Sensor Delay Research Articles

Adaptive estimation of UAV altitude in complex indoor environments using degraded and time-delayed measurements with time-varying uncertainties

A novel approach for robust Unmanned Aerial Vehicle (UAV) altitude estimation relying on laser measurements that is designed for use in complex indoor environments is proposed in this paper. Specifically, we aim to design a system with general usability inside multi-floor buildings. The multi-floor buildings are characterized by areas lacking distinct vertical geometric features to be used as reference by 3D Light Detection and Ranging (LiDAR) localization algorithms, and by areas with either flat floors or limited areas with inconsistent ground elevation. The proposed approach solves the problem of adaptive fusion of data from multiple sources with apriori-unknown confidence dependent on the current environmental properties. Whenever the environment contains enough geometric structure, altitude data from a 3D LiDAR-based Simultaneous Localization and Mapping (SLAM) algorithm are utilized. In environments that are too symmetrical for reliable SLAM operation, the approach relies mostly on measurements from a downward-facing 1D laser rangefinder, while simultaneously detecting inconsistent ground elevation areas. These measurements are fused with barometer data, Inertial Measurement Unit (IMU) data, and information from the UAV position controllers. Furthermore, our approach correctly handles the measurement delay caused by 3D LiDAR data processing that significantly differs from other sensor delays. The performance of the proposed approach has been validated in complex simulations and real-world experiments with the produced altitude estimate utilized in the control loop of the UAV. The proposed approach is released as open-source as part of the MRS UAV System.

Characterization and optimization of 3D-printed, flexible vibration strain sensors with triply periodic minimal surfaces

Flexible piezoresistive strain sensors have promising applications in wearables and soft robotics. For sensing dynamic strains, such as a runner’s gait or a slipping object held by a robotic gripper, the sensor must be capable of vibration strain sensing for a range of amplitudes and frequencies. This article presents the characterization and design optimization of a flexible sensor using Schwarz P-type triply periodic minimal surface (TPMS) structures for vibration strain sensing. Sensors are fabricated using a 3D-printing process, coating a silicone rubber (SR) matrix with graphene nanoplatelets (GNP). Sensors are characterized under uniaxial compressive strain amplitudes from 0% to 10% and frequencies of 10–110 Hz. Frequency and time-domain analyses demonstrate sensor performance and explain the unique deformation mechanisms of the TPMS structure. Low sensor delays of less than 6.3 ms demonstrate its capability for high-frequency sensing. The frequency independence of the sensor is demonstrated, as the mean frequency dependence error is only ±3.89%. In addition, the TPMS sensor was parametrized, and a size optimization was performed on it using a multi-objective adaptive firefly algorithm (MOAFA). The MOAFA converged within 1618 unique function evaluations, a reduction of 85.5% compared to the entire set of feasible solutions. The optimized sensor was fabricated and characterized, displaying a reduced mean frequency dependence error of only ±2.18%.

Comparative analysis of AEB effectiveness based on typical and atypical scenarios of electric two-wheeler accidents in China

Electric two-wheeler (ETW) accidents are one of the most important types of traffic accidents in China. Currently, the China New Car Assessment Program (C-NCAP) has proposed test scenarios to evaluate the effectiveness of autonomous emergency braking (AEB) for ETWs, including several common typical crash scenarios, but other atypical scenarios have not been considered. To determine the performance of AEB in real accidents, 16 in-depth accident cases with typical scenarios and 11 cases with atypical scenarios were selected based on a proposed C-NCAP typical scenario set and reconstructed using the virtual simulation tools MADYMO and PC-Crash. The crashes were re-simulated with a car equipped with an AEB system while varying the sensor field of view (FOV), time-to-collision (TTC), sensor delay time (SDT), and lateral trigger width ( W). The results show that for almost all combinations of AEB parameters, the crash avoidance rate was much higher in the typical scenario than that in the atypical scenario. When using an AEB with a FOV of 90° (±45°), all ETW accidents were avoided in typical scenarios, while even with the most efficient AEB system (FOV = ±60°, TTC trigger value = 1.5 s, SDT = 0.1 s), only 82% of crashes were avoided in atypical scenarios. Further considering the effect of lateral width, increasing the width from 2 to 5 m, the maximum avoidance rate of the AEB system increased by 43% in the typical scenarios and 18% in the atypical scenarios. The findings suggest that the typical AEB test scenarios proposed for C-NCAP were useful for other crash scenarios, but that including additional test scenarios may better reflect real world crash scenarios. It is recommended that atypical scenarios should be considered in C-NCAP, particularly perpendicular crash scenarios with the car or ETW turning, to better describe real accidents and improve vehicle safety.

Event‐based optimal filter for a networked system with multiplicative and auto/cross‐correlated process and measurement noise

SummaryThis article presents an event triggered based optimal filter for a networked system with multiplicative noise in state and measurement matrices. The process and measurement noise are auto/cross‐correlated, which is described as a moving average process. The network in both sensor to estimator (S‐E) and controller to actuator (C‐A) channel is affected by random sensor delay and packet dropouts. These effects are modeled using two random variables for both the S‐E and C‐A channels. Sensor data are transmitted to the estimator through the S‐E channel using an event‐based mechanism. It depends on a time varying threshold to improve trade‐off between tracking performance and utilization of network bandwidth. The process and measurement noise are two‐step auto/cross‐correlated. A comprehensive augmented stochastic model is developed for the networked system by combining the effects of event triggering, network constraints and noise correlation. The filter is designed using an innovation analysis approach based on orthogonal projection. The solution of the filter is given in terms of a Lyapunov and a Riccati equation. An algorithm is presented for the implementation of developed filter. The stability properties of the proposed optimal filter are analyzed. The effectiveness of the proposed filter is shown in the simulation platform and accumulated mean square error performance is evaluated based on Monte Carlo simulation.

A Quantitative Parametric Study on Output Time Delays for Autonomous Underwater Cleaning Operations

Offshore pipelines and structures require regular marine growth removal and inspection to ensure structural integrity. These operations are typically carried out by Remotely Operated Vehicles (ROVs) and demand reliable and accurate feedback signals for operating the ROVs efficiently under harsh offshore conditions. This study investigates and quantifies how sensor delays impact the expected control performance without the need for defining the control parameters. Input-output (IO) controllability analysis of the open-loop system is applied to find the lower bound of the H-infinity peaks of the unspecified optimal closed-loop systems. The performance analyses have shown that near-structure operations, such as pipeline inspection or cleaning, in which small error tolerances are required, have a small threshold for the time delays. The IO controllability analysis indicates that off-structure navigation allow substantial larger time delays. Especially heading is vulnerable to time delay; however, fast-responding sensors usually measure this motion. Lastly, a sensor comparison is presented where available sensors are evaluated for each ROV motion’s respective sensor-induced time delays. It is concluded that even though off-structure navigation have larger time delay tolerance the corresponding sensors also introduce substantially larger time delays.

Partial-Nodes-Based State Estimation for Stochastic Coupled Complex Networks with Random Sensor Delay: An Event-Triggered Communication Method

Partial-Nodes-Based State Estimation for Stochastic Coupled Complex Networks with Random Sensor Delay: An Event-Triggered Communication Method

State estimation of nonlinear time-varying complex networks with time-varying sensor delay for unknown noise distributions

The state estimation for nonlinear time-varying complex networks is a common problem, especially for the systems with time-varying sensor delay. This paper presents an artificial immune strategy based on particle filtering algorithm to estimate the states. Firstly, considering the nonlinear time-varying complex networks with time-varying sensor delay, this paper gives a theoretical proof that particle filter can be used for the state estimation. On this basis, this paper adopts the artificial immune algorithm to solve the problem which is particle impoverishment in particle algorithm. This paper gives the convergence of the proposed algorithm theoretically. Furthermore, for nonlinear time-varying complex networks with unknown noise distributions, this paper gives an improved algorithm which uses an improved kernel density estimation algorithm which has window width varying with probability density to estimate noise distributions. Simulation results demonstrate the effectiveness of the scheme, especially when nonlinear time-varying complex networks have time-varying sensor delay and noise distributions are unknown.

Design of gain-scheduling cascade control for a motion compensated gripper frame

The offshore wind market is growing, resulting in larger wind turbines being installed farther away from the coast and into deeper waters. This trend brings challenges to an industry strongly depending on the use of jack-up vessels. A floating vessel equipped with a Dynamic Positioning (DP) system and a Motion Compensated Gripper Frame (MCGF) provides a more efficient solution to guide XXL monopoles during the installation phase. According to strict regulations, the offset angle of monopiles shall not be over 0.25°. To ensure this, a robust control method is required to deal with gradual change in soil stifness, wave induced motions and sensor delays. The control force exerted by the gripper frame can significantly affect the floating vessel dynamics, hence, this should be optimized.Therefore, the research goal is to analyse the system and design a robustly stable control for the MCGF. In the context of this research, cascade PID control with gain-scheduling is chosen, which is able to control the Gripper Frame during all installation steps for different soil stiffness and to reject wave disturbances.With the use of the lumped multiplicative uncertainty structure, the changing and uncertain soil-characteristics are modelled. The sensor delay to measure the monopile inclination, has a significant impact on system stability. Robust stability criteria are validated using the Nyquist criterion and by analysing the system’s closed-loop poles. With the use of gain-scheduling it is possible to switch control settings during the installation. Time-domain results show that with this design approach the installation criteria are met such as the maximum monopile angle and the applied control force.Ultimately, it is shown that robust PID control with gain-scheduling for a MCGF can ensure safe, efficient and cost-effective installation of the next generation wind turbines. Future research should focus into the definition and implementation of the gain-scheduling switching criteria.

Distributed resilient fusion filtering for nonlinear systems with random sensor delay under round-robin protocol

In this paper, we address the distributed resilient fusion filtering (DRFF) problem for a class of nonlinear multi-sensor networked systems (MSNSs) with random sensor delay (RSD) under round-robin protocol (RRP). The RSD is depicted by the Bernoulli random variable with known occurrence probability. In order to relieve the network congestion, the RRP that can deal with information overload issue of the transmission process from sensor to the estimator is utilised. The major objective of this paper is that the resilient fusion filter is designed for nonlinear MSNSs with RSD and RRP in the light of the inverse covariance intersection approach, where the local upper bound regarding the filtering error covariance is obtained and then minimised by suitably exploiting the local filter gain. Finally, a simulation example that can show the validity of the provided DRFF algorithm is presented.

Output regulation for a first-order hyperbolic PIDE with state and sensor delays

Considering disturbances within domain and at the boundary, a backstepping-based output boundary regulator design is developed for a class of first-order linear hyperbolic partial integro-differential equation (PIDE) in the presence of state and sensor delays. The delays are represented by two transport PDEs, which results in an extended spatial domain where the hyperbolic PIDE, transport PDEs and ordinary differential equation (ODE) are in cascade. The ODE is a finite-dimensional signal model describing exogenous signals. First, a state feedback regulator is realized to achieve a finite time stability by applying an affine Volterra integral transformation. Then, an output regulator is developed on the basis of the nominal plant transfer behavior, which results in an exponential stability. Numerical examples illustrate the performance of the proposed regulators.

Distributed Adaptive Tobit Kalman Filter for Networked Systems Under Sensor Delays and Censored Measurements

Distributed Adaptive Tobit Kalman Filter for Networked Systems Under Sensor Delays and Censored Measurements

Sensor delay-compensated prescribed-time observer for LTI systems

We present an observer for linear time-invariant (LTI) systems with measurement delay. Our design ensures that the observer error converges to zero within a prescribed terminal time. To achieve this, we employ time-varying output gains that approach infinity at the terminal time, which can be arbitrarily short but no shorter than the sensor delay time. We model the sensor delay as a transport partial differential equation (PDE) and build upon the cascade ODE–PDE setting while accounting for the infinite dimensionality of the sensor. To construct our time-varying gains, the observer design needs to be conducted in a particular system representation. For this reason, we employ a sequence of state transformations (and their inverses) mapping the original observer error model into (1) the observer form, (2) a sensor delay-compensated observer error form via backstepping, and (3) a particular diagonal form that is amenable to the selection of time-varying gains for prescribed-time stabilization. Our construction of the time-varying observer gains uses (a) generalized Laguerre polynomials, (b) elementary symmetric polynomials, and (c) polynomial-based Vandermonde matrices. A simulation illustrates the results.

Emerging architecture for heterogeneous smart cyber-physical systems for industry 5.0

Industrial process automation is becoming more advance due to the new technological revolutions like Industrial Internet of things (IIoT), Machine to Machine Communication (M2M), cloud computing, cognitive computing, artificial intelligence, and big little multicore-based ARM processors for embedded applications. Modern IP-enabled sensors, actuators, and controllers are transforming industrial automation into industry 5.0 compliances to attain full autonomy with minimal human intervention. With these emerging technologies, Smart Cyber-Physical System (SCPS) is the most important part of the fourth industrial revolution, where diffident programmed embedded systems are networked together to perform, computation, communication, control, and actuation. In this paper, we proposed a heterogeneous architecture for SCPS, where different electrical, pneumatic, and hydraulic processes can be integrated to execute hybrid process dynamics. The proposed architecture is enabled to separate all aspects like computation, control, communication, and actuation, of a process dynamic by estimating the process disturbances, sensor delay, actuator delay, and conversion delay. The allotment of computational embedded cores to different physical processes is done through voltage frequency islands (VFI) with high modularity which is different for different processes. All the mapped process dynamics are optimized through Dynamic Voltage and Frequency Scaling (DVFS). The best implementation of the proposed architecture is the upcoming era of industry 5.0 where human intervention is also in a fold in various industries like petroleum, fertilizer, paper, cement, space exploration, and automobile manufacturing.

SAW gas sensor based on extremely thin strain-engineered K0.7Na0.3NbO3 films

In this work, we demonstrate that extremely thin strain-engineered K0.7Na0.3NbO3 (KNN) films are ideal candidates for highly sensitive and also potentially selective surface acoustic wave (SAW) sensor applications. The strength of the use of these films in SAW sensors is based on their piezoelectric properties and their thinness. The latter leads to a strong concentration of the SAW energy at the very surface of the sensor's delay line and the generation of higher harmonics with significant amplitudes. Thin epitaxial films of typically 30 nm in thickness are grown via liquid-delivery spin metal-organic vapor phase epitaxy on different (110)-oriented scandate substrates (TbScO3 and GdScO3). The epitaxial strain is induced by the lattice mismatch between a substrate and a film. The SAW signal of thin KNN films and the resulting sensitivity of an SAW thin KNN film sensor are compared with conventional bulk SAW sensors based on LiNbO3 (LN) using identical electrode designs for the generation and detection of the SAW for both systems. Compared to the conventional LN SAW sensor, our KNN-based sensor shows a sensitivity that is approximately 14 times higher. This was achieved using only the third and fifth harmonics. Using even higher harmonics, the improvement could potentially be boosted up to a factor > 40. Moreover, we showed that simultaneous sensor recording of mass loading at different harmonics is possible with the KNN sensor. Similar to other sensor concepts, the resulting multiple signals might provide a fingerprint of the detected material and, thus, lead to a selective detection of the mass load.

