Modified Kalman Filter Algorithm Research Articles

Study on the closed-loop combustion control for a diesel engine by using a dynamic-target online prediction model

In-cylinder pressure-based closed-loop combustion control is a critical technology used to address the issue of cylinder-to-cylinder inhomogeneity throughout the life cycle of a modern diesel engine. Nevertheless, closed-loop combustion control technology (CLCC), which requires a reference or set-point that the control attempts to achieve, causes an increase in the operating maps and the complexity of the calibration effort. In this study, a reference governor-based combustion control (RGCC) is developed in order to reduce the cylinder-to-cylinder combustion dispersions and the calibration effort associated with enabling of the CLCC simultaneously. The RGCC consists of a dynamic-target online prediction model and an indicated mean effective pressure (IMEP) control strategy. The dynamic-target prediction model is designed based on the IMEP of the engine real-time tests; it uses a modified Kalman filter algorithm, which provides the set-point value for the IMEP control strategy. The effectiveness of the proposed RGCC is experimentally validated on a commercial diesel engine test bench under steady-state and transient-state conditions. The experimental results show that the RGCC displayed far better transient tracking performance than the closed-loop IMEP control based on MAP. Furthermore, the controller can effectively reduce cylinder-to-cylinder IMEP variations under steady-state and transient-state conditions.

Smoothed Kalman MRAS observer for speed sensorless indirect vector controlled induction motor drives

This paper proposes a modified Kalman filter algorithm based on fixed lag smoothing which can be used to estimate the rotor speed and flux of three phase induction motor. The behaviour of a typical extended Kalman filter (EKF) algorithm is influenced by the process and measurement error covariance matrices. In EKF, these matrices are chosen by trial and error method and have to be varied according to the varying operating conditions of the motor. In this paper a smoothing-based extended Kalman filter (SKF) algorithm which uses additional data points for estimation is proposed. Since additional data is used, the algorithm is able to give a better estimate, with the same values of process and measurement error covariance matrices. The performance of the proposed algorithm is explored for sensorless indirect vector control application of three-phase induction motor. Finally, in order to validate the superiority of this algorithm, it is compared with the performance of EKF algorithm for various reference speeds in real-time.

Data processing of field measurements of expedition research for mathematical modeling of hydrodynamic processes in the Azov Sea

This paper describes a three-dimensional mathematical hydrodynamic model capable of taking into account the processes of salt and heat transfer in the Azov Sea. The model allows obtaining three-dimensional fields of the vector of water flow rates, pressure, sea water density, salinity and temperature. The model is based on the equations of motion (Navier-Stokes), the continuity equation in the case of variable density, and the equations of heat and salt transport. The boundary and initial conditions are indicated. To approximate the equation of diffusion-convection-reaction in time, we analyzed the schemes with weights. The approximation of the problem of calculating the velocity field of the aquatic environment in terms of spatial variables was carried out on the basis of the balance method taking into account the occupancy ratios of the control areas. The stationary modes of the heat and salt transfer problem were investigated. The initial distribution of the salinity and temperature functions, which have a sufficient degree of smoothness at the points of setting the field values, was calculated using the Laplace equation. Using the interpolation algorithm and by superimposing the boundaries of the region, maps of salinity and temperature of the Sea of Azov were obtained. Based on the monitoring of the water area, three-dimensional mathematical models of the movement of the aquatic environment designed to predict possible scenarios for the development of the Azov Sea ecosystem were constructed in order to avoid the occurrence of anaerobic infection areas and take timely measures for their localization. The full-scale data obtained using different types of measuring instruments was used to develop observation models for prediction of the changes in hydrodynamic processes. A modified Kalman filter algorithm was applied to obtain unbiased minimum-variance state estimation for the dynamic system. A description is given of a software package that allows modeling hydrodynamic processes in shallow water bodies with complex spatial structures of currents, taking into account the transport of salts and heat.

A multifactor transformed diffusion model with applications to VIX and VIX futures

Transformed diffusions (TDs) have become increasingly popular in financial modeling for their model flexibility and tractability. While existing TD models are predominately one-factor models, empirical evidence often prefers models with multiple factors. We propose a novel distribution-driven nonlinear multifactor TD model with latent components. Our model is a transformation of a underlying multivariate Ornstein–Uhlenbeck (MVOU) process, where the transformation function is endogenously specified by a flexible parametric stationary distribution of the observed variable. Computationally efficient exact likelihood inference can be implemented for our model using a modified Kalman filter algorithm and the transformed affine structure also allows us to price derivatives in semi-closed form. We compare the proposed multifactor model with existing TD models for modeling VIX and pricing VIX futures. Our results show that the proposed model outperforms all existing TD models both in the sample and out of the sample consistently across all categories and scenarios of our comparison.

