Precise meshing and three-dimensional monitoring of the marine environment are very important for marine resources exploration, fishery development, and military activities. Therefore, a high-precision, low power consumption, low-cost and multiparameter integrated miniaturized sensor has been in a large-scale demand. Besides, conductivity and temperature are the basic dynamic parameters of the ocean. This paper presents a high-precision conductivity and temperature (CT) integrated sensor based on the micro-electro-mechanical system (MEMS) batch microfabrication technology, which has small size, low cost, high uniformity, and efficiency in meshing and three-dimensional marine measurements. In this study, the miniaturized CT sensor showed excellent linearity of conductivity (R2 ≥ 0.99999) and temperature (R2 ≥ 0.999) measurements. The temperature sensitivity of the CT sensor is 0.0619 ℃/Ω, and the cell constant of the CT sensor is 2.559 cm-1. The temperature test also shows the high repeatability with the variance coefficient of 0.6%. Furthermore, it shows an excellent consistency for batch. The coefficient of variance of cell constants is ±0.019 cm-1 and the 95% confidence intervals for the conductivity is demonstrated to be ±0.0048 mS/cm. The variance coefficient of the temperature sensor is only 1.8%. The results indicate that the batch microfabricated sensors are suitable for a large-scale deployment in the Marine Internet of Things.

Strain sensors are essential to structural health monitoring technology. Commercially available strain gauges currently exhibit limitations such as weak output signal, temperature sensitivity, nonlinearity, instability, and difficulties in mass production. However, the ever-expanding application scenarios of strain gauges have sparked a heightened demand for enhanced sensitivity and testing accuracy in strain detection. In this work, a novel high-sensitive semiconductor strain sensor is developed and manufactured via the MEMS process. The sensor employs dopant silicon as its sensitive material, while its design ingeniously integrates silicon-sensitive grids to amplify the gauge factor and output signal. Through the implementation of a differential output structure, the temperature dependence of the semiconductor resistor can be mitigated. Furthermore, a platinum temperature resistor is integrated to compensate for the impact of temperature variations on the measurement caused by changes in silicon Young’s modulus. The sensor exhibits a nonlinear error of less than 0.5% within the range of 0-196 με. And it possesses a gauge factor of 174, which is approximately 80 times greater than that of existing commercial strain gauges. The proposed sensor exhibits significant performance enhancements and holds promising potential for industrial applications in highly demanding structural health monitoring scenarios.

Microelectromechanical system (MEMS)-based thermal flow sensors are well known for their fast response and compact size in gas flow measurement. Nevertheless, these sensors exhibit superior performance at low gas flow rates, encountering significant limitations in accuracy and anti-disturbance under high gas flow conditions. This paper presents a bypass thermal mass flowmeter (TMF) based on MEMS sensor core that extended the measurement range by optimizing the length and hole diameter of the restrictor to determine the suitable bypass ratio. Meanwhile, by combining the Sage-Husa adaptive robust KF with the strong tracking KF, we propose a Modified Sage-Husa Adaptive Robust-Strong Tracking Kalman Filter (MSHAR-STKF) to ensure high accuracy and rapid response in flow measurement. The results indicate that the flowmeter achieves a measurement range of 0-15 m ² /h with a relative error within 1.5%. By comparing the Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) of static signals under various disturbances, the MSHAR-STKF exceeds the Kalman Filter (KF), Sage-Husa Adaptive Kalman Filter (SHAKF), Strong Tracking Kalman Filter (STKF), and Adaptive Robust Kalman Filter (ARKF) in suppressing outliers and biases. Additionally, the dynamic response time of MSHAR-STKF is also the fastest among above algorithms with about 1s of rise and fall times.