Kalman Filter Research Articles

Speed-sensorless model predictive thrust control of linear induction motor based on improved extended Kalman filter observer

This paper puts forth a strategy for sensorless finite-set model predictive thrust control (FS-MPTC) of linear induction motors (LIM) based on an improved extended Kalman filter (IEKF) observer. It is designed to address the shortcomings of existing LIM sensorless vector control methodologies, namely their lack of robustness and ineffective decoupling due to dynamic end effects. First, a fifth-order nonlinear discrete model of LIM is established in order to meet the requirements of sensorless FS-MPTC. An extended Kalman filter (EKF) observer based on the state-space equations is derived with the objective of achieving synchronous online observation of current, flux linkage and speed. In order to address the limitations of the traditional EKF, which employs a fixed system noise covariance matrix, this paper introduces an adaptive adjustment mechanism based on the difference between the theoretical predictions and actual measurements of the observer. This mechanism allows for the covariance matrix Q to be adjusted in real-time. The enhanced IEKF observer is then put forth. Simulation and experimental outcomes illustrate that, in comparison to the conventional EKF, the IEKF is capable of effectively adapting to internal noise fluctuations resulting from disparate operational conditions, system disturbances and LIM dynamic end effects. It is able to accurately estimate speed and flux linkage, operating stably under both high and low-speed conditions with commendable dynamic and robust performance.

LEO-SOP Differential Doppler/INS Tight Integration Method Under Weak Observability

The utilization of low Earth orbit (LEO) satellites’ signals of opportunity (SOPs) for absolute positioning and navigation in global navigation satellite system (GNSS)-denied environments has emerged as a significant area of research. Among various methodologies, tightly integrated Doppler/inertial navigation system (INS) frameworks present a promising solution for achieving real-time LEO-SOP-based positioning in dynamic scenarios. However, existing integration schemes generally overlook the key characteristics of LEO opportunity signals, including the limited number of visible satellites and the random nature of signal broadcasts. These factors exacerbate the weak observability inherent in LEO-SoOP Doppler/INS positioning, resulting in difficulty in obtaining reliable solutions and degraded positioning accuracy. To address these issues, this paper proposes a novel LEO-SOP Doppler/INS tight integration method that incorporates trending information to alleviate the problem of weak observability. The method leverages a parallel filtering structure combining extended Kalman filter (EKF) and Rauch–Tung–Striebel (RTS) smoothing, extracting trend information from the quasi-real-time high-precision RTS filtering results to optimize the EKF positioning solution for the current epoch. This approach effectively avoids the overfitting problem commonly associated with directly using batch data to estimate the current epoch state. The experimental results validate the improved positioning accuracy and robustness of the proposed method.

ВИКОРИСТАННЯ ФІЛЬТРУ КАЛМАНА У ВЕКТОРНІЙ СИСТЕМІ ЕКСТРЕМАЛЬНОГО КЕРУВАННЯ АСИНХРОННОЮ МАШИНОЮ

In the article the reactive power of an asynchronous machine (AM) in a stable mode is obtained as a function of three vari-ables: the angular speed of rotation of the AM shaft, the module of the rotor flux coupling and the moment of static load on the shaft. The AM power factor decreases in electric drives (ED) of industrial mechanisms in which the load moment while maintaining its direction of action can take a static value less than the nominal value. Therefore, in order to ensure a high power factor of an asynchronous ED in long-term operating modes with a variable load, a vector AM control system was created, in which the magnitude of the rotor flux coupling is regulated by the reactive power channel as a function of the mo-ment of static load on the shaft. To identify the moment of loading, as well as the angular speed of the rotor and its flux cou-pling, a Kalman observer is synthesized, which allows creating a vector control system with simultaneous regulation of the speed and power factor of the AM without physical sensors of the reference vector of the rotor flux coupling and its angular speed of rotation. The significant dependence of the extreme values of the rotor flux coupling for reactive power on the torque on the AM shaft and insignificant dependence on the speed has been proved. Stability and high control quality are achieved by simultaneous use of relay control laws, invariant to coordinate and parametric perturbations, and the effectiveness of the Kalman filter algorithm for identifying variable states in the feedback loop. Thus, the article theoretically substantiates the idea of creating a sensorless relay-vector control system for an asynchronous ED with simultaneous speed regulation and optimization of AM energy indicators. The mathematical model is created as a program written in the Matlab programming language. The operability of the proposed asynchronous ED control system was confirmed by the method of mathematical modeling. References 12, figures 6.

