scholarly journals Multi-Sensor Fusion for Underwater Vehicle Localization by Augmentation of RBF Neural Network and Error-State Kalman Filter

Sensors ◽  
2021 ◽  
Vol 21 (4) ◽  
pp. 1149
Author(s):  
Nabil Shaukat ◽  
Ahmed Ali ◽  
Muhammad Javed Iqbal ◽  
Muhammad Moinuddin ◽  
Pablo Otero
Keyword(s):  
Neural Network ◽  
Kalman Filter ◽  
Sensor Fusion ◽  
Rbf Neural Network ◽  
Underwater Vehicle ◽  
Estimation Accuracy ◽  
Optimization Approach ◽  
Vehicle Localization ◽  
High Nonlinearity ◽  
Error State

The Kalman filter variants extended Kalman filter (EKF) and error-state Kalman filter (ESKF) are widely used in underwater multi-sensor fusion applications for localization and navigation. Since these filters are designed by employing first-order Taylor series approximation in the error covariance matrix, they result in a decrease in estimation accuracy under high nonlinearity. In order to address this problem, we proposed a novel multi-sensor fusion algorithm for underwater vehicle localization that improves state estimation by augmentation of the radial basis function (RBF) neural network with ESKF. In the proposed algorithm, the RBF neural network is utilized to compensate the lack of ESKF performance by improving the innovation error term. The weights and centers of the RBF neural network are designed by minimizing the estimation mean square error (MSE) using the steepest descent optimization approach. To test the performance, the proposed RBF-augmented ESKF multi-sensor fusion was compared with the conventional ESKF under three different realistic scenarios using Monte Carlo simulations. We found that our proposed method provides better navigation and localization results despite high nonlinearity, modeling uncertainty, and external disturbances.

Kalman Filter Based Sensor Fusion for a Mobile Manipulator

2019 ◽  
Author(s):  
Barnaba Ubezio ◽  
Shashank Sharma ◽  
Guglielmo Van der Meer ◽  
Michele Taragna
Keyword(s):  
Kalman Filter ◽  
Sensor Fusion ◽  
Inertial Sensor ◽  
Noise Model ◽  
Estimation Accuracy ◽  
Mobile Manipulator ◽  
End Effector ◽  
Kalman Filter Algorithm ◽  
Working Frequency ◽  
Mems Imu

Abstract End-effector tracking for a mobile manipulator is achieved through Sensor Fusion techniques, implemented with a particular visual-inertial sensor suite and an Extended Kalman Filter algorithm. The suite is composed of an Optitrack motion capture system and a Honeywell HG4930 MEMS IMU, for which a further analysis on the mathematical noise model is reported. The filter is constructed in such a way that its complexity remains constant and independent of the visual algorithm, with the possibility of inserting additional sensors, to further improve the estimation accuracy. Experiments in real-time have been performed with the 12-DOF KUKA VALERI robot, extracting the position and the orientation of the end-effector and comparing their estimates with pure sensor measurements. Along with the physical results, issues related to calibration, working frequency and physical mounting are described.

A hybrid method for lithium-ion batteries state-of-charge estimation based on gated recurrent unit neural network and an adaptive unscented Kalman filter

2021 ◽  
pp. 1-36
Author(s):  
Shuai Xu ◽  
Fei Zhou ◽  
Yucheng Liu
Keyword(s):  
Neural Network ◽  
Kalman Filter ◽  
Hybrid Method ◽  
Unscented Kalman Filter ◽  
Absolute Error ◽  
State Of Charge ◽  
Estimation Accuracy ◽  
Adaptive Unscented Kalman Filter ◽  
Battery State Of Charge ◽  
Gated Recurrent Unit

Abstract Among the battery state of charge estimation methods, the Kalman-based filter algorithms are sensitive to the battery model while the neural network-based algorithms are decided by hyperparameters. In this paper, a hybrid approach composed of a gated recurrent unit neural network and an adaptive unscented Kalman filter method is proposed. A gated recurrent unit neural network is first used to acquire the nonlinear relationship between the battery state of charge and battery measurement signals, and then an adaptive unscented Kalman filter is utilized to filter out the output noise of the neural network to further improve estimation accuracy. The hybrid method avoids the establishment of accurate battery models and the search for optimal hyperparameters. The data of dynamical street test and US06 test are used as training dataset and validation dataset, respectively, while the data collected from the tests under federal urban driving schedules and Beijing driving cycle conditions are taken as testing dataset. As compared with some hybrid methods proposed in other literature, the hybrid method has the best estimation accuracy and generalization for various driving cycles at different ambient temperatures. The root mean square error and the mean absolute error all are less than 1.5%, and the maximum absolute error are less than 2%. In addition, it also exhibits powerful robustness against the abnormal values of the battery signals and can converge to the true value in just 5 seconds.

Intelligent Kalman Filtering for Fault Detection on an Active Magnetic Bearing System

2008 ◽  
Author(s):  
Nana K. Noel ◽  
Kari Tammi ◽  
Gregory D. Buckner ◽  
Nathan S. Gibson
Keyword(s):  
Neural Network ◽  
Kalman Filter ◽  
Fault Detection ◽  
Magnetic Bearing ◽  
Active Magnetic Bearing ◽  
Estimation Accuracy ◽  
False Alarms ◽  
Single Input Single Output ◽  
Single Output ◽  
Modeling Uncertainties

One of the challenges of condition monitoring and fault detection is to develop techniques that are sufficiently sensitive to faults without triggering false alarms. In this paper we develop and experimentally demonstrate an intelligent approach for detecting faults in a single-input, single-output active magnetic bearing. This technique uses an augmented linear model of the plant dynamics together with a Kalman filter to estimate fault states. A neural network is introduced to enhance the estimation accuracy and eliminate false alarms. This approach is validated experimentally for two types of fabricated faults: changes in suspended mass and coil resistance. The Kalman filter alone is shown to be incapable of identifying all fault cases due to modeling uncertainties. When an artificial neural network is trained to compensate for these uncertainties, however, all fault conditions are identified uniquely.

Sign in / Sign up

Export Citation Format

Share Document