Application of Recurrent Neural Network to Modeling Earth's Global Electron Density
Explore the source record for details and available documents.
Sheng Huang↗
SEARCH · Engineering Papers
Results for “recurrence”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Recurrent seascape units identify key ecological processes along the western Antarctic Peninsula
Explore the source record for details and available documents.
Jeff S. Bowman↗
Magnetic fields and flows between 1 AU and 0.3 AU during the primary mission of HELIOS 1
The recurrent flow and field patterns observed by HELIOS 1, and the relation between these patterns and coronal holes are discussed. Four types of recurrent patterns were observed: a large recurrent stream, a recurrent slow (quiet) flow, a rapidly evolving flow, and a recurrent compound stream. There recurrent streams were not stationary, for although the sources recurred at approximately the same longitudes on successive rotations, the shapes and latitudinal patterns changed from one rotation to the next. A type of magnetic field and plasma structure characterized by a low ion temperature and a high magnetic field intensity is described as well as the structures of stream boundaries between the sun at approximately 0.3 AU.
Burlaga, L. F.↗
Predicting Solar Cycle 24 Using a Geomagnetic Precursor Pair
We describe using Ap and F(10.7) as a geomagnetic-precursor pair to predict the amplitude of Solar Cycle 24. The precursor is created by using F(10.7) to remove the direct solar-activity component of Ap. Four peaks are seen in the precursor function during the decline of Solar Cycle 23. A recurrence index that is generated by a local correlation of Ap is then used to determine which peak is the correct precursor. The earliest peak is the most prominent but coincides with high levels of non-recurrent solar activity associated with the intense solar activity of October and November 2003. The second and third peaks coincide with some recurrent activity on the Sun and show that a weak cycle precursor closely following a period of strong solar activity may be difficult to resolve. A fourth peak, which appears in early 2008 and has recurrent activity similar to precursors of earlier solar cycles, appears to be the "true" precursor peak for Solar Cycle 24 and predicts the smallest amplitude for Solar Cycle 24. To determine the timing of peak activity it is noted that the average time between the precursor peak and the following maximum is approximately equal to 6.4 years. Hence, Solar Cycle 24 would peak during 2014. Several effects contribute to the smaller prediction when compared with other geomagnetic-precursor predictions. During Solar Cycle 23 the correlation between sunspot number and F(10.7) shows that F(10.7) is higher than the equivalent sunspot number over most of the cycle, implying that the sunspot number underestimates the solar-activity component described by F(10.7). During 2003 the correlation between aa and Ap shows that aa is 10 % higher than the value predicted from Ap, leading to an overestimate of the aa precursor for that year. However, the most important difference is the lack of recurrent activity in the first three peaks and the presence of significant recurrent activity in the fourth. While the prediction is for an amplitude of Solar Cycle 24 of 65 +/- 20 in smoothed sunspot number, a below-average amplitude for Solar Cycle 24, with maximum at 2014.5+/-0.5, we conclude that Solar Cycle 24 will be no stronger than average and could be much weaker than average.
predictions↗
Some MACSYMA program for solving difference equations
A set of MACSYMA programs are described for finding closed form solutions to linear recurrence relations in equations having either constant or variable coefficients. In the homogenous case, a polymonial equation is obtained and the solution to the recurrence relation can be written as a linear combination of the roots of the polynomial. Exponential generating functions are used to solve variable coefficient relations. Taking successive derivatives and using the recurrence relation, an ordinary differential equation is obtained. Expanding the solution to the differential equation in a Taylor series, shows that the nth term of the series is the solution to the recurrence relation. For second order recurrences, a check is made for those that can be solved in terms of Bessel functions.
Ivie, J.↗
Identification and Control of Aircrafts using Multiple Models and Adaptive Critics
We compared two possible implementations of local linear models for control: one approach is based on a self-organizing map (SOM) to cluster the dynamics followed by a set of linear models operating at each cluster. Therefore the gating function is hard (a single local model will represent the regional dynamics). This simplifies the controller design since there is a one to one mapping between controllers and local models. The second approach uses a soft gate using a probabilistic framework based on a Gaussian Mixture Model (also called a dynamic mixture of experts). In this approach several models may be active at a given time, we can expect a smaller number of models, but the controller design is more involved, with potentially better noise rejection characteristics. Our experiments showed that the SOM provides overall best performance in high SNRs, but the performance degrades faster than with the GMM for the same noise conditions. The SOM approach required about an order of magnitude more models than the GMM, so in terms of implementation cost, the GMM is preferable. The design of the SOM is straight forward, while the design of the GMM controllers, although still reasonable, is more involved and needs more care in the selection of the parameters. Either one of these locally linear approaches outperform global nonlinear controllers based on neural networks, such as the time delay neural network (TDNN). Therefore, in essence the local model approach warrants practical implementations. In order to call the attention of the control community for this design methodology we extended successfully the multiple model approach to PID controllers (still today the most widely used control scheme in the industry), and wrote a paper on this subject. The echo state network (ESN) is a recurrent neural network with the special characteristics that only the output parameters are trained. The recurrent connections are preset according to the problem domain and are fixed. In a nutshell, the states of the reservoir of recurrent processing elements implement a projection space, where the desired response is optimally projected. This architecture trades training efficiency by a large increase in the dimension of the recurrent layer. However, the power of the recurrent neural networks can be brought to bear on practical difficult problems. Our goal was to implement an adaptive critic architecture implementing Bellman s approach to optimal control. However, we could only characterize the ESN performance as a critic in value function evaluation, which is just one of the pieces of the overall adaptive critic controller. The results were very convincing, and the simplicity of the implementation was unparalleled.
Principe, Jose C.↗
Short-term variations of the galactic cosmic ray intensity - 1964-1967
A statistical analysis of time variations in ground-level nucleonic cosmic-ray intensity for the interval from 1964 to 1967 is presented which incorporates synoptic observations of the solar white-light corona as well as indices of photospheric and chromospheric activity. Correlation analysis of solar activity and short-term modulation reveals that all indices vary significantly on time scales near the solar rotation period and that the correlation function exhibits a quasi-sinusoidal variation that maximizes near zero lag. It is found that 27-day variations of the indices were most pronounced in 1966-1967, that recurrent cosmic-ray depressions occurred in conjunction with observed solar-wind disturbances, and that two categories of interplanetary disturbance are sufficient to account for the large recurrent cosmic-ray depressions in 1966-1967. It is suggested that flare-generated shocks were the main source of these recurrent cosmic-ray variations and that the contribution of corotating cosmic-ray disturbances to the observed recurrent modulation was probably very small. This hypothesis is tested against the correlation functions of solar and cosmic-ray indices.
Parker, G. D.↗
On the Bimodality of ENSO Cycle Extremes
On the basis of sea surface temperature in the Nino 3.4 region (5 deg N-5 deg S, 120 deg- 170 deg W) during the interval of 1950-1997, Kevin Trenberth previously has identified some 16 El Nino and 10 La Nina, these 26 events representing the extremes of the quasi-periodic El Nino-Southern Oscillation (ENSO) cycle. Runs testing shows that the duration and recurrence period associated with these extremes vary randomly, as does the sequencing of the extremes. Hence, the frequency of occurrence of these events during the 1990s, especially, for El Nino should not be construed as being significantly different from that of previous epochs. Additionally, the distribution of duration for both El Nino and La Nina looks bimodal, consisting of two preferred modes - about 8 and 16 months in length for El Nino and about 9 and 18 months in length for La Nina. Likewise, the distribution of recurrence period, especially, for El Nino looks bimodal, consisting of two preferred modes - about 21 and 50 months in length. Scatter plots of the recurrence period versus duration for El Nino strongly suggest preferential associations between them, linking shorter (longer) duration with shorter (longer) recurrence period. Because the last known onset of El Nino occurred in April 1997 and the event was of longer than average duration, one infers that the onset of the next expected El Nino will not occur until February 2000 or later.
