A lightweight filter component for real-time signal processing in TinyOS

This paper presents a new component of TinyOS for real-time signal processing. The goal is to provide a fast and flexible filter component for real-time signal processing. We described the component structure and algorithm design. Realistic environment tests and offline simulations are both demonstrated. Implementation and evaluation of this component are presented for vehicle detection based on magnetic sensor. It is shown that this component can easily achieve a detection rate of 92.37%, a memory overhead of merely 46 bytes, a read latency of 174.3us. Performance of detection rate, memory overhead, and latency reveals that our proposed component is efficient and suitable for signal processing on resource-constrained sensor devices. This work of filter components can be applied on any similar signal processing in TinyOS.