A12 Sensors, IoT and Edge Computing

Topic

Internet of Things (IoT) systems integrate three components: sensors that capture a physical quantity, local processing that interprets that data, and wireless communication that transmits it to where it is needed. Many common physical sensors—sensitive to temperature, mechanical strain, or light intensity—operate by changing their electrical resistance in proportion to the quantity being measured (a concept of resistance previously described in the catalog's second model). To convert this small change in resistance into an easily measurable voltage signal, a measurement bridge is typically used to proportionally amplify any imbalance caused by that variation.

Before digitizing a sensor signal, it is advisable to condition it: amplifying and filtering the signal allows for full utilization of the available dynamic range of the converter (described in the ADC/DAC model) and eliminates unwanted noise before it can compete with the signal for that same range.

Choosing a wireless communication protocol for an IoT device always involves a trade-off between range, speed, and energy consumption—a matter of energy management previously covered in the microcontroller model. Different applications favor different protocols depending on which of these three factors is most critical for the specific use case.

Edge computing shifts some intelligent processing—increasingly including machine learning model inference—directly onto the device itself, rather than relying on sending all data to a remote server. This reduces response latency, the energy consumption associated with data transmission, and the reliance on constant connectivity.

Thus, this module—which begins with a sensor that translates a physical quantity into resistance—concludes by demonstrating that the rest of the system (signal conditioning, communication protocols, and edge processing) exists to solve a single fundamental problem: extracting as much useful information as possible from a weak, noisy signal while consuming the least amount of energy in the process.

ID:4070