SPINE2 is a novel framework specifically designed for the development of signal processing applications on WSNs based on the task-oriented paradigm and relying on a platform-independent node-side software architecture.
The task-oriented paradigm provides constructs for specifying an application behavior as a set of interconnected tasks, each of which representing a particular activity ( such as a sensing operation, a processing function or a radio data transmission) by abstracting away low-level details. Designing an application as a composition of elementary blocks enables a more rapid application development, runtime reconfiguration and a simplified software maintenance.
A simple example of SPINE2 task-oriented application specification is depicted in the figure below. The set of interconnected tasks forms a direct graph representing distyributed chains of operations that include sensor data acquisition, data processing and merged result transmission.
Thanks to the software-layering approach (see picture below), the node-side part of the SPINE2 framework can be rapidly ported for supporting different C-like sensor platforms, by simply redefining only the platform-dependent part (such as sensors and radio drivers) of the software architecture.
Specifically, the core framework provides the basic node execution logic (such as task man-agement, application-level message handling and on-board sensor abstraction management) whereas the platform-specific modules are the components needed for accessing the services and the resources provided by a specific sensor node platform on which the software architecture is running.
Apart from the node-side software in charge of the interpretation and execution of the task-based application, the SPINE2 framework includes coordinator-side software components, consisting of a Java-based API and a GUI-based application. Through them developers are able to: (i) managing the sensor network, (ii) defining and deploying the task-based applications, and (iii) gathering pre-elaborated data coming from sensor nodes for eventually passing them to a more specific application which may be in charge of more complex data processing and visualization.
Reference papers:
- N. Raveendranathan et al., “From Modeling to Implementation of Virtual Sensors in Body Sensor Networks,” in IEEE Sensors Journal, vol. 12, no. 3, pp. 583-593, March 2012, doi: 10.1109/JSEN.2011.2121059.
- S. Galzarano, R. Giannantonio, A. Liotta and G. Fortino, “A Task-Oriented Framework for Networked Wearable Computing,” in IEEE Transactions on Automation Science and Engineering, vol. 13, no. 2, pp. 621-638, April 2016, doi: 10.1109/TASE.2014.2365880.
Code Repository
Source code and documentation are available for free download at our GitHub project.
We welcome contribution from all users, including bug reports and fixing, suggestions, and forks of SPINE.
Proposals for new sub-projects and forks should be sent via e-mail at r.gravina@dimes.unical.it by specifying a short workplan with a description of the new functionalities, the impact on existing code, and the expected time for the first stable release.
BSN research prototypes
- Gait-analysis
- Handshake detection