Long-term Drift Research Articles

An ASIC-based system-in-package MEMS gas sensor with impedance spectroscopy readout and AI-enabled identification capabilities

The main advantages of micromachined metal oxide semiconductor (MOX) gas sensors, in particular their low cost, high sensitivity, miniaturization and low power consumption, are compromised by lack of selectivity and long-term sensor signal drifts, which often hinder their application in real-world sensing applications. To increase the sensor data dimensionality, temperature cycled operation (TCO) and impedance spectroscopy (IS) readout was proposed, enabling the acquisition of large datasets from a single sensor, as suitable to be exploited by machine learning techniques. Furthermore, recent studies have shown how the IS readout minimizes the issues represented by the sensor long-term signal drift.The aim of this work is the development of a smart system-in-package, integrating a state-of-the-art micromachined MOX gas sensor based on a wafer-level nanostructured sensing layer with a novel versatile sensor interface chip, capable of high performance IS readout. We propose a miniaturized sensing device, which is a small, lightweight low power consumption chemical sensing micro-system with fast and accurate classification capabilities, suitable to be deployed in long multi-node sensing cables or on unmanned aerial vehicles. A software suite comprising a sensor training wizard was also developed, enabling users with no technical background to train the sensing system.

Research on Phase Stabilization Algorithm of Femtosecond Timing System

This paper presents the design, implementation, and validation of a femtosecond timing system aimed at achieving precise time control and phase synchronization for large particle accelerators. A prototype system utilizing a continuous wave laser was developed, focusing on minimizing timing jitter and long-term phase drift. Key components include an optical delay line for coarse adjustments and a fiber stretcher for fine-tuning, achieving an adjustment precision of 1 femtosecond. The system incorporates a phase detection module with a non-In-phase/Quadrature downconversion approach, enabling high-accuracy phase measurements. A collaborative algorithm was designed to optimize the interplay between the optical delay line and the fiber stretcher, utilizing a proportional-integral-derivative (PID) control algorithm to enhance adjustment precision. A Field Programmable Gate Array (FPGA) served as the core interface converter, facilitating data communication and real-time phase information acquisition. Experimental results demonstrated significant improvements in phase stability, with average phase deviation reduced from 1374.104 fs to 15.782 fs, showcasing the effectiveness of the proposed system in achieving high precision and stability in phase control. This research provides a solid foundation for future advancements in timing systems for high-frequency reference signals.

A data-driven regression model for predicting thermal plant performance under load fluctuations

The global energy demand is still primarily reliant on fossil-fueled thermal power plants. With the growing share of renewables, these plants must frequently adjust their loads. Maintaining, or ideally increasing operational efficiency under these conditions is crucial. Increasing the efficiency of such systems directly reduces associated greenhouse gas emissions, but it requires sophisticated models and monitoring systems. Data-driven models have proven their value here, as they can be used for monitoring, operational state estimation, and prediction. However, they are also sensitive to (1) the training approach, (2) the selected feature set, (3) and the algorithm used. Using operational data, we comprehensively investigate these model parameters for performance prediction in a thermal plant for process steam generation. Specifically, four regression algorithms are evaluated for the prediction of the highly fluctuating live steam flow with two training approaches and three feature subsets of the raw dataset. Furthermore, manual and automatic clustering methods are used to identify different states of operation regarding the fuel amounts used in the combustion chamber. Our results show that the live steam flow is predicted with excellent accuracy for a testing period of one month (R2=0.994 and NMAE=0.55%) when using a dynamic training approach and a comprehensive feature set comprised of 48 features representing the combustion process. It is also seen that the statically trained model predicts various load changes with strong accuracy and that the accuracy of the dynamically trained model can be approached by incorporating the cluster information into the static model. These models reflect the plant’s physical intricacies under varying loads, where deviations from the predicted live steam flow indicate unwanted long-term drifts. They can be directly implemented to help operators detect inefficiencies and optimize plant performance.

Semi-supervised comparative learning compensation method for chemical gas sensor drift.

