Page 29 - 2020 Interconnect Innovations eBook
P. 29

Driver 4: Edge Computing in IoT Devices Further Enhances Wireless Communications
Many years ago, Gordon Moore famously predicted that performance in digital devices would double approximately every 18 months. This prediction known as Moore’s Law has generally held true, to the point that there is now tremendous computing power in the palm of your hand, or in your wearable device (e.g., smart watch). This has enabled edge computing, or the ability to process data at or near the end/edge of the network, rather than send that data in raw form all the way back to a central station for processing.
For a wireless vibration sensor, a perhaps not-so-obvious application of edge computing is calculating the fast Fourier transform (FFT) of a sampled vibration waveform at the sensor itself. In a conventional system, the raw vibration waveform would be sent to the central station as an analog signal and the FFT would be calculated there. With edge computing, the FFT can be calculated at the sensor level and the processed data can be sent back instead. The processed data includes information about the power spectral density of the FFT, including the number of peaks (up to eight total), the total energy, integration size, frequency, and magnitude of each peak, and the temperature measured at the sensor. When compared to sending back raw vibration signals, calculating FFT at the edge reduces bandwidth overhead and battery power drain. But this is only a simple example. Ultimately, much more computing could be done at the sensor. Given the appropriate algorithms, the sensor could even “learn” about the machine it is installed on and when it is running well and when it is not.
The building blocks for truly smart wireless condition-monitoring vibration sensors are now solidly in place. Visit TE Connectivity to learn more.

   27   28   29   30   31