Model-based topography estimation in trolling mode atomic force microscopy

In this study, a novel approach based on a modified Kalman filter algorithm is presented to directly estimate and measure the surface topography of samples by trolling mode atomic force microscopy. Trolling mode atomic force microscopy was introduced as an atomic force microscopy alternative to overcome imaging problems in liquid environments by reducing the liquid-resonator interaction forces. In conventional imaging techniques, the time to reach the steady state periodic motion of the oscillating probe restricts scanning speed. To overcome this limitation, we propose a novel imaging technique for trolling mode atomic force microscopy based on the system dynamics model and using the adaptive fading Kalman filter with forgetting factor. In this approach, the sample height is estimated directly without the need for any closed loop controller. As a result, the scanning speed is improved significantly, and topography is obtained more accurately compared to the conventional imaging method. Moreover, the effects of process noise, scanning speed, and parameter uncertainties on the performance of proposed approach are investigated.

Inverse Analysis of Inconel 718 Laser-Assisted Milling to Achieve Machined Surface Roughness

This manuscript proposes an inverse analysis method for the machined surface roughness in laser-assisted milling on Inconel 718. The method solves the forward problem considering the tool profile and the elastic recovery of machined surface and applies the variance-based recursive method to guide the updating mechanism of process parameters to match the measurements. Subsequently, the inverse analysis identifies four process parameters of feed per tooth, tool tip radius, minimum cutting thickness, and tool tip angle, and finds the optimal solution for target performance, the surface roughness. The measurements are collected under the single beam coaxial laser-assisted milling spindle. The proposed modified Kalman filter algorithm introduces the gain coefficient G when updating the process parameters to improve robustness and accuracy. The inverse analysis is conducted on all measurements, and the average error of target performance is 0.460% when the laser is on and 0.394% when the laser is off. The average difference of process parameters is less than 5%, and the selection process is done in 50 loops within a minute. Therefore, the proposed inverse analysis model is robust, adaptive to different initial guesses and measurements, highly accurate, and saves computation time.

A real-time and robust feature detection method using hierarchical strategy and modified Kalman filter for thick plate seam tracking

Feature detection is an essential and important part in weld seam tracking of automated welding robots. In thick plate seam tracking, a profilometer based on structured light is commonly employed. Features from light stripe of the structured light will be extracted and used as primary information in visual servoing control. The accuracy, robustness, and computational cost are the main aspects of the feature detection. They directly affect the quality of the tracking. In this paper, cross mark created by cross line structured light (CLSL) is taken into account, and handled as the feature. It can be used as a pinpoint for seam tracking. Firstly, the cross mark is hierarchically estimated by coarse-to-fine strategy. In coarse estimation step, the random sample consensus (RANSAC) algorithm is applied to compute the feature position. Subsequently, the mean shift algorithm is used to estimate the precise feature position in the fine estimation step. Finally, the robustness of the detection is improved by the modified Kalman filter algorithm. The experimental results verify that the feature position estimated by the proposed method is robust. Moreover, the coarse-to-fine strategy can reduce a huge computational cost in the detection. And therefore, the detection method is proper for being used in real-time thick plate seam tracking.

A Novel Freeway Traffic Speed Estimation Model with Massive Cellular Signaling Data

With the growing popularity of cell phones, using massive cellular signaling data as probe to track the vehicles movement trajectory and obtain the real-time traffic condition has become one of the most attractive candidate techniques. However, traditional approaches may offer a poor performance in removing noisy data and minimizing deviation of traffic speed in adjacent time intervals. In this paper, a novel approach is proposed to solve these two issues. The authors move noisy data by comparing the cellular signaling data with the trained data set, and adopt a modified Kalman filter algorithm to minimize the deviations. The experiment results show that the accuracy of the approach performs better in comparison to other two traffic speed estimation approaches.

A modified Kalman filter algorithm for fractional system under Lévy noises

This paper considers the filter designing problem for discrete linear fractional order systems with non-Gaussian noises. Taking both the system noise and the measurement noise as Lévy noises instead of Gaussian white noises, through the approximating transformation for Lévy noises, a modified Kalman filter algorithm is proposed for discrete linear fractional order system. The developed results can be used to estimate the system states for discrete linear fractional order system under non-Gaussian Lévy noises with satisfactory accuracy, which are more general than some existing results. Finally, simulation results are provided to verify and illustrate the effectiveness of the obtained results.