The gradual and unpredictable variation in chemo-sensory signal responses when exposed to the same analyte under identical conditions, commonly referred to as sensor drift, has long been recognized as one of the most serious challenges faced by chemical sensors. The traditional drift compensation method is both labor-intensive and expensive, as it requires frequent collection and labeling of gas samples for recalibration. Introducing a small number of meaningful drift calibration samples can be an attractive strategy to reduce the computational load and improve the performance of the updated classifier. However, under the influence of drift, new challenges arise due to the difference in the distribution of source and target domain data. This paper proposes a novel algorithm framework called semi-supervised contrastive learning drift compensation (SSCLDC). The framework automatically extracts high-level abstract features based on a multilayer perceptron to better represent the structure of the source data. In addition, to address the issue of data distribution differences caused by drift between the source and target domains. We add a small number of reference sample pairs into the training for semi-supervised learning. Combining a contrastive loss function that can represent the matching degree of paired samples effectively overcomes the problem of sensor drift. The Kennard-Stone sequential algorithm is used to select the representative reference sample from the set of candidate reference samples. Experiments conducted on a widely used long-term chemical gas sensor drift dataset demonstrate that the proposed method outperforms several classic drift compensation techniques, highlighting its effectiveness and practical applicability.

A graphics-based digital twin (GBDT) framework for accurate UAV localization in GPS-denied environments

Autonomous navigation of Unmanned Aerial Vehicles (UAVs) is crucial for effective assessment of large-scale civil infrastructure, as manual UAV control is time consuming and prone to mishaps. Autonomous navigation outdoors typically employs GPS signals to enable accurate localization and reduce long-term drifts. However, in many civil engineering applications, GPS signals are either poor or unavailable due to interference and/or multi-path effects. Current approaches for UAV localization in GPS-denied environments, track geometric features such as corners; however, these natural features can be misrepresented due to the presence of occlusions and/or image processing errors, resulting in significant localization errors that can grow with time. AprilTags can reduce localization errors, but installation on large-scale civil infrastructure can be challenging. Therefore, this paper proposes a framework that leverages the wealth of visual and geometric information encoded in a Graphics-Based Digital Twin (GBDT) of a target infrastructure to provide accurate localization of a UAV. The GBDT is comprised of a computer graphics model that faithfully represents a target structure’s geometry, structural features, and visual textures. The visual details of the GBDT are exploited to design an object recognition algorithm that detects GBDTtags, which are distinctive objects (or a collection of components) on the target structure in advance; GBDTtags provide functionality similar to AprilTags. When the camera attached to a UAV detects one or more GBDTtags, the vertices of the GBDTtags are mapped from the image plane into the GBDT coordinate system, allowing for the UAV to be localized. The framework, termed herein as GBDTpose, is first validated numerically using Blender, Gazebo, and Mavros software-in-the-loop (SITL). Subsequently, field validation is carried out using the Kavita and Lalit Bahl Smart Bridge at the University of Illinois Urbana-Champaign (UIUC). Results show that localization in GPS-denied environments can be achieved with 5-50 cm accuracy without the need for physical markers being placed on the structure.

TOI-5005 b: A super-Neptune in the savanna near the ridge

The Neptunian desert and savanna have recently been found to be separated by a ridge, an overdensity of planets in the period range of simeq 3--5 days. These features are thought to be shaped by dynamical and atmospheric processes. However, their roles are not yet well understood. Our aim was to confirm and characterize the super-Neptune TESS candidate TOI-5005.01, which orbits a moderately bright (V = 11.8) solar-type star (G2 V) with an orbital period of 6.3 days. With these properties, TOI-5005.01 is located in the Neptunian savanna near the ridge. We used Bayesian inference to analyse 38 HARPS radial velocity measurements, three sectors of TESS photometry, and two PEST and TRAPPIST-South transits. We tested a set of models involving eccentric and circular orbits, long-term drifts, and Gaussian processes to account for correlated stellar and instrumental noise. We computed the Bayesian evidence to find the model that best represents our dataset and infer the orbital and physical properties of the system. We confirm TOI-5005 b to be a transiting super-Neptune with a radius of $R_ p oplus $ ($R_ p J $) and a mass of $M_ p oplus $ ($M_ p J $), which corresponds to a mean density of $ p g \ $. Our internal structure modelling indicates that the core mass fraction (CMF = $0.74^ $) and envelope metal mass fraction ($Z_ env $) of TOI-5005 b are degenerate, but the overall metal mass fraction is well constrained to a value slightly lower than that of Neptune and Uranus ($Z_ planet $). The $Z_ planet $/$Z_ star $ ratio is consistent with the well-known mass-metallicity relation, which suggests that TOI-5005 b was formed via core accretion. We also estimated the present-day atmospheric mass-loss rate of TOI-5005 b, but found contrasting predictions depending on the choice of photoevaporation model ($0.013 oplus $ Gyr$^ $ vs $0.17 oplus $ Gyr$^ $). At a population level, we find statistical evidence ($p$-value = $0.0092^ $) that planets in the savanna such as TOI-5005 b tend to show lower densities than planets in the ridge, with a dividing line around 1 $ g \ $, which supports the hypothesis of different evolutionary pathways populating the two regimes. TOI-5005 b is located in a region of the period-radius space that is key to studying the transition between the Neptunian ridge and the savanna. It orbits the brightest star of all such planets known today, which makes it a target of interest for atmospheric and orbital architecture observations that will bring a clearer picture of its overall evolution.

