Search NASA⌕ Search

SEARCH · Search NASA

Results for “algorithm timings”

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.

At least 1,117 records · Page 62

Combining Machine Learning and Numerical Simulation for High-Resolution PM2.5 Concentration Forecast

Forecasting ambient PM2.5 concentrations with spatiotemporal coverage is key to alerting decision-makers of pollution episodes and preventing detrimental public exposure, especially in regions with limited ground air monitoring stations. The existing methods either rely on chemical transport models (CTMs) to forecast spatial distribution of PM2.5 with nontrivial uncertainty or statistical algorithms to forecast PM2.5 concentration time-series at air monitoring locations without continuous spatial coverage. In this study, we developed a PM2.5 forecast framework by combining the robust Random Forest algorithm with a publicly accessible global CTM forecast product – NASA’s Goddard Earth Observing System “Composition Forecasting” (GEOS-CF), providing spatiotemporally continuous PM2.5 concentration forecasts for the next five days at a 1-km spatial resolution. Our forecast experiment was conducted for a region in Central China including the populous and polluted Fenwei Plain. The forecast for the next two days had overall validation R2 of 0.76 and 0.64, respectively; the R2 was around 0.5 for the following three forecast days. Spatial cross-validation showed similar validation metrics. Our forecast model, with validation normalized mean bias close to zero, substantially reduced the large biases in GEOS-CF. The proposed framework requires minimal computational resources compared to running CTMs at urban scales, enabling near-real-time PM2.5 forecast in resource-restricted environments.

PM2.5↗

Improved local linearization algorithm for solving the quaternion equations

The objective of this paper is to develop a new and more accurate local linearization algorithm for numerically solving sets of linear time-varying differential equations. Of special interest is the application of this algorithm to the quaternion rate equations. The results are compared, both analytically and experimentally, with previous results using local linearization methods. The new algorithm requires approximately one-third more calculations per step than the previously developed local linearization algorithm; however, this disadvantage could be reduced by using parallel implementation. For some cases the new algorithm yields significant improvement in accuracy, even with an enlarged sampling interval. The reverse is true in other cases. The errors depend on the values of angular velocity, angular acceleration, and integration step size. One important result is that for the worst case the new algorithm can guarantee eigenvalues nearer the region of stability than can the previously developed algorithm.

Yen, K.↗

Using the boundary-layer equations in three-dimensional viscous flow simulation

A simple and computationally efficient algorithm for solving the unsteady three dimensional boundary layer equations in a time accurate or relaxation mode is presented. The algorithm is used to compute the separated flow over a 6:1 ellipsoid at angle of attack and the transonic flow over a finite wing with shock induced separation. These results are compared to available experimental data. A method of using the boundary layer algorithm to accelerate a Navier-Stokes computation is also presented. Example computations indicate that a boundary layer algorithm can be used to improve both the accuracy and efficiency of a Navier-Stokes algorithm.

Vandalsem, William R.↗

On the effect of grain burnback on STS-SRM fragment velocity

Concerns raised during the Ulysses Final Safety Analysis Review (FSAR) process called the solid rocket motor (SRM) fragment velocity prediction model into question. The specific area of concern was that there was a section of the SRM casing which was exposed to SRM chamber pressure as the grain (fuel) was consumed. These questions centered on the velocity of fragments which originated from the field joint region given that failure occurred between 37 and 72 seconds mission elapsed time (MET). Two dimensional coupled Eulerian-Lagrangian calculations were performed to assess the hot gas flow field which resulted from SRM casing fragmentation. The fragment to gas interface-pressure time-history obtained from these analyses was reduced to a boundary condition algorithm which was applied to an explicit-time-integration, finite element, three dimensional shell model of the SRM casing and unburned fuel. The results of these calculations showed that the velocity of fragments originating in the field joint was adequately described by the range of velocities given in the Shuttle Data Book (1988). Based on these results, no further analyses were required, and approval was obtained from the Launch Abort Subpanel of the Interagency Nuclear Safety Review Panel to use the SRM fragment velocity environments presented in the Ulysses FSAR (1990).

