The rapid emergence of head-mounted devices such as the Microsoft Holo-lens enables a wide variety continuous vision applications. Such applications often adopt deep-learning algorithms CNN and RNN to extract rich contextual information from first-person-view video streams. Despite high accuracy, use deep learning in mobile raises critical challenges, i.e., processing latency power consumption. In this paper, we propose DeepMon, inference system run inferences purely on device fast...

Target imaging and material identification play an important role in many real-life applications. This paper introduces TagScan, a system that can identify the type image horizontal cut of target simultaneously with cheap commercial off the-shelf (COTS) RFID devices. The key intuition is different materials sizes cause amounts phase RSS (Received Signal Strength) changes when radio frequency (RF) signal penetrates through target. Multiple challenges need to be addressed before we turn idea...

In this paper, we propose cyber foraging: a mechanism to augment the computational and storage capabilities of mobile devices. Cyber foraging uses opportunistically discovered servers in environment improve performance interactive applications distributed file systems on clients. We show how can be improved by staging data at these even though are not trusted. also via remote execution. Finally, present VERSUDS: virtual interface heteregeneous service discovery protocols that used discover servers.

Remote execution can transform the puniest mobile device into a computing giant able to run resource-intensive applications such as natural language translation, speech recognition, face and augmented reality. However, easily partitioning these for remote while retaining application-specific information has proven be difficult challenge. In this paper, we show that automated dynamic repartitioning of reconciled with need exploit knowledge. We useful knowledge about an application relevant...

Cyber foraging is the transient and opportunistic use of compute servers bymobile devices. The short market life such devices makes rapid modification applications for remote execution an important problem. We describe a solution that combines "little language" cyber with adaptive runtime system. report results from user study showing even novice developers are able to successfully modify large, unfamiliar in just few hours. also show quality novice-modified expert-modified comparable most cases.

In this paper, we describe the design, analysis, implementation, and operational deployment of a real-time trip information system that provides passengers with expected fare duration taxi ride they are planning to take. This was built in cooperation operator operates more than 15,000 taxis Singapore. We first overall design then explain efficient algorithms used achieve our predictions based on up 21 months historical data consisting approximately 250 million paid trips. various...

The inclusion of the barometer sensor in smartphones signaled an opportunity for aiding indoor localization efforts. In this paper, we therefore investigate a possible use detecting vertically oriented activities. We start by showing accuracies various commodity measurement devices and challenges they bring forth. then show how to values build predictor that can detect floor changes mode (elevator, escalator, or stairs) used change floors with nearly 100% accuracy. validate these properties...

We explore the use of gesture recognition on a wrist-worn smartwatch as an enabler automated eating activity (and diet monitoring) system. show, using small-scale user studies, how it is possible to accelerometer and gyroscope data from accurately separate episodes similar non-eating activities, additionally identify mode (i.e., spoon, bare hands or chopsticks). Additionally, we investigate likelihood automatically triggering smartwatch's camera capture clear images food being consumed, for...

In this paper, we show how tone mapping techniques can be used to dynamically increase the image brightness, thus allowing LCD backlight levels reduced. This saves significant power as majority of LCD's display is consumed by its backlight. The Gamma function (or equivalent) efficiently implemented in smartphones with minimal resource cost. We describe overcame function's non-linear nature using adaptive thresholds apply different values images differing brightness levels. These allow us...

Real-time monitoring of groups and their rich contexts will be a key building block for futuristic, group-aware mobile services. In this paper, we propose GruMon, fast accurate group system dense complex urban spaces. GruMon meets the performance criteria precise detection at low latencies by overcoming two critical challenges practical spaces, namely (a) high density crowds, (b) imprecise location information available indoors. Using host novel features extracted from commodity smartphone...

In this paper, we examined the impact of project-level configurational choices globally distributed software teams on project productivity, quality, and profits. Our analysis used data from 362 projects four different firms. These spanned a wide range programming languages, application domain, process choices, development sites spread over 15 countries 5 continents. revealed fundamental tradeoffs in choosing that are optimized for and/or particular, achieving higher levels productivity...

Contact tracing is a well-established and effective approach for the containment of spread infectious diseases. While Bluetooth-based contact method using phones has become popular recently, these approaches suffer from need critical mass adoption to be effective. In this paper, we present WiFiTrace, network-centric that relies on passive WiFi sensing with no client-side involvement. Our exploits network logs gathered by enterprise networks performance security monitoring, utilizes them...

