Skip to Main Content
Modern monitoring systems require high-performance instrumentation, low-cost primary sensors, online signal processing, portability, and versatility. Smart sensor utilization is an approach to comply with these features. From an industrial point of view, voltage, electric current, and vibrations are signals commonly analyzed during system monitoring, and wavelet and wavelet packet transforms are popular techniques for extracting time-frequency information from these signals. Here, a system-on-chip-based smart sensor with reconfigurable architecture is proposed for time-frequency analysis applying either wavelet or wavelet packet transform. The proposed smart sensor uses a proprietary hardware processing unit and a custom embedded microprocessor. It can connect to different primary sensors without additional instrumentation and shows obtained results in a 2-D or 3-D view on a video graphics array screen or liquid crystal display. Obtained results from three different study cases demonstrate that the proposed smart sensor is able to show online and in real-time the frequency evolution effects in time-varying signals.