Long-term performance monitoring of a-Si 1200 electronic portal imaging device for dosimetric applications.

Recently, dosimetri applications of the electronic portal imaging device (EPID) in radiotherapy have gained popularity. Confidence in the robust and reliable dosimetric performance of EPID detectors is essential for their clinical use. This study aimed to evaluate the dosimetric performance of the a-Si 1200 EPID and assess the long-term stability of its response. Weekly measurements were performed on two clinically used TrueBeam linear accelerators (linacs) equipped with a-Si 1200 EPID detectors over a 2-year period. They included dark and flood calibration fields, and EPID response to an open field corrected for the long-term machine output drift measured with the secondary absolute dosimeters: an ion chamber and an ion chamber array. All measurements were performed using five photon beam energies and two imaging modes: continuous and dosimetry. The measurements were analyzed for constancy and the presence of long-term trends. Comparisons were made between the two linacs for each beam energy. Pixel sensitivity matrices (PSM) were determined semi-annually and analyzed for long-term constancy for both treatment machines. The long-term variation of the dark and flood field signals, integrated across the EPID plane, over the entire observation period did not exceed 0.17% and 0.79%, respectively. The output-corrected EPID response showed long-term variation from 0.28% to 0.36%, depending on beam energy, while the short-term variation was 0.04%-0.07% for EPID and 0.02%-0.06% for secondary dosimeters. The long-term variation of secondary dosimeters was 0.2%-0.3%. PSMs were found to be stable to within 1% for 97.8% of pixels and 2% for 100% of pixels. Techniques to monitor and assess the long-term performance of the a-Si 1200 EPID as a dosimeter were developed and implemented using two TrueBeam linacs. The long-term variation of the EPID response was within clinical tolerance indicated in AAPM TG-142 report, and the detector was shown to be stable and reproducible for routine clinical dosimetry.

An electronic nose drift compensation algorithm based on semi-supervised adversarial domain adaptive convolutional neural network

Sensor drift is a significant challenge leading to performance degradation in electronic nose(E-nose) systems. Effectively addressing sensor drift represents the most daunting problem in E-nose technology. This work proposes a domain adaptive approach called Semi-Supervised Adversarial Domain Adaptive Convolutional Neural Network (SAD-CNN) to tackle the long-term drift in E-nose and device displacement. SAD-CNN leverages adversarial learning to minimize distribution disparities between the source and target domains. Unlike traditional methods employing projection matrices, SAD-CNN utilizes one-dimensional convolutional neural networks as feature extractors, circumventing the complexities associated with parameter adjustments and matrix calculations in the projection process. During the training process, utilizing pseudo-labels generated by the semi-supervised self-training method to train the model, thereby reducing the labeling costs. Additionally, confidence threshold screening is introduced during the self-training phase to minimize erroneous pseudo-labels. Furthermore, the regenerative Hilbert space’s Maximum Mean Difference combined with Minimum Variance is introduced as a domain constraint function to mitigate distribution discrepancies between domains and enhance feature discriminability across domains. The experimental results demonstrate that the SAD-CNN method outperforms others. Within the long-term drift dataset, the classification accuracies for different scenarios are 78.01 % and 82.53 %, respectively. Meanwhile, the instrument change dataset yields a classification accuracy of 96.45 %.

RLI-SLAM: Fast Robust Ranging-LiDAR-Inertial Tightly-Coupled Localization and Mapping.