Software firms are increasingly distributing their software development effort across multiple locations. In this paper we present the results of a two year field study that investigated effects dispersion on productivity and quality distributed development. We first develop model then use model, along with our empirically observed data, to understand consequences project performance. Our analysis reveals that, even in high process maturity environments, a) significantly reduces has...

Recently, a branch of machine learning algorithms called deep gained huge attention to boost up accuracy variety sensing applications. However, execution algorithm such as convolutional neural network on mobile processor is non-trivial due intensive computational requirements. In this paper, we present our early design DeepSense - GPU-based (CNN) framework. For its design, first explored the differences between server-class and mobile-class GPUs, studied effectiveness various optimization...

In this paper, we present the design and implementation of Focus, a system for effectively efficiently reducing power consumption OLED displays on smartphones. These displays, while becoming exceedingly common still consume significant power. The key idea Focus is that use notion saliency to save display by dimming portions applications are less important user. We envision being especially useful during low battery situations when usability than savings. tested using 15 running Samsung...

Understanding the engagement levels players have with a game is useful proxy for evaluating design and user experience. This particularly important mobile games as an alternative always just easy download away. However, subjective concept usually requires fine-grained highly disruptive interviews or surveys to determine accurately. In this paper, we present EngageMon, first-of-its-kind system that uses combination of sensors from smartphone (touch events), wristband (photoplethysmography...

Driven by the rapid market growth of sensors and beacons that offer Bluetooth Low Energy (BLE) based connectivity, this paper empirically investigates performance characteristics BLE interface on multiple Android smartphones, consequent impact a proposed BLE-based service: continuous indoor location. We first use extensive measurement studies with devices to establish current smartphones is not as "low-energy" nominally expected, such feasible unless we choose moderately large scan interval...

We present VitaMon, a mobile sensing system that can measure the inter-heartbeat interval (IBI) from facial video captured by commodity smartphone's front camera. The continuous IBI measurement is used to compute heart rate variability (HRV), one of most important markers autonomic nervous (ANS) regulation. underlying idea VitaMon recording human face contains multiple cardiovascular pulse signals with different phase shift. Our on 10 participants shows significant time delay (36.79 ms)...

In this paper, we show that characterizing user behaviour from location or smartphone usage traces, without accounting for the interaction of individuals in physical-world groups, can lead to erroneous results. We conducted one largest studies UbiComp domain thus far, involving indoor traces more than 6,000 users, collected over a 4-month period at our university campus, and further studied fine-grained App subset 156 Android users. apply state-of-the-art group detection algorithm annotate...

We present LiveLabs, a mobile experimentation testbed that is currently deployed across our university campus with further deployments at large shopping mall, commercial airport, and resort island soon to follow. The key goal of LiveLabs allow in-situ real-time applications services require context-specific triggers real participants on their actual smart phones. describe how works, then explain the novel R&D required realise it. end description current status (> 700 active date) as well...

Stress and depression are a common affliction in all walks of life. When left unmanaged, stress can inhibit productivity or cause depression. Depression occur independently stress. There has been sharp rise mobile health initiatives to monitor However, these usually require users install dedicated apps multiple sensors, making such solutions hard scale. Moreover, they emphasise sensing individual factors overlook social interactions, which plays significant role influencing while being part...

Wireless networks have become increasingly common and an increasing number of devices are communicating with each other over lossy links. Unfortunately, TCP performs poorly links as it is unable to differentiate the loss due packet corruption from that congestion. We present extension which enables distinguish congestion in environments resulting improved performance. refer this HeAder ChecKsum option (HACK). implemented our algorithm Linux kernel performed various tests determine its...

In this paper, we present the design and evaluation of a near-field communication-based mobile p2p payment application, called mFerio, that is designed to replace cash-based transactions. We first identify criteria systems should satisfy then explain how relative those criteria, improves on limitations systems. next describe mFerio's implementation user interface design, focusing balance between usability security. Finally, results two-phase study, involving total 104 people, shows mFerio...

Advertisements are the de-facto currency of Internet with many popular applications (e.g. Angry Birds) and online services (e.g., YouTube) relying on advertisement generated revenue. However, current economic models mechanisms for mobile advertising fundamentally not sustainable far from ideal. In particular, as we show, which use capable using significant amounts a users' critical resources without being controlled or held accountable. This paper seeks to redress this situation by enabling...