\({\mathbb{T}}\)-Proper Hypercomplex Centralized Fusion Estimation for Randomly Multiple Sensor Delays Systems with Correlated Noises.

The centralized fusion estimation problem for discrete-time vectorial tessarine signals in multiple sensor stochastic systems with random one-step delays and correlated noises is analyzed under different -properness conditions. Based on , , linear processing, new centralized fusion filtering, prediction, and fixed-point smoothing algorithms are devised. These algorithms have the advantage of providing optimal estimators with a significant reduction in computational cost compared to that obtained through a real or a widely linear processing approach. Simulation examples illustrate the effectiveness and applicability of the algorithms proposed, in which the superiority of the linear estimators over their counterparts in the quaternion domain is apparent.

Securing Deep Learning Models with Autoencoder based Anomaly Detection

Deep learning models are on the rise in many scientific fields. Their ability to solve complex and nonlinear tasks has made them very popular. However, in comparison to physical models, they struggle with extrapolation. Hence, it is important that the input data in the production stage is similar to the data seen in training. Deviations to the training often occur in real world applications due to sensor delays and drifts, aging of the system, and communication errors such as noise. Especially in safety relevant applications, securing those models against these influences, which can be seen as anomalies, is essential.
 
 In the past, Autoencoders, especially Variational Autoencoders (VAEs), have been proven useful for anomaly detection. Many researches focus on improving the Autoencoder’s separation ability for an optimal set anomaly threshold. However, the setting of the threshold is not trivial and is crucial for a good anomaly detection. Setting the threshold optimal becomes especially challenging if the anomaly is unknown in the training process, which is often the case in a real world application.
 
 The proposed method combines a deep learning model with an Autoencoder. The input data is handed to the trained Autoencoder which reconstructs the input. If the data is similar to the training data, the Autoencoder should be able to reconstruct the input data accurately. Otherwise, an anomaly is suspected. The reconstruction and the original input data are both passed through the deep learning model, generating two predictions, which are then compared.
 
 For classification and reinforcement tasks with discrete result space the prediction of non-anomalous data should lead to the same class for both samples. For those tasks this allows us to sort out samples as anomalies for which the two results are not the same, hence the threshold becomes obsolete.
 
 For regression and continuous reinforcement tasks, the difference between the two predictions can be interpreted as a safety measure and is easier to grasp than the Autoencoder’s reconstruction error, which is typically used.
 
 The advantage of this method is the distinction between samples that can or cannot be handled by the subsequent application model instead of just deciding if the input is anomalous. This leads to a higher robustness of the joint model and a better usage of resources of the deep leaning model. Moreover, the Autoencoder and the deep learning model are trained separately which makes the training a lot more stable than using coupled training methods.
 
 The proposed method is proven on both, a classification and a regression task. For the classification, the publicly available UEA multivariate time series classification dataset was used. For regression, a dataset simulating a SCR catalyst as part of an automotive exhaust gas aftertreatment system was evaluated. For both tasks common anomalies such as delay and noise were applied to the data.

Enhanced robust nanopositioning control for an X-Y piezoelectric stage with sensor delays: An infinite dimensional [formula omitted] optimization approach

In nanopositioning systems with capacitive displacement sensors, the analog-to-digital (A/D) conversion time is not negligible, particularly for the sub-nano resolution cases with 16 (or more) digital bits. The sensor induced delay will complicate modeling and control of such systems, due to the infinite dimensionality of the delay block. In this paper, we analyze the piezoelectric nanopositioner with a Hammerstein-like model incorporating uncertainties and a measurement delay. An infinite dimensional H∞ control approach is proposed for robust and nanopositioning control of the piezoelectric stage. To eliminate the fragility problem of the H∞ controller on its implementation, an equivalent structure including a Finite Impulse Response (FIR) filter is derived, and the matrix-function-based rational approximation (MFRA) for the FIR filter is presented. Real time experiments with the proposed control method are conducted where the positioning performance, robustness, and hysteresis compensation capability are comprehensively evaluated. Comparative studies demonstrate significant improvements over conventional methods such as Proportional–Integral–Derivative (PID) control and the finite dimensional robust control.

