Understanding mushroom farm environment using TinyML-based monitoring devices

Abstract The optimization of environmental conditions in mushroom cultivation is pivotal for maximizing yield and quality. A Smart Environmental Monitoring System for Mushroom Farms is presented in this paper that makes use of advanced Tiny Machine Learning (TinyML) and Internet of Things (IoT) technologies for evaluating and controlling key parameters that impact the growth of mushrooms. The rapid growth of the worldwide mushroom markets indicates how important these efforts are economically. This study uses more developed instruments for tracking the temperature, humidity level, carbon dioxide concentration in the atmosphere, intensity of light, moisture content of the soil as well as pH and temperature values found within the soil itself. On the one hand, the study employed SCD30 Sensirion sensor mostly for gauging atmospheric conditions and the other (Grove-Digital sensor) for measuring various parameters specific to soils (such as moisture content, pH level, or temperature). The latter is then connected to an XIAO ESP32-S3 microprocessor chip which supports different types of connections such as WiFi or Bluetooth connections while it can also run TinyML models to enable immediate processing of data. The authors set up the system to gather environmental data on time, using the Edge Impulse platform for data analysis and model training. TinyML-enabled microcontroller processes the data locally, autonomously controlling actuators like humidifiers, heaters, and fans hence maintaining the best conditions for plants. The experimental design included situating sensors at various locations in the mushroom farm environment to monitor important parameters continually and record them. The system’s effectiveness in maintaining ideal conditions for breeding mushrooms has been carefully examined through detailed analysis. The mushroom cultivation system’s temperature and humidity were controlled between 15–22 °C and 85%–90% respectively, which led to a higher crop yield and quality improvements. By using TinyML, it enabled doing fast on-device processing without relying heavily on cloud solutions, hence reducing latency.