Experimental and Analytical Studies on Improved Feedforward ML Estimation Based on LS-SVR

Discrete Dynamics in Nature and Society ◽

10.1155/2013/192021 ◽

2013 ◽

Vol 2013 ◽

pp. 1-10

Author(s):

Xueqian Liu ◽

Hongyi Yu

Keyword(s):

Frequency Estimation ◽

Additive White Gaussian Noise ◽

Estimation Method ◽

Threshold Value ◽

Small Sample ◽

Support Vector ◽

Linear Filter ◽

Ml Estimation ◽

Single Tone ◽

Estimation Performance

Maximum likelihood (ML) algorithm is the most common and effective parameter estimation method. However, when dealing with small sample and low signal-to-noise ratio (SNR), threshold effects are resulted and estimation performance degrades greatly. It is proved that support vector machine (SVM) is suitable for small sample. Consequently, we employ the linear relationship between least squares support vector regression (LS-SVR)’s inputs and outputs and regard LS-SVR process as a time-varying linear filter to increase input SNR of received signals and decrease the threshold value of mean square error (MSE) curve. Furthermore, it is verified that by taking single-tone sinusoidal frequency estimation, for example, and integrating data analysis and experimental validation, if LS-SVR’s parameters are set appropriately, not only can the LS-SVR process ensure the single-tone sinusoid and additive white Gaussian noise (AWGN) channel characteristics of original signals well, but it can also improves the frequency estimation performance. During experimental simulations, LS-SVR process is applied to two common and representative single-tone sinusoidal ML frequency estimation algorithms, the DFT-based frequency-domain periodogram (FDP) and phase-based Kay ones. And the threshold values of their MSE curves are decreased by 0.3 dB and 1.2 dB, respectively, which obviously exhibit the advantage of the proposed algorithm.

Download Full-text

Comparative Analysis of Various Single-tone Frequency Estimation Techniques in High-order Instantaneous Moments Based Phase Estimation Method

10.1063/1.3426140 ◽

2010 ◽

Author(s):

G. Rajshekhar ◽

Sai Siva Gorthi ◽

Pramod Rastogi ◽

Pramod K. Rastogi ◽

Erwin Hack

Keyword(s):

Comparative Analysis ◽

Frequency Estimation ◽

Estimation Method ◽

High Order ◽

Phase Estimation ◽

Tone Frequency ◽

Single Tone ◽

Estimation Techniques

Download Full-text

Drowsiness Estimation Using Electroencephalogram and Recurrent Support Vector Regression

Information ◽

10.3390/info10060217 ◽

2019 ◽

Vol 10 (6) ◽

pp. 217 ◽

Cited By ~ 1

Author(s):

Izzat Aulia Akbar ◽

Tomohiko Igasaki

Keyword(s):

Support Vector Regression ◽

Estimation Method ◽

Electrode Site ◽

Estimation Methods ◽

Support Vector ◽

Physical Damage ◽

Prediction Ability ◽

Estimation Performance ◽

Eeg Data ◽

Electroencephalogram Eeg

As a cause of accidents, drowsiness can cause economical and physical damage. A range of drowsiness estimation methods have been proposed in previous studies to aid accident prevention and address this problem. However, none of these methods are able to improve their estimation ability as the length of time or number of trials increases. Thus, in this study, we aim to find an effective drowsiness estimation method that is also able to improve its prediction ability as the subject’s activity increases. We used electroencephalogram (EEG) data to estimate drowsiness, and the Karolinska sleepiness scale (KSS) for drowsiness evaluation. Five parameters (α, β/α, (θ+α)/β, activity, and mobility) from the O1 electrode site were selected. By combining these parameters and KSS, we demonstrate that a typical support vector regression (SVR) algorithm can estimate drowsiness with a correlation coefficient (R2) of up to 0.64 and a root mean square error (RMSE) of up to 0.56. We propose a “recurrent SVR” (RSVR) method with improved estimation performance, as highlighted by an R2 value of up to 0.83, and an RMSE of up to 0.15. These results suggest that in addition to being able to estimate drowsiness based on EEG data, RSVR is able to improve its drowsiness estimation performance.

Download Full-text

Widely Linear Adaptive Instantaneous Frequency Estimation in Vector Hydrophones

Sensors ◽

10.3390/s18103348 ◽

2018 ◽

Vol 18 (10) ◽

pp. 3348 ◽

Cited By ~ 1

Author(s):

Panpan Peng ◽

Liang An

Keyword(s):

Instantaneous Frequency ◽

Frequency Estimation ◽

Linear Method ◽

Estimation Method ◽

Frequency Resolution ◽

Linear Filter ◽

Instantaneous Frequency Estimation ◽

Time Frequency ◽

Short Time ◽

Widely Linear

To solve the problem that the time-frequency resolution of Short-Time Fourier Transform (STFT) is constrained by the window length and the moving step of the short time window, and to utilize the merits of a widely linear method, a novel instantaneous frequency estimation method in vector hydrophone was proposed. In this paper, a complex variable was constructed. It is composed of sound pressure and particle velocity as its real part and imaginary part, respectively. The constructed variable was approved to be second order noncircular (improper). For the modelling of noncircular signals, the standard linear estimation is not adequate and the pseudo-covariance matrix should also be taken into consideration. As a result, a widely linear adaptive instantaneous frequency estimation algorithm and its three solutions based on the augmented complex least mean square (ACLMS) method are presented to estimate the instantaneous frequency in vector hydrophones. The results of simulations and laboratory experiments prove that this approach based on a widely linear model performs better compared to STFT and strict linear filter methods.