Simultaneous localization and mapping (SLAM) is an essential component for smart robot operations in unknown confined spaces such as indoors, tunnels and underground. This paper proposes a novel tightly-coupled ranging-LiDAR-inertial simultaneous localization and mapping framework, namely RLI-SLAM, which is designed to be high-accuracy, fast and robust in the long-term fast-motion scenario, and features two key innovations. The first one is tightly fusing the ultra-wideband (UWB) ranging and the inertial sensor to prevent the initial bias and long-term drift of the inertial sensor so that the point cloud distortion of the fast-moving LiDAR can be effectively compensated in real-time. This enables high-accuracy and robust state estimation in the long-term fast-motion scenario, even with a single ranging measurement. The second one is deploying an efficient loop closure detection module by using an incremental smoothing factor graph approach, which seamlessly integrates into the RLI-SLAM system, and enables high-precision mapping in a challenging environment. Extensive benchmark comparisons validate the superior accuracy of the proposed new state estimation and mapping framework over other state-of-the-art systems at a low computational complexity, even with a single ranging measurement and/or in a challenging environment.

Nanoporous Carbon Materials as Solid Contacts for Microneedle Ion-Selective Sensors.

Continuous sensing of biomarkers, such as potassium ions or pH, in wearable patches requires miniaturization of ion-selective sensor electrodes. Such miniaturization can be achieved by using nanostructured carbon materials as solid contacts in microneedle-based ion-selective and reference electrodes. Here we compare three carbon materials as solid contacts: colloid-imprinted mesoporous (CIM) carbon microparticles with ∼24-28 nm mesopores, mesoporous carbon nanospheres with 3-9 nm mesopores, and Super P carbon black nanoparticles without internal porosity but with textural mesoporosity in particle aggregates. We compare the effects of carbon architecture and composition on specific capacitance of the material, on the ability to incorporate ion-selective membrane components in the pores, and on sensor performance. Functioning K+ and H+ ion-selective electrodes and reference electrodes were obtained with gold-coated stainless-steel microneedles using all three types of carbon. The sensors gave near-Nernstian responses in clinically relevant concentration ranges, were free of potentially detrimental water layers, and showed no response to O2. They all exhibited sufficiently low long-term potential drift values to permit calibration-free, continuous operation for close to 1 day. In spite of the different specific capacitances and pore architecture of the three types of carbon, no significant difference in potential stability for K+ ion sensing was observed between electrodes that used each material. In the observed drift values, factors other than the carbon solid contact are likely to play a role, too. However, for pH sensing, electrodes with CIM as a carbon solid contact, which had the highest specific capacitance and best access to the pores, exhibited better long-term stability than electrodes with the other carbon materials.

Characterising and tackling thermally induced zero-drift in displacement measuring interferometry using temperature-controlled enclosure

Abstract Our research efforts in displacement measurement interferometry focused on long-term drifts initiated an extended experimental investigation in the interferometric assemblies of our design. We aimed to analyze, characterize and tackle the long-term measurement stability, expressed as the zero-drift, with particular attention to the thermal effects. For the experimentation, we developed a thermostatic chamber equipped with active temperature regulation, an array of sensors and control electronics. With either the finely stabilized temperature or with the thermal cycling, we can carry out a range of investigations: verification of modified design or prototype interferometers, testing of production pieces, characterization of integrated assemblies and units in terms of the zero drift and the susceptibility to thermal effects—the temperature sensitivity ( δ L / δ T , expressed in nm ⋅ K−1). With these experimental studies, we demonstrate the potential of the zero drift studies to contribute to the development and broader expansion of interferometric instrumentation.

High-Performance Humidity Sensor Capable for Monitoring Low-Moisture Levels in Energy Industries

Humidity sensors are extensively used in industrial processing and environmental control. There are two categories of humidity sensors: High-humidity (relative humidity) measurement and Low-humidity (moisture) measurement. The proper units for low-humidity measurement are dew point and parts per million in volume (ppmv) or parts per billion in volume (ppbv). It is very important to monitor moisture levels in natural gas during processing and transportation in pipelines. It is of critical importance to monitor moisture levels in lithium-ion battery manufacturing. In these cases, the moisture levels are limited to be less than -50°C dew point or 40 ppmv. The conventional relative humidity sensors cannot be used because of their low sensitivity. The dew point sensors on the market are polymer-based and gamma-phase-alumina based. The polymer-based sensors have low sensitivity, which can only accurately detect moisture levels of -60°C dew point and higher. The gamma-phase alumina sensors suffer from long-term drift due to phase change.We present novel high-performance humidity sensors based on porous alpha-phase alumina and silicon dioxide nanostructures, which are capable of detecting moisture levels as low as -100°C dew point or 13 ppbv with excellent long-term stability. The porous alpha-phase alumina films were produced by anodic-spark-deposition on aluminum substrates in which electric sparks were produced due to high-voltage anodization (>120V). Local high-temperature (about 2200°C) at the points of sparks converted the amorphous alumina into alpha-phase alumina (Sapphire). It provides an extremely stable structure for humidity sensors. Hydrophilic silicon dioxide films deposited by atomic layer deposition was used as sensing materials. The sensors’ sensitivity, temperature coefficient, response speed, and long-term stability were studied in details. The sensors have great promise for monitoring moisture levels in natural gas industry and lithium-ion battery manufacturing.

