Episode 28: Training a Neural Network with the BL5340 (Vibration Demo)
Published on August 27, 2021
The BL5340 features a dual-core ARM Cortex M33 processor, the main core of which (and the focus of this demo) is a high performance 128 MHz/64 MHz CPU. This core features a single-precision floating point unit and the ARM DSP instruction set – both ideal for running neural networks.
In this video, Field Application Engineer Mark Duncombe uses the BL5340 Development Kit's accelerometer to measure vibration frequencies from a vibration generator, and uses Edge Impulse to train a neural network on the frequencies used. With this dataset, the network can identify vibration frequencies on its own, useful for industrial applications where the frequency of equipment's vibration might be an indicator of machine condition and whether equipment must be investigated or serviced.
For more on the BL5340, visit: https://www.lairdconnect.com/bl5340-series
For the full documentation of this demo, see the guide: