Faulty Sensor Research Articles

A sensor fault detection, isolation, and estimation method for intelligent vehicles

Accurate and reliable sensor information is crucial for intelligent vehicles to ensure safety and reliability. It is significant to devise effective solutions to diagnose possible sensor faults. In this paper, a sensor fault detection, isolation, and estimation approach is developed for intelligent vehicle-integrated motion systems. The Luenberger observer is designed and embedded into the fault diagnosis framework to real-time estimate the system states. Interval observers are constructed for vehicle subsystems to estimate the interval bounds of the corresponding observation errors. The estimation errors enter into the corresponding intervals generated by the adaptive laws when the sensor fault is free. In the event of a sensor fault, it can be detected efficiently and the faulty sensor can be isolated. Moreover, occurred sensor fault can be estimated by modifying the interval observer structure. The simulation results in the standard J-Turn test scenario are provided to evaluate and verify the effectiveness of the proposed method.

Dual-Gate Organic Thin-Film Transistor and Multiplexer Chips for the Next Generation of Flexible EG-ISFET Sensor Chips.

Ion-sensitive field-effect transistors (ISFETs) are used as elementary devices to build many types of chemical sensors and biosensors. Organic thin-film transistor (OTFT) ISFETs use either small molecules or polymers as semiconductors together with an additive manufacturing process of much lower cost than standard silicon sensors and have the additional advantage of being environmentally friendly. OTFT ISFETs' drawbacks include limited sensitivity and higher variability. In this paper, we propose a novel design technique for integrating extended-gate OTFT ISFETs (OTFT EG-ISFETs) together with dual-gate OTFT multiplexers (MUXs) made in the same process. The achieved results show that our OTFT ISFET sensors are of the state of the art of the literature. Our microsystem architecture enables switching between the different ISFETs implemented in the chip. In the case of sensors with the same gain, we have a fault-tolerant architecture since we are able to replace the faulty sensor with a fault-free one on the chip. For a chip including sensors with different gains, an external processor can select the sensor with the required sensitivity.

Federated Learning-Based Spectrum Occupancy Detection.

Dynamic access to the spectrum is essential for radiocommunication and its limited spectrum resources. The key element of dynamic spectrum access systems is most often effective spectrum occupancy detection. In many cases, machine learning algorithms improve this detection's effectiveness. Given the recent trend of using federated learning, we present a federated learning algorithm for distributed spectrum occupancy detection. This idea improves overall spectrum-detection effectiveness, simultaneously keeping a low amount of data that needs to be exchanged between sensors. The proposed solution achieves a higher accuracy score than separate and autonomous models used without federated learning. Additionally, the proposed solution shows some sort of resistance to faulty sensors encountered in the system. The results of the work presented in the article are based on actual signal samples collected in the laboratory. The proposed algorithm is effective (in terms of spectrum occupancy detection and amount of exchanged data), especially in the context of a set of sensors in which there are faulty sensors.

Self-detection based fault diagnosis for wireless sensor networks

Wireless sensor network (WSN) is a collection of low-power and low-cost sensor nodes (SNs) deployed stochastically in the monitoring area to support various applications. The random occurrence of faulty sensor nodes affects the quality of service in WSN. Most of the existing fault detection algorithms are based on statistics, thresholds, majority voting, hypothetical tests, comparison, or machine learning and depend on the neighbor's sensing information to detect the fault status of SNs. Hence, these fault detection algorithms suffer from low detection accuracy (DA) and a high false alarm rate (FAR). Moreover, existing fault detection techniques are not scalable for large-scale WSNs, as they suffer from high energy overhead and detection latency. Furthermore, existing fault diagnosis algorithms burden an SN's resources while estimating the fault status of the WSN. This paper proposes a deep belief network (DBN) based self-detection algorithm to address these issues. The performance of the proposed DBN based self-detection algorithm has been evaluated using google colab, and MATLAB. The simulation results of the proposed algorithm has been compared with the existing fault detection algorithms. Simulation results show that the proposed algorithm performs better than the existing algorithms in terms of DA, FAR, and false positive rate. The performance of the proposed algorithm has also been evaluated in other network parameters such as energy consumption, and fault detection latency.

