A two-stage method for real-time baseline drift compensation in gas sensors

Baseline drift caused by slowly changing environment and other instability factors affects significantly the performance of gas sensors, resulting in reduced accuracy of gas classification and quantification of the electronic nose. In this work, a two-stage method is proposed for real-time sensor baseline drift compensation based on estimation theory and piecewise linear approximation. In the first stage, the linear information from the baseline before exposure is extracted for prediction. The second stage continuously predicts changing linear parameters during exposure by combining temperature change information and time series information, and then the baseline drift is compensated by subtracting the predicted baseline from the real sensor response. The proposed method is compared to three efficient algorithms and the experiments are conducted towards two simulated datasets and two surface acoustic wave sensor datasets. The experimental results prove the effectiveness of the proposed algorithm. Moreover, the proposed method can recover the true response signal under different ambient temperatures in real-time, which can guide the future design of low-power and low-cost rapid detection systems.

Online Inertial Machine Learning for Sensor Array Long-Term Drift Compensation

The sensor drift problem is objective and inevitable, and drift compensation has essential research significance. For long-term drift, we propose a data preprocessing method, which is different from conventional research methods, and a machine learning framework that supports online self-training and data analysis without additional sensor production costs. The data preprocessing method proposed can effectively solve the problems of sign error, decimal point error, and outliers in data samples. The framework, which we call inertial machine learning, takes advantage of the recent inertia of high classification accuracy to extend the reliability of sensors. We establish a reasonable memory and forgetting mechanism for the framework, and the choice of base classifier is not limited. In this paper, we use a support vector machine as the base classifier and use the gas sensor array drift dataset in the UCI machine learning repository for experiments. By analyzing the experimental results, the classification accuracy is greatly improved, the effective time of the sensor array is extended by 4–10 months, and the time of single response and model adjustment is less than 300 ms, which is well in line with the actual application scenarios. The research ideas and results in this paper have a certain reference value for the research in related fields.

Recent Progress in Smart Electronic Nose Technologies Enabled with Machine Learning Methods

Machine learning methods enable the electronic nose (E-Nose) for precise odor identification with both qualitative and quantitative analysis. Advanced machine learning methods are crucial for the E-Nose to gain high performance and strengthen its capability in many applications, including robotics, food engineering, environment monitoring, and medical diagnosis. Recently, many machine learning techniques have been studied, developed, and integrated into feature extraction, modeling, and gas sensor drift compensation. The purpose of feature extraction is to keep robust pattern information in raw signals while removing redundancy and noise. With the extracted feature, a proper modeling method can effectively use the information for prediction. In addition, drift compensation is adopted to relieve the model accuracy degradation due to the gas sensor drifting. These recent advances have significantly promoted the prediction accuracy and stability of the E-Nose. This review is engaged to provide a summary of recent progress in advanced machine learning methods in E-Nose technologies and give an insight into new research directions in feature extraction, modeling, and sensor drift compensation.

Factory Oriented Technique for Thermal Drift Compensation in MEMS Capacitive Accelerometers

Capacitive MEMS accelerometers have a high thermal sensitivity that drifts the output when subjected to changes in temperature. To improve their performance in applications with thermal variations, it is necessary to compensate for these effects. These drifts can be compensated using a lightweight algorithm by knowing the characteristic thermal parameters of the accelerometer (Temperature Drift of Bias and Temperature Drift of Scale Factor). These parameters vary in each accelerometer and axis, making an individual calibration necessary. In this work, a simple and fast calibration method that allows the characteristic parameters of the three axes to be obtained simultaneously through a single test is proposed. This method is based on the study of two specific orientations, each at two temperatures. By means of the suitable selection of the orientations and the temperature points, the data obtained can be extrapolated to the entire working range of the accelerometer. Only a mechanical anchor and a heat source are required to perform the calibration. This technique can be scaled to calibrate multiple accelerometers simultaneously. A lightweight algorithm is used to analyze the test data and obtain the compensation parameters. This algorithm stores only the most relevant data, reducing memory and computing power requirements. This allows it to be run in real time on a low-cost microcontroller during testing to obtain compensation parameters immediately. This method is aimed at mass factory calibration, where individual calibration with traditional methods may not be an adequate option. The proposed method has been compared with a traditional calibration using a six-sided orthogonal die and a thermal camera. The average difference between the compensations according to both techniques is 0.32 mg/°C, calculated on an acceleration of 1 G; the maximum deviation being 0.6 mg/°C.