Proliferation of sensors into everyday environments is resulting in a connected world that generates large volumes of complex data. This data is opening new scientific and commercial investigations in fields such as pollution monitoring and patient health monitoring. Parallel to this development, deep learning has matured into a powerful analytics technique to support these investigations.However, computing and resource requirements of deep learning remain a challenge, often forcing analysis to be carried at remote third-party data centres. In this article, we describe an alternative computing as a service model where available smart devices opportunistically form micro-data centres that can support deep learning-based investigations of data streams generated by sensors. Our model enables smart homes, smart buildings, smart offices, and other types of smart spaces to become providers of powerful computation as a service, enabling edge analytics and other applications that require pervasive (in-space) decisioning.

Wearable thermal imaging is emerging as a powerful and increasingly affordable sensing technology. Current thermal imaging solutions are mostly based on uncooled forward looking infrared (FLIR), which is susceptible to errors resulting from warming of the camera and the device casing it. To mitigate these errors, a blackbody calibration technique where a shutter whose thermal parameters are known is periodically used to calibrate the measurements. This technique, however, is only accurate when the shutter's temperature remains constant over time, which rarely is the case. In this paper, we contribute by developing a novel deep learning based calibration technique that uses battery temperature measurements to learn a model that allows adapting to changes in the internal thermal calibration parameters. Our method is particularly effective in continuous sensing where the device casing the camera is prone to heating. We demonstrate the effectiveness of our technique through controlled benchmark experiments which show significant improvements in thermal monitoring accuracy and robustness.

While mobile apps have become an integral part of everyday life, little is known about the factors that govern their usage. Particularly the role of geographic and cultural factors has been understudied. This article contributes by carrying out a large-scale analysis of geographic, cultural, and demographic factors in mobile usage. We consider app usage gathered from 25, 323 Android users from 44 countries and 54, 776 apps in 55 categories, and demographics information collected through a user survey. Our analysis reveals significant differences in app category usage across countries and we show that these differences, to large degree, reflect geographic boundaries. We also demonstrate that country gives more information about application usage than any demographic, but that there also are geographic and socio-economic subgroups in the data. Finally, we demonstrate that app usage correlates with cultural values using the Value Survey Model of Hofstede as a reference of cross-cultural differences.

With great power comes - besides great responsibility - big energy drain, especially where Internet of Things (IoT) devices are considered. Indeed, despite significant improvements in design and manufacturing, energy efficiency remains a critical design consideration for IoT, particularly for devices operating continuous sensing. The energy footprint of these devices has traditionally been measured using hardware power monitors (such as Monsoon power meter), which provide an accurate view of instantaneous power use. However, power meters require direct connection with the device's power source (such as battery) and hence can be used to measure energy drain only on devices with detachable power sources.

Energy-efficiency remains a critical design consideration for mobile and wearable systems, particularly those operating continuous sensing. Energy footprint of these systems has traditionally been measured using hardware power monitors (such as Monsoon power meter) which tend to provide the most accurate and holistic view of instantaneous power use. Unfortunately applicability of this approach is diminishing due to lack of detachable batteries in modern devices. In this paper, we propose an innovative and novel approach for assessing energy footprint of mobile and wearable systems using thermal imaging. In our approach, an off-the-shelf thermal camera is used to monitor thermal radiation of a device while it is operating an application. We develop the general theory of thermal energyefficiency, and demonstrate its feasibility through experimental benchmarks where we compare energy estimates obtained through thermal imaging against a hardware power monitor. CCS CONCEPTS• Computer systems organization → Embedded systems; • Hardware → Power and energy; Power estimation and optimization; Temperature monitoring.

While the use of smartphones in extreme temperatures does not necessarily occur every day nor in all parts of the world, numerous use cases can be highlighted where the use of smartphones in cold temperatures is mandatory. Modern smartphones are designed to function in a wide range of temperatures, but when exposed to extreme cold temperatures the performance and reliability can significantly suffer. This paper presents a controlled laboratory experiment, using a clinical cold chamber to expose seven smartphone models to both medium cold (0 • C to −20 • C) and extreme cold (−30 • C) environments. The results showcase the smartphones' sensing software's lack of awareness of the cold environment, as well as reliability issues in the form of device crashes across the whole range of tested devices. We present a strategy for implementing monitoring application designs to both appropriately sense the effect of cold environments, as well as predicting device shutdowns in extreme cold.