SNRepair: Systematically Addressing Sensor Faults and Self-Calibration in IoT Networks

An accurate and robust technique for sensor fault diagnosis proves useful for an uninterrupted supply of correct monitoring data across the Internet of Things (IoT) network. The manual checking and calibration of thousands of sensors deployed in IoT network is a challenging task. Further, most calibration techniques require additional hardware support for calibration. To address these issues, a unique IoT-based framework, SNRepair, has been proposed that uses modified Deep Reinforcement Learning (DRL) for detecting different types of sensor faults, such as bias, drift, complete failure and precision degradation. This technique is also capable of self-calibrating the faulty sensors automatically using the historical healthy data of the same sensor. The fault detection module identifies four types of sensor faults with 96.17% accuracy. The self-calibration module can calibrate the faulty sensors within a few seconds, thereby ensuring the uninterrupted availability of accurate monitoring information. The proposed model is robust as it can work efficiently even in noisy environments with accurate results.

Research on designated calibration method of fault sensor in photovoltaic thermal heat pump system based on fault detection and virtual calibration

With the increasingly serious global energy problem, clean energy sources such as solar energy have become the mainstream focus of development. Among these, solar Photovoltaic thermal (PVT) heat pump systems have promising prospects. However, incorrect data transmitted by sensors can significantly impact the operation and control of the entire heat pump system, leading to reduced efficiency. Given the specific nature of PVT heat pump systems, their internal sensors are prone to errors during operation. To address these challenges, the Autoencoder Virtual in-situ calibration (AE-VIC) is applied to PVT heat pump systems. Preliminary studies have shown that this method can effectively reduce systematic and random errors in sensors. However, the current AE-VIC method faces certain issues, including unclear calibration targets and inefficient calibration of multiple sensors simultaneously, making it difficult to implement in practical systems. In order to overcome the limitations, fault detection is integrated with AE-VIC. By combining the feature extraction capability of Autoencoder with a Softmax classifier, sensors with faults can be identified before the overall calibration process, making the calibration objective of the AE-VIC more targeted. Following fault detection, inputs of the AE model are optimized using the mRMR algorithm for the identified faulty sensors. This optimization alleviates the difficulty of calibration. Through validation with actual system, the improved method effectively diagnoses and locates faulty sensors, and subsequently calibrates them. After calibration, the sensor system error can be reduced by over 95%. The improved calibration method surpasses the original AE-VIC method in terms of both time and accuracy.

Fault Detection Strategy for Fork Displacement Sensor in Dual Clutch Transmission via Deep Long Short-Term Memory Network

Detecting the fork displacement sensor fault is critical for ensuring the reliability and safety of a dual clutch transmission (DCT). In this paper, a deep learning method is proposed to monitor the state of the fork displacement sensor. Firstly, the fork displacement prediction algorithm is developed based on the deep long short-term memory (LSTM) network using the driving data of a DCT vehicle. Secondly, the synchronizer control system model is constructed to imitate the fork displacement sensor fault as the experimental vehicle works in normal condition and the collected data lacks of faulty sensor signal. Finally, the residual is obtained by comparing the predicted fork displacement and the measured sensor information. The sensor fault is detected as the residual exceeds the predetermined threshold. Results show the fork displacement prediction algorithm can accurately estimate the synchronizer position. And the fault detection method can detect the fork displacement sensor fault timely and accurately.

Estimation‐Based Event‐Triggered Adaptive Terminal Sliding Mode Control without Pressure Sensors for a Polymer Electrolyte Membrane Fuel Cell Air Feeding System

Appropriate control for a polymer electrolyte membrane fuel cell air feeding system is vulnerable to load current disturbance, parameter uncertainties, measurement noise, faulty sensors, and limited communication resources. Considering these practical issues, this study presents a novel estimation‐based event‐triggered nonlinear control scheme without pressure sensors to regulate the oxygen excess ratio (OER) at its optimal reference. First, to exactly reconstruct the unmeasured OER, a uniform robust exact differentiator and an adaptive high gain observer are developed to estimate the supply manifold pressure and the cathode pressure, respectively, thus redundant sensors and development cost can be saved. Then, a reduced‐order high gain observer is constructed to approximate the unmodeled lumped disturbance consisting of external disturbances and system uncertainties, which helps to decrease the control gain and improve the robustness. Further, an event‐triggered adaptive terminal sliding mode controller with disturbance compensation is proposed to guarantee accurate and fast OER tracking control with integral absolute error (IAE) of 0.447 and average settling time of 0.9 s while avoiding chattering effect and unnecessary communication in the controller‐to‐actuator channel. The comparison results illustrate that the proposed approach achieves superior estimation and control of OER with satisfactory robustness to parameter perturbations (less than 4% deviation in IAE).

Fault tolerant control for a class of nonlinear systems with multiple faults using neuro-dynamic programming

Considering the multiple fault scenario, i.e., both actuator and sensor faults occur simultaneously, this paper develops a neuro-dynamic programming (NDP)-based fault tolerant control (FTC) scheme for a class of nonlinear systems. By combining a descriptor observer with an adaptive observer, system states and multiple faults are estimated simultaneously. For the nominal system, i.e., the fault-free system, a critic neural network (NN) is employed to solve the Hamilton–Jacobi-Bellman (HJB) equation, and the approximate optimal control policy is obtained. To eliminate the influence of sensor faults, the accurate estimations of system states are used instead of system states from faulty sensors to construct the approximate optimal control policy. Then, by combining the estimations of actuator faults with the approximate optimal control policy, an FTC law is developed to suppress the influence of actuator faults. The stability of the closed-loop nonlinear system is analyzed to be uniformly ultimately bounded via the Lyapunov stability theorem. The effectiveness of the present FTC scheme is demonstrated by two simulation examples.

Virtual Sensor Design for Replacement the Faulty Physical Sensors

Abstract: The paper considers the problem of virtual sensor design for nonlinear dynamic systems with non-smooth nonlinearities described by continuous-time models for faulty physical sensor replacement. It is assumed that to solve the problem, the system is equipped by diagnostic system allowing detecting and isolating the faulty sensor. For every such a sensor, the virtual sensor generating estimate replacing the faulty sensor is designed. To solve the problem, so-called logic-dynamic approach is used which does not guarantee optimal solution but uses only methods of linear algebra to solve the problem for systems with non-smooth nonlinearities. The virtual sensor can be designed in the identification canonical form or Jordan canonical form. The advantage of the first form is a standard procedure of the virtual sensor design while Jordan form allows obtaining a simpler solution. The relations allowing designing the virtual sensor both in identification and in Jordan canonical form are derived.

Joint Model-Order and Robust DoA Estimation for Underwater Sensor Arrays

The direction-of-arrival (DoA) estimation algorithms have a fundamental role in target bearing estimation by sensor array systems. Recently, compressive sensing (CS)-based sparse reconstruction techniques have been investigated for DoA estimation due to their superior performance relative to the conventional DoA estimation methods, for a limited number of measurement snapshots. In many underwater deployment scenarios, the acoustic sensor arrays must perform DoA estimation in the presence of several practical problems such as unknown source number, faulty sensors, low values of the received signal-to-noise ratio (SNR), and access to a limited number of measurement snapshots. In the literature, CS-based DoA estimation has been investigated for the individual occurrence of some of these errors but the estimation under joint occurrence of these errors has not been studied. This work investigates the CS-based robust DoA estimation to account for the joint impact of faulty sensors and low SNR conditions experienced by a uniform linear array of underwater acoustic sensors. Most importantly, the proposed CS-based DoA estimation technique does not require a priori knowledge of the source order, which is replaced in the modified stopping criterion of the reconstruction algorithm by taking into account the faulty sensors and the received SNR. Using Monte Carlo techniques, the DoA estimation performance of the proposed method is comprehensively evaluated in relation to other techniques.

Parallel crosschecking neural network based fault-tolerant flight parameter estimation and faulty sensor identification

Reliable flight parameter estimation is crucial for the safe operation of unmanned air vehicles. Recent studies on flight parameter estimation based on distributed airflow sensors have shown promising results; however, how to extract reliable flight parameters and locate faults in case of sensor failures are not studied yet. This paper proposes a novel fault-tolerant flight parameter estimation method with capabilities to identify and isolate faulty sensors for a flight parameter estimation system based on distributed airflow sensors. First, the distributed sensors are grouped into subsets using combinations. Second, feedforward backpropagation neural networks of the same structure are designed and used by all subsets for flight parameter estimation. The crosschecking between these parallel networks improves the robustness of the estimation. In case of sensor failures, the sensor groups that contain any faulty sensors will result in wrong flight parameter estimations. Especially, a directional random sample consensus algorithm is developed and used to identify the subsets that contain faulty sensors. Then, the ratio of the faulty subsets is used to identify the number of faulty sensors. Finally, the faulty sensors will be identified according to the frequencies of their appearance in the subsets that contain faulty sensors. The effectiveness of the proposed method is validated by using wind tunnel tests. The results show that the proposed method can effectively provide reliable flight parameters and identify faulty sensors.