Optimal linear estimation with square-based sampling

This paper is concerned with the problem of event-based sampling for the purpose of reducing the number of data transferred in control systems and the corresponding modified Kalman filter. A mathematical description based upon square-based sampling is proposed, in which the sampling occured when the accumulated difference between two adjacent sampling points satisfies the specified event condition. Then, by utilizing the square based sampling method, a modified Kalman filter algorithm is adopted, in which the measurement is sent when the output exceeds a given event. The simulation results illustrate the advantage of the new event-based sampling scheme and the effectiveness of modified Kalman filter.

Double-filter model with modified Kalman filter for baseband signal pre-processing with application to ultra-tight GPS/INS integration

In federated design of ultra-tight GPS/INS integrated system, the baseband signal pre-processing is completed in a single pre-filter assigned for each channel. As the state space model of this single pre-filter includes the code tracking errors coupled with carrier tracking errors, ionospheric errors and normalized signal amplitude, the carrier tracking process may be destroyed. Also, the measurement noises are not independent any longer after passing through the code and carrier discriminators. Therefore, we propose a double-filter-based pre-filter model that distributes the carrier and code tracking into two independent filters: a conventional pre-filter, where the normalized signal amplitude is excluded from the state space and tracks only the code signal, and a 3-dimension state filter, tracking the carrier signal. The measurement information from both filters is a scalar quantity, which removes most of the noise correlation. To further improve the performance of the double-filter-based pre-filter model, we propose a modified Kalman filter algorithm. Simulation and field tests have been conducted, and the performance analysis has been done for the following configurations in a vector-tracking mode: double-filter model with modified Kalman filter, double-filter model with conventional Kalman filter and traditional single-filter model. The preliminary analysis indicates that the double-filter model with modified Kalman filter shows the best performance in tracking and navigation domains, while the traditional single-filter model shows a sub-optimal performance.

Modified Kalman filter for networked monitoring systems employing a send-on-delta method

In this paper, we consider a networked estimation problem in which sensor data are transmitted only if their values change more than the specified value. When this send-on-delta method is used, no sensor data transmission implies that the sensor value does not change more than the specified value from the previously transmitted sensor value. Using this implicit information, we propose a modified Kalman filter algorithm. The proposed filter reduces sensor data traffic with relatively small estimation performance degradation. Through experiments, we demonstrate the feasibility of the proposed filter algorithm.

Use of a modified kalman filter for a visually coupled system application

Spatial tracking devices are frequently used in virtual environments, such as in the case of helmet mounted displays, to dynamically determine the user's viewpoint and line of sight. Temporal distortion effects are perceived by the user as a result of the lag between head movement and visual feedback. Measurements of phase lag have been made and to help alleviate these problems, predictive filtering techniques are frequently used. We report on studies that have been made in the use of a modified Kalman filter algorithm. The implementation provides favourable results in terms of reduction on the effect of phase lag.

State and Coefficients Estimation for the Activated Sludge Process Using a Modified Kalman Filter Algorithm

An advanced identification algorithm is proposed for the on-line estimation of the non-measurable state variables and coefficients of the IAWPRC biodegradation dynamic model for the Activated Sludge process. The new algorithm employs real-time measurable data. This paper represents a first step in a mathematical methodology which intends to improve the present techniques used for model calibration and validation. The algorithm is based on a recursive non-linear extended Kaiman filter and it has been verifiedrby using simulated values of the measurable data. These simulations describe the behaviour of a specific physical configuration of the AS process under steady and transient operational conditions. When the coefficients are assumed to be known, the algorithm estimates the state rapidly and accurately even when the measurements are affected by a very high degree of noise and the initial estimates are very far from the real values. The speed of convergence of the algorithm decreases when neither the states nor the coefficients of the model are known.

A structured state space approach to computing the likelihood of an ARIMA process and its derivatives

Kohn and Ansley (1984a) define a likelihood for a general nonstationary Gaussian ARIMA process having missing observations anywhere in the series, and obtain a modified Kalman filter algorithm to compute the likelihood.This paper shows how to efficiently implement the modified Kalman filter and its derivatives so as to compute the Kohn and Ansley (1984a) likelihood and its first two derivatives.We do so by demonstrating that most of the computation is done by a set of basic algorithms.