Wilson, Robert M.↗
On the Bimodality of ENSO Cycle Extremes
On the basis of sea surface temperature in the El Nino 3.4 region (5 deg. N.,-5 deg. S., 120-170 deg. W.) during the interval of 1950-1997, Kevin Trenberth previously has identified some 16 El Nino and 10 La Nina, these 26 events representing the extremes of the quasi-periodic El Nino-Southern Oscillation (ENSO) cycle. Runs testing shows that the duration, recurrence period, and sequencing of these extremes vary randomly. Hence, the decade of the 1990's, especially for El Nino, is not significantly different from that of previous decadal epochs, at least, on the basis of the frequency of onsets of ENSO extremes. Additionally, the distribution of duration for both El Nino and La Nina looks strikingly bimodal, each consisting of two preferred modes, about 8- and 16-mo long for El Nino and about 9- and 18-mo long for La Nina, as does the distribution of the recurrence period for El Nino, consisting of two preferred modes about 21- and 50-mo long. Scatterplots of the recurrence period versus duration for El Nino are found to be statistically important, displaying preferential associations that link shorter (longer) duration with shorter (longer) recurrence periods. Because the last onset of El Nino occurred in April 1997 and the event was of longer than average duration, onset of the next anticipated El Nino is not expected until February 2000 or later.
Wilson, Robert M.↗
On The Bimodality of ENSO Cycle Extremes
On the basis of sea surface temperature in the El Nino 3.4 region (5N.-5S., 120-170W.) during the interval of 1950-1997, Kevin Trenberth previously has identified some 16 El Nino and 10 La Nina, these 26 events representing the extremes of the quasi-periodic El Nino-Southern Oscillation (ENSO) cycle. Runs testing shows that the duration, recurrence period, and sequencing of these extremes vary randomly. Hence, the decade of the 1990's, especially for El Nino, is not significantly different from that of previous decadal epochs, at least, on the basis of the frequency of onsets of ENSO extremes. Additionally, the distribution of duration for both El Nino and La Nina looks strikingly bimodal, each consisting of two preferred modes, about 8- and 16-months long for El Nino and about 9- and 18-months long for La Nina, as does the distribution of the recurrence period for El Nino, consisting of two preferred modes about 21- and 50- mo long. Scatterplots of the recurrence period versus duration for El Nino are found to be statistically important, displaying preferential associations that link shorter (longer) duration with shorter (longer) recurrence periods. Because the last onset of El Nino occurred in April 1997 and the event was of longer than average duration, onset of the next anticipated El Nino is not expected until February 2000 or later.
Wilson, Robert M.↗
An Accreting White Dwarf near the Chandrasekhar Limit in the Andromeda Galaxy
The iPTF (Intermediate Palomar Transient Factory) detection of the most recent outburst of the recurrent nova system RX J0045.4+4154 in the Andromeda Galaxy has enabled the unprecedented study of a massive (mass is greater than 1.3 solar masses) accreting white dwarf (WD). We detected this nova as part of the near daily iPTF monitoring of M31 to a depth of R (red band-pass filter) approximately equal to magnitude 21 and triggered optical photometry, spectroscopy and soft X-ray monitoring of the outburst. Peaking at an absolute magnitude of MR (red, mid-infrared band-pass filter) equals magnitude −6.6, and with a decay time of 1 magnitude per day, it is a faint and very fast nova. It shows optical emission lines of He/N and expansion velocities of 1900 to 2600 kilometers per second 1-4 days after the optical peak. The Swift monitoring of the X-ray evolution revealed a supersoft source (SSS) with kT (energy: Boltzmann constant times temperature) (sub eff (effective)) approximately equal to 90-110 electronvolts that appeared within 5 days after the optical peak, and lasted only 12 days. Most remarkably, this is not the first event from this system, rather it is a recurrent nova with a time between outbursts of approximately 1 year, the shortest known. Recurrent X-ray emission from this binary was detected by ROSAT in 1992 and 1993, and the source was well characterized as a mass greater than 1.3 solar masses WD SSS. Based on the observed recurrence time between different outbursts, the duration and effective temperature of the SS phase, MESA models of accreting WDs allow us to constrain the accretion rate to mass greater than 1.7x10 (sup −7) solar masses per year and WD mass greater than 1.30 solar masses. If the WD keeps 30 percent of the accreted material, it will take less than a million years to reach core densities high enough for carbon ignition (if made of C/O) or electron capture (if made of O/Ne) to end the binary evolution.
