<![CDATA[Dissolved Oxygen (DO) is an important parameter for maintaining water quality in aquaculture systems. The accuracy of DO sensors significantly affects the reliability of Internet of Things (IoT)-based monitoring systems. This study aimed to calibrate the DO sensor using a two-point calibration method and evaluate the accuracy of the sensor readings compared with those of a reference device (standard DO meter). A key novelty of this study lies in its multi-media calibration, performed directly on six distinct aquaculture water types, providing field-realistic validation conditions not commonly explored in previous studies. Furthermore, the accuracy of the calibrated sensor is evaluated quantitatively using MAE, RMSE, and percentage deviation to ensure rigorous performance assessment. The system was developed using an ESP32 microcontroller, DO sensor (SEN0237), DS18B20 temperature sensor, and ADS1115 ADC module. Testing was performed on six types of aquaculture water media and compared with a standard DO meter using a comparative approach. In total, n = 6 field measurement points (one stabilized reading per water medium) were used to compute MAE, RMSE, and percentage deviation. The comparison results showed that the calibrated sensor had high accuracy, with a Mean Absolute Error (MAE) of 0.1083 mg/L and a Root Mean Square Error (RMSE) of 0.2654 mg/L. Significant deviations occurred only in one type of water medium, whereas the other five showed results consistent with the reference device, indicating stable sensor readings. These findings confirm that proper calibration can improve the accuracy and reliability of IoT systems used for water-quality monitoring. Regular calibration is required to maintain the sensor performance, particularly for long-term use in dynamic aquaculture water environments.]]>
Copyrights © 2025
