Introduction

SPINE (Signal Processing in Node Environment) is [an Open Source] software framework for the design of Body Sensor Network (BSN) applications.
Our intent is for SPINE to become a part of the OpenCare Project in order to take advantage of SPINEs support for hardware platforms such as telosb motes with spine sensor board, telosb motes with biosensor sensor board, micaz motes with mts300 board, SHIMMER motes (SHIMMER motes are supported as of SPINE v1.2) aswell as the potential offered by SPINES digital signalprocessing libraries.

About SPINE

SPINE is a free software and is distributed by Telecom Italia, the copyright holder, in open source software under the terms of the LGPL (Lesser General Public Licence Version 2). The latest version of SPINE is SPINE 1.2 released on September 22nd 2008 and can be downloaded here. The latest release now supports Intel SHIMMER, as noted in the SPINE news archive.

Telecom Italia Lab is the R&D branch of the Telecom Italia Group and is responsible for promoting technological innovation by scouting new technologies, carrying out and assessing feasibility studies, and developing prototypes and emulators of new services and products.

Arcitechture

Figure 1 shows two representations of the SPINE Network Architecture with nodes in a star topology around a coordinator. A software application using the SPINE Framework must implement two main software components:
  • One to be executed on the sensor nodes
  • One to be executed on a BSN coordinator.

spine_sensors.PNG
Figure 1. SPINE Network Architecture (v1.0)

The sensor node component, designed in the TinyOS environment and written in nesC language, includes several utilities for signal processing such as data storage buffers, mathematical function libraries and common feature extractors used in signal processing. Furthermore, it includes an over-the-air communication protocol to transfer data from the sensor nodes to the coordinator [1].

SPINE takes pride in being able to exploit implementation tradeoffs in a distributed architecture. While a localized implementation features raw data being sent from the nodes to the coordinator, a distributed implementation allows for computation and decisions to be taken directly on the node itself, saving power and optimizing wireless bandwidth. For such a distributed architecture, the communication between node and coordinator can be controlled in a more granulated manner using the Feature Selection Protocol (FSP), as shown on figure 2. The FSP and Feature Extraction in general is described in more detail the SPINE white paper [2].

spine_overview.PNG
Figure 2. Functional components of the SPINE Architecture (v1.0)

SPINE v1.2 has been redesigned to allow easier addition of new sensors, platforms and functionality.

SPINE Advantages

  • Open source
  • Interoperability (Lightweight Java "server side" API allows porting to PCs/Mobile phones)
  • Distributed implementation (Perform signal processing directly on nodes to reduce the amount of data to be transmitted and save energy)
  • Reusability (Coordinators can recognize functions of an already configured sensor node, thus allowing for the reuse of nodes in different applications)

Related work

Other initiatives exist in order to reduce development time of sensory applications. Biomobius, developed by the TRILCentre, is a portfolio of hardware and software dedicated to the rapid creation of information technology solutions for biomedical research. Through a Drag'n'drop Graphical Development Environment (GDE) called EyesWeb, domain experts are able to quickly piece together biosignal applications using components developed by biosignal- and Platform Development engineers. BioMobius is not Open Source, but is an open and shareable technology platform. It's "openness" is based on the ability to extend EyesWeb's component libraries through components written in C++. Because these components are made available to domainexperts through applicationspecific catalogues only understood by EyesWeb, BioMobius is a fixed solution with little option for integration with other frameworks.

Getting more information

  • The TinyOS documentation Wiki is an extensive collection of TinyOS tutorials, manuals and reference documents pertaining to TinyOS. Community contributions are encouraged, and the TinyOS developers prefer all documentation to live here.
  • spine-news and spine-develop are two mailinglists that represent SPINEs user forum. Sign up for these when downloading the source code.


REFERENCES