IoT platform where physical and digital objects are accessible by applications crossing different organization and domains, and are based on IP-overlay architecture. These solutions inherit the constraints of the current internet, especially in terms of naming, heterogeneity, mobility and security. In this paper, we propose a new Information-Centric Network (ICN) based IoT middleware to address these challenges by leveraging various promising features of ICN, such as naming. We elaborate the functions of ICN-based IoT middleware by integrating with the two future internet architectures, namely Named-Data Networking and MobilityFirst. Moreover, we evaluate the efficiency of service discovery (one of the functions in the proposed ICN-based IoT middleware) and demonstrate the feasibility of the proposed ICNbased IoT middleware .

Exploiting ICN for Realizing Service-Oriented Communication in IoT

The rapid growth in Internet of Things (IoT) deployment has posed unprecedented challenges to the underlying network design. We envision that tomorrows global-scale IoT systems will focus on service-oriented data sharing and processing rather than point-to-point data collection. Requirements such as global reachability, mobility, communication diversity, and security will also develop naturally along with the change in the communication patterns. However, existing (IP-based) networks focus only on locations and point-to-point channels. The mismatch between the dynamic requirements and the functionalities provided by the network renders IoT communication inefficient and inconvenient. Information-centric networking (ICN) shifts the focus from location to the (identity of) information. This paradigm can naturally be adapted to IoT communication since service can also be identified as a type of information. To demonstrate the potential of ICN in IoT communication, this article adopts and modifies a particular example of ICN called MobilityFirst, and shows that the target architecture can satisfy the requirements posed by IoT communication. Similar adaptations can also be used by other ICN architectures such as NDN and XIA.

A Security Framework for the Internet of Things in the Future Internet Architecture

The Internet of Things (IoT) is a recent trend that extends the boundary of the Internet to include a wide variety of computing devices. Connecting many stand-alone IoT systems through the Internet introduces many challenges, with security being front-and-center since much of the collected information will be exposed to a wide and often unknown audience. Unfortunately, due to the intrinsic capability limits of low-end IoT devices, which account for a majority of the IoT end hosts, many traditional security methods cannot be applied to secure IoT systems, which open a door for attacks and exploits directed both against IoT services and the broader Internet. This paper addresses this issue by introducing a unified IoT framework based on the MobilityFirst future Internet architecture that explicitly focuses on supporting security for the IoT. Our design integrates local IoT systems into the global Internet without losing usability, interoperability and security protection. Specifically, we introduced an IoT middleware layer that connects heterogeneous hardware in local IoT systems to the global MobilityFirst network. We propose an IoT name resolution service (IoT-NRS) as a core component of the middleware layer, and develop a lightweight keying protocol that establishes trust between an IoT device and the IoT-NRS.

Monitoring a Person's Heart Rate and Respiratory Rate on a Shared Bed Using Geophones

Using geophones to sense bed vibrations caused by ballistic force has shown great potential in monitoring a persons heart rate during sleep. It does not require a special mattress or sheets, and the user is free to move around and change position during sleep. Earlier work has studied how to process the geophone signal to detect heartbeats when a single subject occupies the entire bed. In this study, we develop a system called VitalMon, aiming to monitor a persons respiratory rate as well as heart rate, even when she is sharing a bed with another person. In such situations, the vibrations from both persons are mixed together. VitalMon first separates the two heartbeat signals, and then distinguishes the respiration signal from the heartbeat signal for each person. Our heartbeat separation algorithm relies on the spatial difference between two signal sources with respect to each vibration sensor, and our respiration extraction algorithm deciphers the breathing rate embedded in amplitude fluctuation of the heartbeat signal. We have developed a prototype bed to evaluate the proposed algorithms. A total of 86 subjects participated in our study, and we collected 5084 geophone samples, totaling 56 hours of data. We show that our technique is accurate -- its breathing rate estimation error for a single person is 0.38 breaths per minute (median error is 0.22 breaths per minute), heart rate estimation error when two persons share a bed is 1.90 beats per minute (median error is 0.72 beats per minute), and breathing rate estimation error when two persons share a bed is 2.62 breaths per minute (median error is 1.95 breaths per minute). By varying sleeping posture and mattress type, we show that our system can work in many different scenarios.

MF-IoT: A MobilityFirst-Based Internet of Things Architecture with Global Reach-Ability and Communication Diversity

The rapid growth in IoT deployment has posed unprecedented challenges to the underlying network design. We envision tomorrows global-scale IoT systems should support global device reach-ability, mobility, diverse communication patterns, and resource efficiency. Existing solutions either rely on IP protocols that do not have efficient mobility support, or seek application-layer optimizations that incur high computing and deployment overhead. To fill this void, we propose to adopt cleanslate network architecture in the core network that decouples locators from network addresses, and towards this end, we propose MF-IoT, which is centered around the MobilityFirst architecture that focuses on mobility handling. MF-IoT enables efficient communication between devices from different local domains, as well as communication between devices and the infrastructure network. The MF-IoT network layer smoothly handles mobility during communication without interrupting the applications. This is achieved through a transparent translation mechanism at the gateway node that bridges an IoT domain and the core network. In the core network, we leverage MobilityFirst functionalities in providing efficient mobility support for billions of mobile devices through long, persistent IDs, while in the local IoT domain, we use short, local IDs for energy efficiency. By seamlessly translating between the two types of IDs, the gateway organically stitches these two parts of the network. Through large scale simulation studies, we show that MFIoT is able to achieve the four features we envisioned for an IoT system, with performance optimizations in reducing network traffic load, avoiding congestion, and ensuring fast and timely packet delivery.