Source-linked AI summary
Dynamic Configuration of Sensors Using Mobile Sensor Hub in Internet of Things Paradigm
Charith Perera, Prem Jayaraman, Arkady Zaslavsky, Peter Christen, Dimitrios Georgakopoulos
TL;DR
Mobile sensor hubs can reduce communication by processing sensor data locally, but middleware may not know which sensors are connected or what streams to expect. The paper proposes MoSHub and autonomous configuration between it and GSN; its evaluation reports DGCW as more efficient overall across four considered factors, despite longer configuration time.
Problem
Middleware may lack information about sensors attached to a mobile phone and their data streams, making autonomous configuration to accept those streams a research challenge.
Source: §III · p. 2Method
MoSHub collects sensor data and communicates with GSN, while runtime code generation configures middleware communication for each hub.
Source: §III · p. 3Results
DGCW was more efficient when all four factors were considered together, although it takes more time to configure each MoSHub than SPSW.
Source: §VII · p. 6Takeaways & Limitations
Customized wrappers can reduce data-stream processing time by up to 18% and support future addition of context discovery to GSN.
Source: §VII · pp. 5–6
Abstract
from arXiv · showhide
Internet of Things (IoT) envisions billions of sensors to be connected to the Internet. By deploying intelligent low-level computational devices such as mobile phones in-between sensors and cloud servers, we can reduce data communication with the use of intelligent processing such as fusing and filtering sensor data, which saves significant amount of energy. This is also ideal for real world sensor deployments where connecting sensors directly to a computer or to the Internet is not practical. Most of the leading IoT middleware solutions require manual and labour intensive tasks to be completed in order to connect a mobile phone to them. In this paper we present a mobile application called Mobile Sensor Hub (MoSHub). It allows variety of different sensors to be connected to a mobile phone and send the data to the cloud intelligently reducing network communication. Specifically, we explore techniques that allow MoSHub to be connected to cloud based IoT middleware solutions autonomously. For our experiments, we employed Global Sensor Network (GSN) middleware to implement and evaluate our approach. Such automated configuration reduces significant amount of manual labour that need to be performed by technical experts otherwise. We also evaluated different methods that can be used to automate the configuration process.
I. INTRODUCTION
The introduction frames sensor networks as layered systems in which data moves from low-end sensors toward the cloud. Because communication consumes more energy than local processing, device capabilities matter when deciding where to process data.
I. INTRODUCTION
- Communication from sensors to cloud applications costs significantly more energy than local processing, motivating filtering and fusion to reduce network traffic.The paper relates this opportunity to the expected growth in sensor deployments and data generation.Source: §I · p. 1
- Data flows from low-end sensor nodes through sink nodes and computational devices before reaching the cloud for sharing, storage, and further processing.The network layers are identified by device capabilities, and IoT may require additional sublayers for its varied sensing capabilities.Source: §I · p. 1, Figure 1 · p. 1
- Capabilities increase across six identified layers, but the first three lack enough energy and processing power for comprehensive data processing.Mobile phones can fuse and filter data to reduce communication costs.Source: §I · p. 1
II. BACKGROUND AND RELATED WORK
The background situates the work in sensor middleware and prior wrapper-generation efforts. It describes GSN as the experimental middleware and contrasts its data-processing capabilities with commercial smart-environment systems.
II. BACKGROUND AND RELATED WORK
- TWINE, Ninja Blocks, and Smart Things focus on If-THEN event detection, while the authors say they lack GSN-like complex query processing and support only a limited number of devices.The paper also notes that its plugin architecture can add capabilities through app stores.Source: §II · p. 2
- GSN provides heterogeneous sensor integration, discovery, combination, querying, and filtering through an XML-based language, and was selected for the experiments.The authors state that their findings remain open to other middleware solutions that use mobile devices as sensor hubs.Source: §II · p. 2
III. MOTIVATIONS AND OUR CONTRIBUTION
The paper addresses autonomous connection of mobile sensor hubs to IoT middleware that does not know the attached sensors or their data streams. It proposes MoSHub and runtime middleware configuration, then compares configuration approaches and preliminarily examines plugin scalability.
III. MOTIVATIONS AND OUR CONTRIBUTION
- The research question is how middleware can autonomously configure itself to accept streams from phones when it lacks information about their locally connected sensors.Phones are presented as capable of collecting, processing, and communicating sensor data.Source: §III · p. 2
- MoSHub is a lightweight Android application that connects to external sensors over WiFi or Bluetooth and retrieves their data.An Android Interface Definition Language plugin architecture supports plug-and-play configuration between sensors and MoSHub.Source: §III · p. 3
- The proposed model configures communication between phone-based MoSHub and server-based GSN by generating code at runtime, reducing manual work for technical experts.The approach extends and applies ASCM4GSN.Source: §III · p. 3
- The authors compare two automation approaches and conduct preliminary evaluations of plugin-architecture scalability.Mobile phones are motivated as sensor-processing devices by their computing and wireless capabilities.Source: §III · pp. 2–3
IV. OUR APPROACH
The approach section outlines the system overview, client- and server-side autonomous configuration, and the MoSHub application. It also says the paper considers alternatives and justifies its choices through experiments.
IV. OUR APPROACH
- The section presents a high-level system overview, explains client- and server-side configuration, describes MoSHub, and evaluates alternative approaches experimentally.The authors state that the experiments are used to justify their approach choices.Source: §IV · p. 3