scholarly journals An Iterative Nonlinear Filter Using Variational Bayesian Optimization

Sensors ◽  
2018 ◽  
Vol 18 (12) ◽  
pp. 4222
Author(s):  
Yumei Hu ◽  
Xuezhi Wang ◽  
Hua Lan ◽  
Zengfu Wang ◽  
Bill Moran ◽  
...  
Keyword(s):  
Lower Bound ◽  
Nonlinear Filter ◽  
Nonlinear Filters ◽  
Bayesian Optimization ◽  
System State ◽  
Step Size ◽  
Variational Bayesian ◽  
Penalty Factor ◽  
Leibler Divergence ◽  
Bound Optimization

We propose an iterative nonlinear estimator based on the technique of variational Bayesian optimization. The posterior distribution of the underlying system state is approximated by a solvable variational distribution approached iteratively using evidence lower bound optimization subject to a minimal weighted Kullback-Leibler divergence, where a penalty factor is considered to adjust the step size of the iteration. Based on linearization, the iterative nonlinear filter is derived in a closed-form. The performance of the proposed algorithm is compared with several nonlinear filters in the literature using simulated target tracking examples.

Towards features updating selection based on the covariance matrix of the SLAM system state

Robotica ◽  
2010 ◽  
Vol 29 (2) ◽  
pp. 271-282 ◽  
Author(s):  
Fernando A. Auat Cheein ◽  
Fernando di Sciascio ◽  
Gustavo Scaglia ◽  
Ricardo Carelli
Keyword(s):  
Mobile Robot ◽  
Selection Criterion ◽  
Selection Procedure ◽  
Mobile Robot Navigation ◽  
Monte Carlo Experiment ◽  
Features Selection ◽  
System State ◽  
Localization And Mapping ◽  
Leibler Divergence ◽  
Slam Algorithm

SUMMARYThis paper addresses the problem of a features selection criterion for a simultaneous localization and mapping (SLAM) algorithm implemented on a mobile robot. This SLAM algorithm is a sequential extended Kalman filter (EKF) implementation that extracts corners and lines from the environment. The selection procedure is made according to the convergence theorem of the EKF-based SLAM. Thus, only those features that contribute the most to the decreasing of the uncertainty ellipsoid volume of the SLAM system state will be chosen for the correction stage of the algorithm. The proposed features selection procedure restricts the number of features to be updated during the SLAM process, thus allowing real time implementations with non-reactive mobile robot navigation controllers. In addition, a Monte Carlo experiment is carried out in order to show the map reconstruction precision according to the Kullback–Leibler divergence curves. Consistency analysis of the proposed SLAM algorithm and experimental results in real environments are also shown in this work.

Fracture-Directed Steerable Needles

2019 ◽  
Vol 04 (01) ◽  
pp. 1842002 ◽  
Author(s):  
Fan Yang ◽  
Mahdieh Babaiasl ◽  
John P. Swensen
Keyword(s):  
Lower Bound ◽  
Soft Tissues ◽  
Target Location ◽  
Needle Insertion ◽  
Small Radius ◽  
Radius Of Curvature ◽  
Step Size ◽  
Tissue Phantoms ◽  
Kinematic Models ◽  
Steerable Needles

Steerable needles hold the promise of improving the accuracy of both therapies and biopsies as they are able to steer to a target location around obstructions, correct for disturbances, and account for movement of internal organs. However, their ability to make late-insertion corrections has always been limited by the lower bound on the attainable radius of curvature. This paper presents a new class of steerable needle insertion where the objective is to first control the direction of tissue fracture with an inner stylet and later follow with the hollow needle. This method is shown to be able to achieve radius of curvature as low as 6.9[Formula: see text]mm across a range of tissue stiffnesses and the radius of curvature is controllable from the lower bound up to a near infinite radius of curvature based on the stylet/needle step size. The approach of “fracture-directed” steerable needles indicates the promise of the technique for providing a tissue-agnostic method of achieving high steerability that can account for variability in tissues during a typical procedure and achieve radii of curvature unattainable through current bevel-tipped techniques. A variety of inner stylet geometries are investigated using tissue phantoms with multiple stiffnesses and discrete-step kinematic models of motion are derived heuristically from the experiments. The key finding presented is that it is the geometry of the stylet and the tuning of the bending stiffnesses of both the stylet and the tube, relative to the stiffness of the tissue, that allow for such small radius of curvature even in very soft tissues.

scholarly journals Novel SINS Initial Alignment Method under Large Misalignment Angles and Uncertain Noise Based on Nonlinear Filter

2017 ◽  
Vol 2017 ◽  
pp. 1-14 ◽  
Author(s):  
Bo Yang ◽  
Xiaosu Xu ◽  
Tao Zhang ◽  
Jin Sun ◽  
Xinyu Liu
Keyword(s):  
Kalman Filter ◽  
Nonlinear Filter ◽  
Nonlinear Filters ◽  
Alignment Method ◽  
Initial Alignment ◽  
Numerical Accuracy ◽  
Alignment Problem ◽  
Sampling Problem ◽  
Simulation Results ◽  
Better Than

For the SINS initial alignment problem under large misalignment angles and uncertain noise, two novel nonlinear filters, referred to as transformed unscented quadrature Kalman filter (TUQKF) and robust transformed unscented quadrature Kalman filter (RTUQKF), are proposed in this paper, respectively. The TUQKF sets new deterministic sigma points to address the nonlocal sampling problem and improve the numerical accuracy. The RTUQKF is the combination ofH∞technique and TUQKF. It improves the accuracy and robustness of state estimation. Simulation results indicate that TUQKF performs better than traditional filters when misalignment angles are large. Turntable and vehicle experiments results indicate that, under the condition of uncertain noise, the performances of RTUQKF are better than other filters and more robust. These two methods can effectively further increase precision and convergence speed of SINS initial alignment.

Approximate Nonlinear Filters and Deterministic Filter Gains

1972 ◽  
Vol 94 (1) ◽  
pp. 57-63 ◽  
Author(s):  
A. K. Bejczy ◽  
R. Sridhar
Keyword(s):  
Performance Evaluation ◽  
Differential Equations ◽  
Evaluation Method ◽  
Nonlinear Filter ◽  
Nonlinear Filters ◽  
New Method ◽  
Performance Evaluation Method ◽  
And Performance

A simple nonlinear filter construction and performance evaluation method is described and illustrated on several examples by comparing it to more complex nonlinear filter schemes. In the new method, the filter gain is a precomputed, deterministic quantity (possibly a constant) and, the filter’s performance is (approximately) described by deterministic differential equations which can be solved off-line.

Cramer–Rao lower bound optimization of an EM-CCD-based scintillation gamma camera

2013 ◽  
Vol 58 (8) ◽  
pp. 2641-2655 ◽  
Author(s):  
Marc A N Korevaar ◽  
Marlies C Goorden ◽  
Freek J Beekman
Keyword(s):  
Lower Bound ◽  
Gamma Camera ◽  
Scintillation Gamma ◽  
Cramer Rao Lower Bound ◽  
Bound Optimization
Sign in / Sign up

Export Citation Format

Share Document