Machine learning-assisted heat transport modelling for full-scale emplacement experiment at Mont Terri underground laboratory

Machine learning (ML)-assisted modelling of deep geological repositories (DGR) is of emerging interest and can help to improve the safe and reliable operation of DGRs as well as the public acceptance. Here, a concept of ML-assisted physical-based 3D heat transport model for the Full-scale Emplacement (FE) experiment performed at Mont Terri Underground Laboratory (URL), is presented. The FE experiment is a 1:1 scale mock-up of a DGR tunnel where heaters simulate emplaced high level waste. ML is applied to sparse sensor data of the water saturation degree in the granulated bentonite material (GBM) yielding a physically and neural network (NN) based surrogate model for the thermal conductivity of the GBM, needed to calculate the temperature evolution in the FE tunnel near-field. In order to investigate the dominant parameters influencing the temperature evolution in the vicinity of the FE experiment, the results of 32 orthogonal test cases have been analysed systematically.For the ML predicted water saturation degree, three NN methods are tested. The Elman NN (with a Pearson's r coefficient of 0.9911, a mean squared error (MSE) of 4.62, and mean absolute error (MAE) of 1.44) operates better than the back propagation (BP) and cascade BP (CBP) NN methods. Results of ML-assisted heat transport calculations are validated with the large experimental dataset of 137 × 10 temperature sensor data points (errors range within 7%). Parameters uncertainty ranges of ±10% for thermal conductivities of the GBM λGBM and bentonite block λblock are analysed and bands of temperature uncertainties are compared with temperature sensor data, which allows sensor data assessment including identification of the faulty sensor data. The ML-assisted physical modelling framework can be applied to future DGRs.

Virtual sensors for erroneous data repair in manufacturing a machine learning pipeline

Manufacturing converts raw materials into finished products using machine tools for controlled material removal or deposition. It can be observed using sensors installed within and around machine tools. These sensors measure quantities, such as vibrations, cutting forces, temperature, currents, power consumption, and acoustic emission, to diagnose defects and enable zero-defect manufacturing as part of the Industry 4.0 vision. The continuity of high-quality sensor data streams is fundamental to predicting phenomena, such as geometric deformations, surface roughness, excessive coolant use, and imminent tool wear with adequate accuracy and appropriate timing. However, in practice, data acquired by some sensors can be of poor quality and unsuitable for prediction due to sensor faults stemming from environmental factors. In this paper, we answer if we can repair erroneous data in a faulty sensor based on data simultaneously available in redundant sensors that observe the same process. We present a machine learning pipeline to synthesize virtual sensors that can step in for faulty sensors to maintain reasonable quality and continuity in sensor data streams. We have validated the synthesized virtual sensors in four industrial case studies.

A Smart Battery Management System for Electric Vehicles Using Deep Learning-Based Sensor Fault Detection

Battery sensor data collection and transmission are essential for battery management systems (BMS). Since inaccurate battery data brought on by sensor faults, communication issues, or even cyber-attacks can impose serious harm on BMS and adversely impact the overall dependability of BMS-based applications, such as electric vehicles, it is critical to assess the durability of battery sensor and communication data in BMS. Sensor data are necessary for a BMS to perform every operation. Effective sensor fault detection is crucial for the sustainability and security of electric vehicle battery systems. This research suggests a system for battery data, especially lithium ion batteries, that allows deep learning-based detection and the classification of faulty battery sensor and transmission information. Initially, we collected the sensor data, and preprocessing was carried out using z-score normalization. The features were extracted using sparse principal component analysis (SPCA), and enhanced marine predators algorithm (EMPA) was used for feature selection. The BMS’s safety and dependability may be enhanced by the suggested incipient bat-optimized deep residual network (IB-DRN)-based false battery data identification and classification system. Simulations using MATLAB (2021a), along with statistics, machine learning, and a deep learning toolbox, along with experimental research, were used to show and assess how well the suggested strategy performs. It is shown to be superior to traditional approaches.

Challenges, limitations, and measurement strategies to ensure data quality in deep-sea sensors

In this paper we give an overview of factors and limitations impairing deep-sea sensor data, and we show how automatic tests can give sensors self-validation and self-diagnostic capabilities. This work is intended to lay a basis for sophisticated use of smart sensors in long-term autonomous operation in remote deep-sea locations. Deep-sea observation relies on data from sensors operating in remote, harsh environments which may affect sensor output if uncorrected. In addition to the environmental impact, sensors are subject to limitations regarding power, communication, and limitations on recalibration. To obtain long-term measurements of larger deep-sea areas, fixed platform sensors on the ocean floor may be deployed for several years. As for any observation systems, data collected by deep-sea observation equipment are of limited use if the quality or accuracy (closeness of agreement between the measurement and the true value) is not known. If data from a faulty sensor are used directly, this may result in an erroneous understanding of deep water conditions, or important changes or conditions may not be detected. Faulty sensor data may significantly weaken the overall quality of the combined data from several sensors or any derived model. This is particularly an issue for wireless sensor networks covering large areas, where the overall measurement performance of the network is highly dependent on the data quality from individual sensors. Existing quality control manuals and initiatives for best practice typically recommend a selection of (near) real-time automated checks. These are mostly limited to basic and straight forward verification of metadata and data format, and data value or transition checks against pre-defined thresholds. Delayed-mode inspection is often recommended before a final data quality stamp is assigned.