dwarf↗
Artifacts produced during electrical stimulation of the vestibular nerve in cats
Evidence is presented to indicate that evoked potentials in the recurrent laryngeal, the cervical sympathetic, and the phrenic nerve, commonly reported as being elicited by vestibular nerve stimulation, may be due to stimulation of structures other than the vestibular nerve. Experiments carried out in decerebrated cats indicated that stimulation of the petrous bone and not that of the vestibular nerve is responsible for the genesis of evoked potentials in the recurrent laryngeal and the cervical sympathetic nerves. The phrenic response to electrical stimulation applied through bipolar straight electrodes appears to be the result of stimulation of the facial nerve in the facial canal by current spread along the petrous bone, since stimulation of the suspended facial nerve evoked potentials only in the phrenic nerve and not in the recurrent laryngeal nerve. These findings indicate that autonomic components of motion sickness represent the secondary reactions and not the primary responses to vestibular stimulation.
Tang, P. C.↗
Implementation of the block-Krylov boundary flexibility method of component synthesis
A method of dynamic substructuring is presented which utilizes a set of static Ritz vectors as a replacement for normal eigenvectors in component mode synthesis. This set of Ritz vectors is generated in a recurrence relationship, which has the form of a block-Krylov subspace. The initial seed to the recurrence algorithm is based on the boundary flexibility vectors of the component. This algorithm is not load-dependent, is applicable to both fixed and free-interface boundary components, and results in a general component model appropriate for any type of dynamic analysis. This methodology was implemented in the MSC/NASTRAN normal modes solution sequence using DMAP. The accuracy is found to be comparable to that of component synthesis based upon normal modes. The block-Krylov recurrence algorithm is a series of static solutions and so requires significantly less computation than solving the normal eigenspace problem.
Carney, Kelly S.↗
Pseudogradient Training For A Class Of Neural Networks
Developmental second-order recurrent neural networks of special type modified to enhance stability in face of inputs beyond range of inputs on which trained. Second-order recurrent neural networks contain product feedback units and can be trained, by use of example inputs and outputs, to act as finite-state automatons. Particular second-order recurrent neural networks in question learn grammars in sense they are trained to generate binary responses to input training sequences of ones and zeros, each sequence being marked "legal" or "illegal" according to grammar to be learned.
Zeng, Zheng↗
Improvements in Block-Krylov Ritz Vectors and the Boundary Flexibility Method of Component Synthesis
A method of dynamic substructuring is presented which utilizes a set of static Ritz vectors as a replacement for normal eigenvectors in component mode synthesis. This set of Ritz vectors is generated in a recurrence relationship, proposed by Wilson, which has the form of a block-Krylov subspace. The initial seed to the recurrence algorithm is based upon the boundary flexibility vectors of the component. Improvements have been made in the formulation of the initial seed to the Krylov sequence, through the use of block-filtering. A method to shift the Krylov sequence to create Ritz vectors that will represent the dynamic behavior of the component at target frequencies, the target frequency being determined by the applied forcing functions, has been developed. A method to terminate the Krylov sequence has also been developed. Various orthonormalization schemes have been developed and evaluated, including the Cholesky/QR method. Several auxiliary theorems and proofs which illustrate issues in component mode synthesis and loss of orthogonality in the Krylov sequence have also been presented. The resulting methodology is applicable to both fixed and free- interface boundary components, and results in a general component model appropriate for any type of dynamic analysis. The accuracy is found to be comparable to that of component synthesis based upon normal modes, using fewer generalized coordinates. In addition, the block-Krylov recurrence algorithm is a series of static solutions and so requires significantly less computation than solving the normal eigenspace problem. The requirement for less vectors to form the component, coupled with the lower computational expense of calculating these Ritz vectors, combine to create a method more efficient than traditional component mode synthesis.