Controlling optical-cavity locking using reinforcement learning

This study applies an effective methodology based on Reinforcement Learning to a control system. Using the Pound–Drever–Hall locking scheme, we match the wavelength of a controlled laser to the length of a Fabry-Pérot cavity such that the cavity length is an exact integer multiple of the laser wavelength. Typically, long-term drift of the cavity length and laser wavelength exceeds the dynamic range of this control if only the laser’s piezoelectric transducer is actuated, so the same error signal also controls the temperature of the laser crystal. In this work, we instead implement this feedback control grounded on Q-Learning. Our system learns in real-time, eschewing reliance on historical data, and exhibits adaptability to system variations post-training. This adaptive quality ensures continuous updates to the learning agent. This innovative approach maintains lock for eight days on average.

The evaluation of the long-term drift of electrical capacitance measurement standards by two methods

Measurements of electrical capacitance are essential in various fields of electrical engineering, electronics and other fields. Capacitance measurements are necessary for correct design of electrical circuits and devices. Such measurements help ensure the stability and reliability of electrical systems such as power supplies, filters, capacitors, etc. In some applications, such as radio transmitters, filters, and other electronic devices, it is needed to maintain certain frequency characteristics. It is critical and appropriate to consider various factors related to the drift of the measuring instrument or measurement standard when performing measurements to ensure the required accuracy and reliability of the measurements. The study of the time drift of the measurement standard is mandatory when carrying out comparisons of national measurement standards. The types of the drift and main methods of its evaluation for measuring instruments and measurement standards between their calibrations were analysed. A conventional method of the long-term drift analysis involves the use of regression models followed by their detailed analysis. Such models are specific mathematical functions that describe theoretical values that best represent the underlying bias in the time series for the long-term drift. Exponentially Weighted Moving Average (EWMA) charts reduces the lag inherent in conventional moving averages by giving more weight to recent observations. The results of the evaluation of the long-term drift of measurement standards of electrical capacitance for high-precision calibration of measurement standards by polynomial regression diagrams and EWMA charts are given. Polynomials of the 2nd degree were sufficient to approximate the drift of electrical capacitance measurement standards under consideration. The application of EWMA charts showed greater sensitivity to drift changes over the past few years of observations compared to the regression analysis. Consistent results were obtained.

An empirical evaluation of meta residual network for classifying sensor drift samples

With the advancements in materials and pattern recognition algorithms, electronic nose (E-nose) has been continuously developing and finding increasingly widespread applications. However, E-nose is prone to drift during practical usage. Currently, the primary methods for addressing the issue of sensor drift in sample classification are semi-supervised learning and unsupervised learning. The implementation of supervised learning has been difficult in the past due to the requirement for a large number of up-to-date samples to retrain the models. But now, this challenge can be overcome with the help of meta-learning. We applied the model-agnostic meta-learning (MAML) architecture from meta-learning and designed a residual classifier with learnable per-parameter learning rates. Leveraging the ability of ‘learning to learn’, our method do not require a large amount of target domain data and there is no need to repeat extensive computations when transferring to any other target domain compared to other algorithms, resulting it highly deployable. Through the evaluation of multiple experiments, we demonstrate the feasibility of our approach and achieve very high classification accuracy on a publicly available long-term drift dataset compared with other methods.

Drift approximation by the modified Boris algorithm of charged-particle dynamics in toroidal geometry

In this paper, we study the dynamics of charged particles under a strong magnetic field in toroidal axi-symmetric geometry. Using modulated Fourier expansions of the exact and numerical solutions, the long-term drift motion of the exact solution in toroidal geometry is derived, and the error analysis of the large-stepsize modified Boris algorithm over long time is provided. Numerical experiments are conducted to illustrate the theoretical results.