Download Full-text

Direct Instantaneous Frequency Rate Estimation to Improve the Carrier Estimation Performance in Mars Entry, Descent, and Landing Flight

Remote Sensing ◽

10.3390/rs10081259 ◽

2018 ◽

Vol 10 (8) ◽

pp. 1259 ◽

Cited By ~ 1

Author(s):

Wanhong Hao ◽

Xiaowei Cui ◽

Jianguang Feng ◽

Guangliang Dong ◽

Zhiyong Zhu

Keyword(s):

Instantaneous Frequency ◽

Search Algorithm ◽

Frequency Estimation ◽

Estimation Method ◽

Current Approach ◽

Estimation Accuracy ◽

Cubic Phase ◽

Estimation Performance ◽

Frequency Rate ◽

Coherent Integration

This paper focuses on the carrier estimation performance improvement in Mars entry, descent, and landing (EDL) flights. Carrier reconstruction could be used for trajectory derivation and Martian atmosphere profile inversion, and is the critical information for mission operations, as it helps determine the flight status of the spacecraft, demodulate the downlink information. The current approach is maximum likelihood estimation based on a two-dimensional (2D) maximum energy search algorithm, which computes the grid energy over all the combinations of frequency cells and frequency rate cells among the search space. Although it has good performance on robust estimation, the frequency estimation accuracy is limited due to the short coherent integration. An instantaneous frequency rate tracking approach based on the cubic phase function (CPF) is proposed that directly estimates the instantaneous frequency rate over the frequency rate cells, followed by the frequency estimation among the frequency cells. A sequential estimation method is introduced to propose the sequential CPF statistics, which uses the a priori Doppler phase information to suppress the noise squaring loss inherent in the standard CPF statistics. Simulations have been made on the released Mars Science Laboratory EDL trajectory for the two approaches, which show that considerable estimation improvement has been achieved for aerobraking flight by the new algorithm.

Download Full-text

A Simple and Effective NDA Carrier Frequency Estimation Method for MPSK Signals in Unknown Fading Channel

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.403-408.3108 ◽

2011 ◽

Vol 403-408 ◽

pp. 3108-3113

Author(s):

Zhe Yan ◽

Hua Xu ◽

Ping Li

Keyword(s):

Impulse Response ◽

Carrier Frequency ◽

Fading Channel ◽

Frequency Estimation ◽

Research Result ◽

Estimation Method ◽

Joint Estimation ◽

Estimation Performance ◽

Channel Effects ◽

Carrier Frequency Estimation

The problem of the frequency estimation of MPSK signal in unknown fading channel is too complicated and there is no research result for this problem at present, while we often meet these problems in practice. The method mentioned in this paper compensates the channel effects before the estimation the carrier frequency. This method is more simple and efficient than the method which uses frequency and channel impulse response for the joint estimation, and it is more convenient to remove modulation for MPSK signal. Compared with the frequency estimation research of non-fading channel, the estimation performance falls only about 3dB. So, this method can satisfy most application needs.

Download Full-text

Estimating Frequency by Interpolation Using Least Squares Support Vector Regression

Mathematical Problems in Engineering ◽

10.1155/2015/648707 ◽

2015 ◽

Vol 2015 ◽

pp. 1-10

Author(s):

Changwei Ma

Keyword(s):

Least Squares ◽

Support Vector Regression ◽

High Frequency ◽

Fourier Coefficients ◽

Frequency Estimation ◽

Signal To Noise Ratio ◽

Threshold Value ◽

Estimation Accuracy ◽

Support Vector ◽

Size Number

Discrete Fourier transform- (DFT-) based maximum likelihood (ML) algorithm is an important part of single sinusoid frequency estimation. As signal to noise ratio (SNR) increases and is above the threshold value, it will lie very close to Cramer-Rao lower bound (CRLB), which is dependent on the number of DFT points. However, its mean square error (MSE) performance is directly proportional to its calculation cost. As a modified version of support vector regression (SVR), least squares SVR (LS-SVR) can not only still keep excellent capabilities for generalizing and fitting but also exhibit lower computational complexity. In this paper, therefore, LS-SVR is employed to interpolate on Fourier coefficients of received signals and attain high frequency estimation accuracy. Our results show that the proposed algorithm can make a good compromise between calculation cost and MSE performance under the assumption that the sample size, number of DFT points, and resampling points are already known.

Download Full-text