Carney, Kelly Scott↗
Risk-based Classification of Incidents
As the penetration of software into safety-critical systems progresses, accidents and incidents involving software will inevitably become more frequent. Identifying lessons from these occurrences and applying them to existing and future systems is essential if recurrences are to be prevented. Unfortunately, investigative agencies do not have the resources to fully investigate every incident under their jurisdictions and domains of expertise and thus must prioritize certain occurrences when allocating investigative resources. In the aviation community, most investigative agencies prioritize occurrences based on the severity of their associated losses, allocating more resources to accidents resulting in injury to passengers or extensive aircraft damage. We argue that this scheme is inappropriate because it undervalues incidents whose recurrence could have a high potential for loss while overvaluing fairly straightforward accidents involving accepted risks. We then suggest a new strategy for prioritizing occurrences based on the risk arising from incident recurrence.
Greenwell, William S.↗
Education and Training Module in Alertness Management
The education and training module (ETM) in alertness management has now been integrated as part of the training regimen of the Pilot Proficiency Awards Program ("WINGS") of the Federal Aviation Administration. Originated and now maintained current by the Fatigue Countermeasures Group at NASA Ames Research Center, the ETM in Alertness Management is designed to give pilots the benefit of the best and most recent research on the basics of sleep physiology, the causes of fatigue, and strategies for managing alertness during flight operations. The WINGS program is an incentive program that encourages pilots at all licensing levels to participate in recurrent training, upon completion of which distinctive lapel or tie pins (wings) and certificates of completion are awarded. In addition to flight training, all WINGS applicants must attend at least one FAA-sponsored safety seminar, FAA-sanctioned safety seminar, or industry recurrent training program. The Fatigue Countermeasures Group provides an FAA-approved industry recurrent training program through an on-line General Aviation (GA) WINGS ETM in alertness management to satisfy this requirement. Since 1993, the Fatigue Countermeasures Group has translated fatigue and alertness information to operational environments by conducting two-day ETM workshops oriented primarily toward air-carrier operations subject to Part 121 of the Federal Aviation Regulations pertaining to such operations. On the basis of the information presented in the two-day ETM workshops, an ETM was created for GA pilots and was transferred to a Web-based version. To comply with the requirements of the WINGS Program, the original Web-based version has been modified to include hypertext markup language (HTML) content that makes information easily accessible, in-depth testing of alertness-management knowledge, new interactive features, and increased informational resources for GA pilots. Upon successful completion of this training module, a participant receives a computer- screen display of a certificate of completion. The certificate, which includes the pilot s name and an identifying number, can be printed out and submitted, for ground training credit, with the pilot s WINGS application.
Mallis, M. M.↗
An efficient parallel algorithm for the solution of a tridiagonal linear system of equations
Tridiagonal linear systems of equations are solved on conventional serial machines in a time proportional to N, where N is the number of equations. The conventional algorithms do not lend themselves directly to parallel computations on computers of the ILLIAC IV class, in the sense that they appear to be inherently serial. An efficient parallel algorithm is presented in which computation time grows as log sub 2 N. The algorithm is based on recursive doubling solutions of linear recurrence relations, and can be used to solve recurrence relations of all orders.
Stone, H. S.↗