Individual cylinder fuel blend estimation in a dual-fuel engine using an in-cylinder pressure based observer

Cylinder-to-cylinder combustion dispersion in internal combustion engines might be caused by various factors (e.g. manufacturing variations of the injectors or nozzle coking) which can result in an increase in pollutant emissions. When dealing with low temperature combustion concepts such as premixed dual-fuel, the system might suffer from an additional source of dispersion due to port fuel injection distribution. Conventional cylinder fuel concentration estimation is based on look-up tables previously calibrated and saved in the ECU. The aging and the fuel distribution characterization of the injectors are a challenging task when relying only on a single UEGO sensor placed at the exhaust. In-cylinder pressure sensors offer a powerful solution to evaluate the energy released by the fuel with one cycle resolution. The present work proposes to combine the information provided by such sensor together with conventional sensors, in this case air mass flow and lambda sensor, for estimating the fuel concentration and blending ratio entering each cylinder in a heavy-duty dual-fuel engine. A Kalman filter was designed to tackle the dynamics of the system, e.g. lambda sensor delay and port fuel distribution, and validated in both conventional diesel and dual-fuel combustion. The output of the filter was then used to update the injectors look-up table in order to cope with aging and possible bias over time.

Partial-nodes-based state estimation for linear complex networks with randomly occurring sensor delay and stochastic coupling strength

This paper is concerned with the partial-nodes-based state estimation (PNBSE) problem for a class of time-varying complex networks with randomly occurring sensor delay (ROSD) and stochastic coupling strength (SCS). We assume that only partial outputs of nodes can be measured and utilized in the estimation algorithm design. The ROSD is expressed by a set of Bernoulli distributed random variables and the occurrence probabilities are certain. Moreover, the SCS is represented by a set of random variables which obeys the uniform distribution. The aim of this paper is to design a recursive state estimator based on the framework of the Kalman filter, where the optimization problem of the upper bound of the estimation error covariance (EEC) is discussed by employing the variance-constrained method. It can be seen that the gain matrix of each node can be obtained by solving two Riccati-like difference equations. In addition, the performance of the PNBSE method is characterized by discussing the relationship between the occurrence probabilities of ROSD and the upper bound of the EEC, where the related mathematical proof is provided. Finally, some simulations are given to show the validity and correctness of the presented PNBSE approach.

Shear-horizontal surface acoustic wave sensor for non-invasive monitoring of dynamic cell spreading and attachment in wound healing assays

A Love-wave based biosensor is introduced for analyzing a standardized wound healing assay by observing cell growth and quantifying cell detachment processes. Utilizing the piezoelectric material LiTaO3 36° XY-cut with a thin SiO2-cover layer, shear horizontal surface acoustic waves (SAW) are excited and detected by a set of Interdigital Transducers. Epithelial cells, being cultivated on the substrate and invading the sensors delay line cause a phase shift in the transmitted SAW signal. This phase shift correlates exactly with the surface coverage of the invading cells. After wound healing, emerging fluctuations in the phase shift signal provide information about the cell growth in a confluent cell layer. Additionally, the signal slope allows to quantify the cell detachment process induced by apoptosis, necrosis or cell lysis substances, respectively. Furthermore, culture conditions like temperature or osmolality can be simultaneously monitored by SAW. Based on a theoretical approach and using FEM simulations, we identified the acoustoelectric interaction as the main reason for the phase shift in various frequency- and time-dependent studies. Our model is validated by experimental data and allows predicting the phase change caused by variations in the cell-substrate distance or the volume ratio of the nucleus and the complete cell.