Plug-and-Play environment for sensor configuration and data collection

Abstract:

The objective of this research is to investigate and implement a software architecture to improve productivity in the development of rapidly deployable, robust, real-time situational awareness and response (R3SAR) applications. The approach is to allow a commodity to device to dynamically configure itself for a pre-instrumented sensor environment, such as a truck with attached road temperature, humidity, and moisture sensors. This device will then collect data and send it to a central server whose server environment can also be rapidly configured to accept the data. Emergency response applications, typified by human and natural disasters such as hurricanes or chemical spills, are distinguished by the need to be rapidly deployable, to maintain robustness under potential disruptions in the network, and to provide real-time communication guarantees. The system is intended for a multi-modal domain in which mobile sensors attached to vehicles or persons are capable of constantly streaming data and that first-response users require painless data collection from their local devices and emergency managers require a high-level overview of the complete picture of data. This past few years have seen a revolution of technologies available for sensor systems. Inexpensive Android-based mobile phones and tablets have gotten a significant share of the commercial market. Android and its devices provide pre-packaged versions of many of the software and hardware components needed by a typical emergency response sensor system including an on-platform relational database, camera, accelerometer, microphone, video camera, and most importantly, a cellular radio and global positioning system (GPS). In a typical ad-hoc sensor environment, these devices must be attached to an embedded device at greater expense and less reliability than is available by the typical Android phone. Interfaces between the Android, Arduino, and USB-based sensors are now available, including libraries supported by Google. Our focus therefore to building the R3SAR infrastructure as layer upon the Android API. Our primary driving problem is to build an extensible infrastructure for vehicle-based sensors connected to a central geo-analytics platform. A server side sensor meta-model within the RENCI Geoanalytics Framework can then be configured to host collected data from the various sensor environments. The focus of the project over the past year has been the management of sensor data within this framework. A key issue is to integrate the collected sensor data with other datatypes in the geoanalytics framework to provide a uniform interface for analysis, visualization, and data transmission. The framework provides SOS, WFS, WMS, REST endpoints for data. It supports interactive exploration and analysis of data with R, and Ipython Notebook. It also facilitates the deployment of web and mobile applications for understanding, federating, and analyzing sensor data with thematic data.