GPS Phase Integer Ambiguity Resolution Based on Eliminating Coordinate Parameters and Ant Colony Algorithm.

Correctly fixing the integer ambiguity of GNSS is the key to realizing the application of GNSS high-precision positioning. When solving the float solution of ambiguity based on the double-difference model epoch by epoch, the common method for resolving the integer ambiguity needs to solve the coordinate parameter information, due to the influence of limited GNSS phase data observations. This type of method will lead to an increase in the ill-posedness of the double-difference solution equation, so that the fixed success rate of the integer ambiguity is not high. Therefore, a new integer ambiguity resolution method based on eliminating coordinate parameters and ant colony algorithm is proposed in this paper. The method eliminates the coordinate parameters in the observation equation using QR decomposition transformation, and only estimates the ambiguity parameters using the Kalman filter. On the basis that the Kalman filter will obtain the float solution of ambiguity, the decorrelation processing is carried out based on continuous Cholesky decomposition, and the optimal solution of integer ambiguity is searched using the ant colony algorithm. Two sets of static and dynamic GPS experimental data are used to verify the method and compared with conventional least squares and LAMBDA methods. The results show that the new method has good decorrelation effect, which can correctly and effectively realize the integer ambiguity resolution.

System identification and fault reconstruction in solar plants via extended Kalman filter-based training of recurrent neural networks.

This article proposes using the extended Kalman filter (EKF) for recurrent neural network (RNN) training and fault estimation within a parabolic-trough solar plant. The initial step involves employing an RNN to model the system. Given the challenge of fault discernibility in the collectors, parallel EKFs are employed to reconstruct the parameters of the faults. The parameters are used independently to estimate the system output, and the type of fault is isolated based on the estimation errors using another feedforward neural network. To evaluate the effectiveness of the methodology, simulations are conducted on a loop of the ACUREX plant with irradiances from sunny and cloudy days. The results reveal a fault classification accuracy of approximately 90% and a fault reconstruction error below 3%, with even better accuracies in the cloudy dataset than in the sunny dataset.

KalmanFormer: using transformer to model the Kalman Gain in Kalman Filters.

Tracking the hidden states of dynamic systems is a fundamental task in signal processing. Recursive Kalman Filters (KF) are widely regarded as an efficient solution for linear and Gaussian systems, offering low computational complexity. However, real-world applications often involve non-linear dynamics, making it challenging for traditional Kalman Filters to achieve accurate state estimation. Additionally, the accurate modeling of system dynamics and noise in practical scenarios is often difficult. To address these limitations, we propose the KalmanFormer, a hybrid model-driven and data-driven state estimator. By leveraging data, the KalmanFormer promotes the performance of state estimation under non-linear conditions and partial information scenarios. The proposed KalmanFormer integrates classical Kalman Filter with a Transformer framework. Specifically, it utilizes the Transformer to learn the Kalman Gain directly from data without requiring prior knowledge of noise parameters. The learned Kalman Gain is then incorporated into the standard Kalman Filter workflow, enabling the system to better handle non-linearities and model mismatches. The hybrid approach combines the strengths of data-driven learning and model-driven methodologies to achieve robust state estimation. To evaluate the effectiveness of KalmanFormer, we conducted numerical experiments in both synthetic and real-world dataset. The results demonstrate that KalmanFormer outperforms the classical Extended Kalman Filter (EKF) in the same settings. It achieves superior accuracy in tracking hidden states, demonstrating resilience to non-linearities and imprecise system models.

Validation and Analysis of Recreational Runners' Kinematics Obtained from a Sacral IMU.

