Synchronization Of Wireless Accelerometer Sensors For Industrial Application

More Info
expand_more

Abstract

Wireless sensor networks play a vital role in major technological developments. The success of such networks depend on the quality and reliability of data acquisition. Despite a lot of research involving network clock synchronization, the area of synchronous sampling has not been dealt in much detail. This thesis aims at studying and implementing synchronization at the sensor level on wireless accelerometer sensors. The designated application for this thesis involves machine condition monitoring using accelerometer sensors that requires high synchronization accuracy between the samples. Two approaches are presented to obtain synchronous sampling, namely: real-time and subsequent synchronization. The sensor nodes designed with commercially-off-the-shelf components are used for the implementation of the two methods. The embedded software was developed as a platform to realize the proposed synchronous data acquisition techniques. The real-time synchronization uses a software solution and provides an interrupt-based triggering to align the sampling instant of the sensors. An external reliable clock source is required for the accelerometer to implement this technique. The signal propagation delay were minimized by employing interrupts whenever possible. In subsequent synchronization, the samples are collected asynchronously by the sensor nodes. Each sample from the sensor is assigned a timestamp according to its local clock. As a post-process, the time shift between the samples collected by the sensor nodes are estimated and realigned using cross-correlation, interpolation and re-sampling. The subsequent synchronization technique can be used when real-time network synchronization is not possible. The testing environment was designed to emulate the real application. A mechanical shaker was used to provide controlled input signals to the sensors in order to synchronously reconstruct the signals in time-domain.

Files