Low-Cost Data Acquisition (DAQ) with Arduino and Binho for Machine Learning | DigiKey

Data acquisition (DAQ) devices are normally very expensive and require proprietary software and add-on sensors. We can develop our own low-cost DAQ quite easily using a few different techniques. First, we’ll explore using an Arduino UNO as a DAQ device. We’ll write a quick script that reads raw data from an accelerometer and relays the readings out over the Serial port. We then write a Python script in Jupyter Notebook to read these measurements from the Serial port and save them in a Numpy array. Then, we take a look at using a professional host adapter tool, the Binho Nova, to perform the exact same collection process. Because the Binho Nova works with CircuitPython, we can rely on community-written libraries to handle the raw communication for us. This saves us the step of having to write low-level C/C++ firmware. Being able to read from sensors quickly and easily is an important step in many data collection systems, but is proving increasingly so in machine learning. While most machine learning (ML) and artificial intelligence (AI) systems focus on audio and visual elements, we will likely begin to see a rise in ML applied to other types of sensor data. For example, we show how to collect data from an accelerometer in this video for someone who might want to make a sort of “magic wand” interface device.