Our aim was to validate a sacral-mounted inertial measurement unit (IMU) for reconstructing running kinematics and comparing movement patterns within and between runners. IMU data were processed using Kalman and complementary filters separately. RMSE and Bland-Altman analysis assessed the validity of each filtering method against a motion capture system. Running data from 24 recreational runners were analyzed using Fourier transform coefficients, PCA, and k-means clustering. High agreement was found for Kalman-filtered data in the frontal, sagittal, and transverse planes, with a Bland-Altman bias of ~2 mm on average, compared to a bias of ~10.5 mm for complementary-filtered data. Pelvic angles calculated from Kalman-filtered data had superior agreement, with systematic biases of ~0.3 versus 3.4 degrees for complementary-filtered data. Our findings suggest that inertial sensors are viable alternatives to motion capture for reconstructing pelvic running kinematics and movement patterns. In the second part of our study, negligible intra-individual differences were observed with changes in speed, while inter-individual differences were large. Two clusters of runners were identified, each showing distinct movement patterns and ranges of motion. These observations highlight the potential usefulness of inertial sensors for performance analysis and rehabilitation as they may permit the use of individual-specific and cluster-specific practice programs.

Experimental Validation of Virtual Torque Sensing for Wind Turbine Gearboxes Based on Strain Measurements

ABSTRACTIn efforts to reduce the operation and maintenance cost of wind turbines, there is an increasing interest to monitor key turbine quantities such as the torque load on the gearbox. Monitoring the torque paves the way for the calculation of remaining useful lifetime, leading to cost reductions through improved reliability and maintenance planning. In order to avoid expensive direct torque sensors, this paper investigates the potential of virtual torque sensing, a technique based on 3 basic components: first, a set of non‐intrusive sensors installed on the gearbox. Three groups of strain gauges on the gearbox as well an angular encoder are considered in this paper. Second is a physics‐based model, capable of predicting the response of aforementioned sensors. These models are constructed with a purposeful balance between accuracy and computational cost. Model validation and updating are performed to ensure efficient and accurate prediction of the sensor output. Finally, an augmented extended Kalman filter (AEKF) is used to combine the measured response with predictions from the model to infer the gearbox input torque. Since a key factor determining the performance of the AEKF is the tuning of the AEKF covariance matrices, multiple methods are introduced to systematically tune the covariance matrices. Experimental validation results show that the virtual torque sensor can detect the load torque with a normalized mean absolute error (NMAE) between 3.41% and 7.47%, depending on the sensor set. The influence of the amount of sensors used and the tuning method are also investigated.

A comparative analysis of perceptual noise in lateral and depth motion: Evidence from eye tracking.

The characterization of how precisely we perceive visual speed has traditionally relied on psychophysical judgments in discrimination tasks. Such tasks are often considered laborious and susceptible to biases, particularly without the involvement of highly trained participants. Additionally, thresholds for motion-in-depth perception are frequently reported as higher compared to lateral motion, a discrepancy that contrasts with everyday visuomotor tasks. In this research, we rely on a smooth pursuit model, based on a Kalman filter, to quantify speed observational uncertainties. This model allows us to distinguish between additive and multiplicative noise across three conditions of motion dynamics within a virtual reality setting: random walk, linear motion, and nonlinear motion, incorporating both lateral and depth motion components. We aim to assess tracking performance and perceptual uncertainties for lateral versus motion-in-depth. In alignment with prior research, our results indicate diminished performance for depth motion in the random walk condition, characterized by unpredictable positioning. However, when velocity information is available and facilitates predictions of future positions, perceptual uncertainties become more consistent between lateral and in-depth motion. This consistency is particularly noticeable within ranges where retinal speeds overlap between these two dimensions. Significantly, additive noise emerges as the primary source of uncertainty, largely exceeding multiplicative noise. This predominance of additive noise is consistent with computational accounts of visual motion. Our study challenges earlier beliefs of marked differences in processing lateral versus in-depth motions, suggesting similar levels of perceptual uncertainty and underscoring the significant role of additive noise.

