Kalman Filtering and Neural Networks

Haykin, S.

doi:10.1002/0471221546

Cited by 1,206 publications

(31 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…A “filter” is, therefore, a transformation (linear or non-linear) of an n-dimensional input signal to an m-dimensional output signal. With this terminology, low-pass, band-pass, or high-pass filters are included in the definition of “filters”, as well as the short-time Fourier transform (Allen, 1977) or classification and regression methods, such as linear discriminant analysis (Hastie et al, 2011), support vector regression (Vapnik and Chervonenkis, 1974) and Kalman filter (Kalman, 1960; Haykin, 2001). …”

Section: Methodsmentioning

confidence: 99%

“…For classification, we tested filter pipelines implementing linear discriminant analysis (LDA) (Hastie et al, 2011) and support vector machines (SVM) (Vapnik and Chervonenkis, 1974). For regression we tested the linear filter (LF), support vector regression (SVR) and the Kalman filter (KF) (Kalman, 1960; Haykin, 2001). We used libsvm library (Chang and Lin, 2011) to implement SVM and SVR, and the GNU Scientific Library (Gough, 2009) for linear algebra operations.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Low-latency multi-threaded processing of neuronal signals for brain-computer interfaces

et al. 2014

View full text Add to dashboard Cite

Brain-computer interfaces (BCIs) require demanding numerical computations to transfer brain signals into control signals driving an external actuator. Increasing the computational performance of the BCI algorithms carrying out these calculations enables faster reaction to user inputs and allows using more demanding decoding algorithms. Here we introduce a modular and extensible software architecture with a multi-threaded signal processing pipeline suitable for BCI applications. The computational load and latency (the time that the system needs to react to user input) are measured for different pipeline implementations in typical BCI applications with realistic parameter settings. We show that BCIs can benefit substantially from the proposed parallelization: firstly, by reducing the latency and secondly, by increasing the amount of recording channels and signal features that can be used for decoding beyond the amount which can be handled by a single thread. The proposed software architecture provides a simple, yet flexible solution for BCI applications.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Low-latency multi-threaded processing of neuronal signals for brain-computer interfaces

et al. 2014

View full text Add to dashboard Cite

show abstract

“…These heuristics are listed in Table 1. 19 Mark x i as invalid; In the table, it is apparent that all heuristics are based on the spline model, which is general and problem-independent. The use of splines allows the algorithm to accurately predict parameter values, while preserving the first and second derivatives of the parameters across the sequence without generating oscillations.…”

Section: Extrapolation Heuristicsmentioning

confidence: 99%

“…The underlying process is assumed to be a linear dynamical system. Extended Kalman Filters [19] (EKF) generalize this solution for nonlinear dynamical systems. However, none of the existing methods provide a generalized solution for offline optimization of nonlinear dynamical systems.…”

Section: Related Workmentioning

confidence: 99%

Spline-based parallel nonlinear optimization of function sequences

Ben-Nun

Barak

Raviv

2016

Journal of Parallel and Distributed Computing

View full text Add to dashboard Cite

“…The P k equation is extended and written as follows: (24) By setting the previous equation to zero and solving it to obtain the optimal, (25) is obtained: (25) The special T k K obtained to minimize the mean square error of estimate in (25) is called the Kalman gain. The covariance matrix related to this estimate can be computed and obtained as (26) by substituting the optimal value in the overall relationship.…”

Section: Kalman Filter Algorithmmentioning

confidence: 99%

Locating a two-wheeled robot using extended Kalman filter

Javad

Cai

2015

Teh. vjesn.

View full text Add to dashboard Cite

Original scientific paper The Kalman filter is a very general method of filtering which can solve problems such as optimal estimation, prediction, noise filtering, and optimal control. A problem with detection of correct path of moving objects is the received noisy data. Therefore, it is possible that the information is incorrectly detected. There are Different methods to extract the correct data from the received information. This paper aims to detect the path of a two-wheeled robot using extended Kalman filter. For this purpose, triangular, circular, elliptical, and Sinusoidal paths were used to explore various scenarios. The results showed that the Kalman filter optimally detects the correct path with less than 3 % error rate. These results also show that error rate related to detect circular and triangular paths has the highest and lowest value, respectively, using the extended Kalman filter; in addition, the results showed that the error rate strongly depends on path changes. Keywords: path detection; the extended Kalman filter tracking; two-wheeled robot Lociranje robota s dva kotača primjenom proširenog Kalman filteraIzvorni znanstveni članak Kalman filter je vrlo općenita metoda filtriranja kojom se mogu riješiti problemi kao što su optimalna procjena, predviđanje, filtriranje buke i optimalna kontrola. Problem koji se javlja kod detekcije ispravne putanje gibajućih predmeta je prijem bučnih podataka. Stoga je moguće da se informacije neispravno pronađu. Postoje razne metode za dobivanje ispravnih podataka iz primljene obavijesti. U ovom radu cilj je otkriti putanju robota s dva kotača primjenom proširenog Kalman filtera. U tu su se svrhu rabile trokutaste, kružne, eliptične i sinusoidne putanje u istraživanju raznih scenarija. Rezultati pokazuju da se Kalmanovim filterom optimalno pronalazi ispravna putanja s manje od 3 % učestalosti pogreške. Ti rezultati također pokazuju da učestalost pogreške kod pronalaženja kružnih i trokutastih putanja ima najvišu i najnižu vrijednost primjenom Kalman filtera; uz to, rezultati su pokazali da učestalost pogreške uvelike ovisi o promjenama putanje.Ključne riječi: obilježavanje proširenim Kalman filterom; pronalaženje putanje; robot s dva kotača

show abstract

Kalman Filtering and Neural Networks

Cited by 1,206 publications

References 35 publications

Low-latency multi-threaded processing of neuronal signals for brain-computer interfaces

Low-latency multi-threaded processing of neuronal signals for brain-computer interfaces

Spline-based parallel nonlinear optimization of function sequences

Locating a two-wheeled robot using extended Kalman filter

Contact Info

Product

Resources

About