SWAN-Lake is a framework for building distributed, sensor-based apps for mobile devices equipped with Android. The framework extends SWAN by adding support for opportunistic distributed sensing, which makes way for many new context aware applications. It also extends Swan-song in order to support gathering of data from nearby devices, which makes it possible to accurately identify the context in situations where using only the sensors available on the phone is not enough.
Smartphones are equipped with various hardware sensors to enrich the user experience. SWAN is a middleware framework that supports easy collection and processing of sensor data. However, the limited resources of the smartphones prevent the apps from supporting big data applications that need to store and operate on large historical datasets, leading to the need for implementing a cloud solution. To this end we present the SWAN-Fly framework, a generic and flexible mechanism to ease the application developers’ task of sending sensor data from the device to their preferred cloud solution for additional storage and processing. We evaluate the ease of use of SWAN-Fly on a Crowd Monitoring application and the flexibility is tested on two different cloud solutions.
SWAN expressions offer a flexible way of getting and managing sensor data. SWAN acts as middleware, providing and managing access to hardware or software sensors through it’s powerful expression evaluation features. Until now the expressions was designed to work with unique sensors, which are mostly available on the phone. After extending the available sensors, it became clear that current expression format is not flexible enough to work with multi-sensor models.
The project pursues to extend the SWAN framework by providing support for a new type of sensors, that use the LoRaWan(low range wide area network) technology. The limitations of these new sensors, their functionality compared to their power efficiency and providing of a real life application that can highlight the strong points of using this technology represents the focus of the project.
The goal of the project is to help developer’s/user’s build SWAN application in the web. To acheive we do the following • Develop SWAN web client and Plugins/Client in SWAN (Android) • Develop a communication mechanism between the phone and the web • Integrate API based sensors (Facebook, Gmail) • Identify and build actuators usable for the SWAN • Develop a Notification Mechanism • Evaluate the system for performance/usability implications of plugin vs client/standalone approach.
This project studies how to scale up and enhance distributed mobile sensing using smart phones. The sensors from large numbers of users can in theory be combined to gather valuable data and information about people, environments, and cities, but doing this efficiently and in a privacy-preserving way is a major challenge. Based on preliminary work in COMMIT, we will exploit direct phone-to-phone communication to support interaction between nearby groups of users (e.g., Wi-Fi Direct is a high- performance energy-efficient protocol for direct communication without access points). The result will be an advanced scalable distributed sensing framework that will be used by the ONVZ health insurance company in its Vitality Lab. In particular, ONVZ wants to exploit more and more novel sensors (especially from modern wearables) from groups of users to give real-time health feedback and advice. The project combines extensive research experience in distributed and mobile systems (VU Computer Science/HPDC) with deep knowledge of digital business innovation (VU Economics and Business Administration/KIN) and the innovation agenda from a major company (ONVZ).
The increasing number of sensors in a smartphone can be put to good use with smart sensor frameworks. Such a sensor framework is SWAN, but it is lacking general internet sensors and the expressions are very hard to program by the non programmer. In this work we propose two general use internet sensors (JSON and RSS sensors) which we test on both battery and data usage. To make it easy to use these new, but also the existing sensors, we propose a new expression builder which we compare with the old ways to build an expression.