Corrigendum to ‘Pressure drop prediction for R407C fluid during flow evaporation in horizontal pipes using Kalman filter’ [International Journal of Refrigeration 168 (2024) 149–158

Corrigendum to ‘Pressure drop prediction for R407C fluid during flow evaporation in horizontal pipes using Kalman filter’ [International Journal of Refrigeration 168 (2024) 149–158

Noninteracting optimal and adaptive torque control using an online parameter estimation with help of polynomials in EKF for a PMSM.

This paper addresses a non-interacting torque control strategy to decouple the d- and q-axis dynamics of a permanent magnet synchronous machine (PMSM). The maximum torque per ampere (MTPA) method is used to determine the reference currents for the desired torque. To realize the noninteracting control, knowledge concerning the inductances Ld and Lq of the electrical machine is necessary. These two inductances are estimated by two extended Kalman filters (EKFs), which use a univariate polynomial as a model to describe the saturation effects of the PMSM. The Kalman filters (KF) are realized within a noninteracting control system to improve the observability of the inductance. Despite the non-perfect decoupling, thanks to the structural stochastic nature of the KFs, noninteracting cancellation errors are represented with its process noise and the inductances are estimated sufficiently well. In this sense, we can speak about KFs for and within noninteracting control. Estimating inductances is fundamental for optimal torque control, which is a viable approach to reducing mechanical vibration and disturbance. Moreover, the control strategy of model-based techniques must be adaptively tuned to work properly. Starting from the existing literature, a viable control structure is proposed in which the stability of the control loop using a Proportional Integral (PI) controller is shown for the resulting time-varying system. In fact, the model is represented as a time-varying system because of the presence of the variable inductances Ld and Lq and because of the presence of the velocity of the rotor which is not considered as a state. In this paper, a forward Euler discretization is used to realize the observer in the discrete experimental setup. Measures realized with hardware in the loop (HIL) show interesting results in the context of inductance estimation, due to the advantage of the reduction of the dimensions of the two decoupled EKFs resulting from the noninteraction control. Using the HIL simulator, the proposed torque control strategy is investigated, showing promising results in terms of increasing observability due to decoupling. This and the usage of univariate polynomials in EKF calculations lead to significant reduction of measurement points, reduction of oscillations and ripples, deviation between desired and achieved torque and reduction of disturbances. Moreover, the proposed control strategy using a very limited calculation load, at the same time, maintains the ripples inside the technical limits of the obtained torque. Both effects are due to the decoupled EKFs with simplified and reduced order of the models using univariate polynomials, which require significantly fewer measuring points in the run-up to the creation of the model of the inductances Ld and Lq.

Modelling COVID-19 in the North American region with a metapopulation network and Kalman filter

Modelling COVID-19 in the North American region with a metapopulation network and Kalman filter

Heterogeneous sensor-based rotor attitude detection for a spherical motor using distributed Kalman filtering method

Permanent magnet spherical motors (PMSpMs) are new special motors with multiple degrees of freedom capability in a limited space. This paper proposes a heterogeneous sensor-based rotor attitude detection (RAD) approach for the PMSpM using the distributed Kalman filtering method. An industrial camera is used to detect the rotor attitude by integrating the Kalman filter (visual method), in order to improve short-time occlusion ability. Meanwhile, a MEMS sensor is introduced to detect the PMSpM rotor attitude simultaneously by using the Kalman filtering method (MEMS method). On the basis of this, the heterogeneous sensor network consists of the camera and the MEMS sensor, and the proposed RAD approach is implemented through the Kalman filtering fusing method. To verify the proposed heterogeneous sensor-based RAD method, the test bench is developed and the one-dimensional and three-dimensional fusing experiments are conducted. The results indicate that the heterogeneous sensor-based RAD method possesses the ability of visual resistance to long-term interference with good accuracy.

