Continuous measurement of oxygen concentration is essential in production processes, environmental analysis, medical applications, etc. An electrochemical oxygen sensor measures oxygen concentration by measuring produced current on the basis of electrochemical principles. It has a relatively long response time.
A moving average filter is not sensitive to whether the phase response is linear or not, and therefore it is used in many applications for noise processing in various measurement systems. There are several types of moving average filters including simple moving average (SMA) filter, linear decay moving average (LDMA) filter and raised cosine moving average (RCMA) filter. The disadvantage of these moving average filters is that they cannot achieve sufficient denoising effects if the filter order is limited. A special case of the moving average filter is an exponential moving average (EMA) filter.
Han Chol Hak, a section head at the Faculty of Automatics, proposed an improved exponential moving average filter (EMA) by determining the optimal coefficient to suit smooth noisy signals of sensors.
First, he analyzed the variation of deterministic signal accuracy and noise reduction characteristics according to the coefficient value of EMA filter. Through simulation, he determined the optimal coefficient which depends on the signal to noise ratio (SNR) and the frequency of input signals by approximation. Then, he set sample frequency constraints to obtain good filtering effect. Finally, he designed an improved EMA filter that changes the optimal coefficient according to input signals, and compared it with a simple moving average filter (SMA) for oxygen sensor signals.
The results show that the proposed algorithm outperforms other moving average filters for smooth noisy signals.
If more information is needed, you can refer to his paper “An Improved Exponential Moving Average Filter Design and Application in Electrochemical Oxygen Sensor Signal Denoising” in “Proceedings of KUTIC-2025”.