Improving the error compensation capability in self-mixing interferometry

A scheme to improve the error compensation capability in self-mixing interferometry is proposed. Two beams of different frequencies are used as the measurement and reference light to compensate for the error. The short-term drift of the system is 12 nm and the long-term drift is 59 nm, and the experimental results show the effectiveness of the compensation. The performance is also evaluated, and the linearity of the frequency measurement can reach the order of 10−5, and the maximal standard deviation of the amplitude precision is 4.31 nm. The optical structure of this scheme is simple, the optical paths of measurement and reference light are easy to adjust, and the system has high robustness, which can be used as a sensor for real-time monitoring in practical engineering.

The estimation of the long-term drift of the inductance measurement standards

Inductance measurements are important in many fields, especially in electronics, electrical engineering, radio engineering, and other areas. The inductance is often an important parameter in a wide range of applications such as radio transmitters, power circuits, magnetic resonance pulsed sources, etc. The accuracy of the inductance affects the quality of products, especially in devices where inductors are used, such as filters, transformers, inverters, etc. High-precision inductance measurements are used for the product quality control to ensure that manufactured devices meet established specifications and standards.
 Drift is an undesirable property of all measuring instruments and measurement standards during their life cycle. The analysis of the instrumental drift of measurement standards is of great importance in metrology. Reliable drift accounting plays an important role in maintaining measurement accuracy. For electrical measurement standards, the long-term drift is predictable. The drift types and main methods of its estimation for measurement standards between their calibrations were analysed. The drift uncertainty can be evaluated from the history of successive calibrations, and in the absence of such history, the order of magnitude of the calibration uncertainty can be estimated.
 The results of the estimation of the long-term drift of the inductance measurement standards for high-precision calibration of measuring instruments and measurement standards by two methods, polynomial regression curves and Exponentially Weighted Moving Average (EWMA) schemes, are given. The EWMA schemes reduce the lag inherent in traditional moving averages by giving more weight to recent observations. It is shown that the use of the EWMA schemes compared to the regression analysis shows greater sensitivity to the drift changes in the last years of observations. This allows the laboratory to take this factor into account when calibrating measuring instruments and measurement standards.

Contrastive domain generalization convolution neural network correcting the drift of gas sensors

The drift problems of gas sensors caused by aging ingredients and unexpected environmental impacts, which can severely decrease the performance and service life of electronic noses, have been widely discussed. Many scholars believe that the essential reason for sensor drift can be attributed to the different data distributions in the latent feature space. However, traditional drift compensation methods like OSC, GasNet, SniffMultinose, and SniffConv are costly and complex. This paper has introduced an algorithm called contrastive domain generalization convolution neural network (CDCNN) to resolve this problem. For the first time, domain generalization and contrastive learning were used as the drift compensation methods for the gas sensor. A novel data augmentation was proposed to enrich the datasets. Feature generation module is used to simulate the drift of gas sensors. Contrastive learning adapts to the unseen areas in the latent feature space. The artificially generated features and the original features are drawn closer in the feature space to improve the algorithm's generalization ability. Experiments on long-term drift data show that CDCNN achieves high accuracy (0.7230). The experimental results show that the CDCNN algorithm is more suitable for practical applications due to less resource consumption and looser constraints.

Open-set adversarial domain match for electronic nose drift compensation and unknown gas recognition

In practical applications, electronic noses (EN) must tackle not only sensor drift but also resist interference from unknown gases. In prolonged open environments, electronic noses often encounter unknown gases that cannot be predicted in advance. Current gas detection methods typically rely on prior knowledge of target analytes but struggle to comprehensively address sensor drift and interference from unknown gases. Such as indoor air quality monitoring, long-term stability is crucial, and not all relevant analytes can be predetermined. However, the recent research cannot solve both the sensor drift and unknown gas intrusion problem simultaneously well. In this work, we unify above problems in to the open-set risk boundary. We propose an open-set adversarial domain match (OSADM) model and introduce the considers of open-set domain adaptation (OSDA). OSDA trains a target classifier through matching the domain distribution to recognize the known and unknown gases. First, a binary adversarial loss divides the class boundary. Secondly, adversarial domain adaptation unifies the distribution of different domains. Compared with the metric methods, it avoids complex distribution computation and parameter adjustment to reduce negative transfer. Extensive experimental results on two benchmark datasets, Gas Sensor Array Drift and Twin gas sensor arrays Dataset show that OSADM outperformance of the existing open-set models.