A Zero-Shot Soft Sensor Modeling Approach Using Adversarial Learning for Robustness Against Sensor Fault

Soft sensors are widely used in many industrial systems to monitor key variables that are difficult to measure, using measurements from other available physical sensors. Because physical sensors are susceptible to faults, it is crucial for soft sensor models to be robust against them. Recently, deep learning has shown promising results in developing data-driven soft sensors for various applications. However, existing learning-based soft sensors are still vulnerable to sensor faults which could deteriorate the performance of the models. In this paper, we propose a deep learning-based modeling framework for developing soft sensor models that are robust to sensor faults. Due to the difficulty in obtaining datasets that cover all possible sensor fault characteristics, the proposed framework is developed to be zero-shot such that the model can be trained with only fault-free dataset without requiring any sensor fault patterns, thus greatly saving the time and resources needed to collect such data. Instead, adversarial examples are used as a proxy for faulty sensor inputs so that the model can learn to be adaptive through the proposed two-stage, uncertainty-aware recurrent neural network architecture. We demonstrate our approach on the TE benchmark process and a real industrial multiphase flow process, and show that robustness is achieved as the accuracy does not degrade significantly when sensor faults are present during model evaluation.

Active Power Decoupling Control for PWM Converter Considering Sensor Failures

The power converters integrated with active power decoupling (APD) may handle the ripple power in dc link and then reduce the volume of the filter capacitor dramatically. Generally, voltage and current signals of the LC branch are required for the closed-loop APD control. However, sensor failures could degrade the control performance and even lead to the breakdown of the converter system. This article proposes a detection method for sensor failures in the APD circuit of the single-phase converter. The voltage and current of the LC branch in the APD circuit are estimated by a dual sliding mode observer (SMO). The residuals of the estimated and measured signals are calculated, and they have high sensitivity to specific sensor faults, namely, the current residual is sensitive to the fault of the current sensor and the voltage residual is sensitive to the fault of the voltage sensor, which could be employed to locate the faulty sensors. Finally, the effectiveness of the proposed method is verified by the simulation and experimental results.

Sensor fault detection and isolation for smart irrigation wireless sensor network based on parity space

<p>In the recent years, wireless sensor network technology (WSN) has been widely adopted in precision agriculture for determining the needs of the soil in term of water by monitoring some environmental parameters. To do this, WSN is constructed using several sensor nodes; these small sensing devices are prone to failure and may produce erroneous measurements. To ensure good management of freshwater, the network service quality is necessary. To avoid the degradation of service, the detection of the faulty sensor in WSN is required. In this paper, a fault detection and isolation (FDI) algorithm derived from a parity space approach and based on direct redundancy is proposed toward detecting and isolating sensor fault in WSN. In laboratory experiments, the proposed FDI algorithm proved its effectiveness.</p>

Optimization of PZT Wafers’ placement with enhanced level of robustness using multiple interacting networks

Continuous monitoring and in particular structural health monitoring (SHM) of structures is used, nowadays, by asset managers to ensure a high level of safety and extend the lifetime of their assets. One of the main bottlenecks faced in the development of an SHM system is the design and implementation of sensor networks. This includes selecting the sensor type and optimizing the number and positions of sensors. This study proposes a novel approach to account for the potential loss in sensing elements within the network, by having multiple independent networks but at the same time interacting for enhanced performance. The developed model considers the PZT locations as continuous variables where any point on the surface can be a sensor location. The model can handle complex structures and customize to different wave characteristics and coverage levels.The objective function in the developed model aims at maximizing a weighted sum of two dependent terms: the coverage which is the percentage of covered control points within the network, and the robustness which is the average number of networks covering each covered control point. The proposed model was solved using a genetic algorithm and was tested on several geometries. To mimic a faulty sensor within the network, a complete network was removed in each case, and it was found that the coverage was not impacted. The effect of the partitioning of large sections on the performance of the optimizer was explored.Experimental validations were carried out to evaluate the model's accuracy in damage localization within the optimized sensor networks and to emphasize the importance of each of the coverage and robustness. The results demonstrated the proficiency of the model developed in distributing the piezoelectric ceramics (PZTs) on the tested specimens, and its ability to detect damages more precisely in more robust configurations where multiple networks participate in covering wide regions.