Performance analysis of an activity monitoring system using the SPINE framework

SPINE is an Open Source Framework for the design of signal processing intensive (SPI) WSNs. It supports the construction of WSN applications through high-level abstractions and libraries, and allows designers to quickly explore implementation tradeoffs through fast prototyping. This paper describes the architecture of SPINE and presents implementation parameters, such as processing time, memory, bandwidth usage and power consumption that are most relevant for application developers to set tunable parameters and analyze system performance. Finally, the paper presents the performance and the resource usage of a SPINE based posture recognition system for elderly health monitoring.