A Neural-Metaheuristic Kalman Filter for Moving Microburst Wind Shear Identification

Microburst wind shears (MB) are powerful localized three dimensional (3D) columns of wind that occur when cooled air drops from the base of thunderstorms at high speeds. They eventually hit the ground and spread out in all directions. As such MBs are considered a major atmospheric hazard for aircrafts (ACs), especially in terminal flight phases. Though pilots train regularly to survive it, yet the most recommended practice within the aviation community is to avoid its encounter upon detection. Some modern aircrafts have predictive wind shear warning systems, but they have limited short range capability and do not provide quantitative data for engineering application. In this sense, accurate onboard detection and identification of MBs and its characteristics is of importance for flight safety, whose success can lead to design of automatic flight control systems (FCSs) that reduce the risk of aircraft crash landings and accidents. Even though there are some studies on FCS designs for MB encounter, research on MB identification is rare but ongoing. To this aim, the present study is among the firsts that focuses on online identification of multiple and moving MB parameters using onboard aircraft air data and inertial sensors. The identification task is initially performed using the aircraft six degrees of freedom (6DoF) equations of motion (EOM) integrated with a 3D MB model via the utility of nonlinear Kalman filtering (KF) algorithms. Subsequently, an enhanced neural metaheuristic Kalman filter (NMKF) is proposed that improves the estimation accuracy of the MB model parameters. The results demonstrate the efficacy of the proposed NMKF in comparison with previous studies.

Online identification method of power grid load sensitivity based on adaptive Kalman filter

Online identification method of power grid load sensitivity based on adaptive Kalman filter

The use of Kalman filters in human gait analysis using depth sensors

The use of Kalman filters in human gait analysis using depth sensors

Accurate state of charge estimation for UAV-centric lithium-ion batteries using customized unscented Kalman filter

Accurate state of charge estimation for UAV-centric lithium-ion batteries using customized unscented Kalman filter

Robust Face Tracking Under Challenging Conditions Using Linear Regression and YOLO algorithm

Face detection and tracking play a crucial role in various computer vision applications, including surveillance, fault face detection systems, artificial intelligence, etc. The objective of this paper is to enhance the precision of face detection and tracking through the introduction of an innovative approach centered on the linear regression algorithm. The effectiveness of the proposed method was compared to the traditional Kalman filter approach. Additionally, the study explored the integration of the YOLO algorithm for face detection with the linear regression tracking algorithm to further enhance accuracy. The proposed algorithm's performance is assessed through comprehensive experiments on annotated images and video sequences affected by occlusions or other issues such as poor lighting conditions and motion blur. These experiments utilize the COCO dataset, operating at a speed of 60 FPS. The experimental results show that the proposed method can accurately track the human face in different facial positions

A collaborative surface target detection and localization method for an unmanned surface vehicle swarm

A single unmanned surface vehicle (USV) designed for marine missions suffers from limited payload, low efficiency and weak intelligence, while a swarm of USVs shows significant advantages in mission flexibility, diverse payload and task efficiency. One of the key issues for an USV swarm is how to achieve highly efficient collaborative perception. To address this issue, a method framework of collaborative surface target detection and localization based on multiple sensors for a swarm including 4 USVs is designed. First, perception systems are constructed, a joint calibration method for different sensors is proposed, and a lightweight target detection method improved with attention mechanism and lightweight adaptive spatial feature fusion is designed. Second, a specialized fusion method using sensor principles based on an extended Kalman filter (EKF) is proposed for a single USV to obtain a target state model. Third, the obtained target models from different USVs are registered with fuzzy matching and integrated into the complete model in a geographic coordinate system. The proposed method is applied to the collaborative perception system on our developed 4 USV swarm and verified in real marine environment and simulation. Experimental results show that our proposed method framework significantly improves the accuracy, efficiency, and reliability of the target detection and localization. The proposed LAF-YOLOv8-s reduces the model size by 5.1M, while the mean average precision (mAP) reaches 68.7%, which is significantly superior to other methods. The average collaborative localization error is reduced by 2.9m. The dataset is available at https://github.com/maochenyu1/WSLight.

An effective UWB/IMU joint positioning based on improved anti-NLoS adaptive extended Kalman filter

An effective UWB/IMU joint positioning based on improved anti-NLoS adaptive extended Kalman filter