Eck, Marshall B.↗

A Fast Implementation of the ISOCLUS Algorithm

Unsupervised clustering is a fundamental tool in numerous image processing and remote sensing applications. For example, unsupervised clustering is often used to obtain vegetation maps of an area of interest. This approach is useful when reliable training data are either scarce or expensive, and when relatively little a priori information about the data is available. Unsupervised clustering methods play a significant role in the pursuit of unsupervised classification. One of the most popular and widely used clustering schemes for remote sensing applications is the ISOCLUS algorithm, which is based on the ISODATA method. The algorithm is given a set of n data points (or samples) in d-dimensional space, an integer k indicating the initial number of clusters, and a number of additional parameters. The general goal is to compute a set of cluster centers in d-space. Although there is no specific optimization criterion, the algorithm is similar in spirit to the well known k-means clustering method in which the objective is to minimize the average squared distance of each point to its nearest center, called the average distortion. One significant feature of ISOCLUS over k-means is that clusters may be merged or split, and so the final number of clusters may be different from the number k supplied as part of the input. This algorithm will be described in later in this paper. The ISOCLUS algorithm can run very slowly, particularly on large data sets. Given its wide use in remote sensing, its efficient computation is an important goal. We have developed a fast implementation of the ISOCLUS algorithm. Our improvement is based on a recent acceleration to the k-means algorithm, the filtering algorithm, by Kanungo et al.. They showed that, by storing the data in a kd-tree, it was possible to significantly reduce the running time of k-means. We have adapted this method for the ISOCLUS algorithm. For technical reasons, which are explained later, it is necessary to make a minor modification to the ISOCLUS specification. We provide empirical evidence, on both synthetic and Landsat image data sets, that our algorithm's performance is essentially the same as that of ISOCLUS, but with significantly lower running times. We show that our algorithm runs from 3 to 30 times faster than a straightforward implementation of ISOCLUS. Our adaptation of the filtering algorithm involves the efficient computation of a number of cluster statistics that are needed for ISOCLUS, but not for k-means.

Memarsadeghi, Nargess↗

Collision detection for spacecraft proximity operations

Collision Detection for Spacecraft Proximity Operations This thesis describes the development of a new collision detection algorithm to be used when two spacecraft are operating in the same vicinity. The two spacecraft are modelled as unions of convex polyhedra, where the polyhedron resulting from the union may be either convex or nonconvex. The relative motion of the two spacecraft is assumed to be such that one vehicle is moving with constant linear and angular velocity with respect to the other. The algorithm determines if a collision is possible and, if so, predicts the time when the collision will take place. The theoretical basis for the new collision detection algorithm is the C-function formulation of the configuration space approach recently introduced by researchers in robotics. Three different types of C-functions are defined that model the contacts between the vertices, edges, and faces of the polyhedra representing the two spacecraft. These C-functions are used to formulate three "collision" conditions. The first of these conditions limits the points representing potential collisions to the zeros of the C-functions. The new algorithm is fundamentally a search for the smallest zero of any C-function that satisfies the second and third collision conditions. The C-functions are shown to be transcendental functions of time for the assumed trajectory of the moving spacecraft. The zeros of these functions cannot be expressed in dosed form. Therefore, numerical search procedures are developed to find aLl of the zeros of a C-function in specified bounded intervals of time. These bounded intervals of time are found by examining the second and third collision conditions. The capabilities of the new algorithm are demonstrated for several example cases. These include examples of collisions determined by zeros of each of the three different types of C-functions. In addition to predicting the time of first contact of the polyhedra, the algorithm identifies the features of the two polyhedra that are touching at this time. The new collision detection algorithm is the first such algorithm that is capable of solving the collision detection problem exactly for the case where the moving object has constant linear and angular velocities. This is a significant improvement on previous collision detection algorithms described in the literature. In particular, the ability to handle constant angular velocity represents a more realistic type of rotational motion than those which have been used in other algorithms.

Robin M Vaughan↗

Robust Optimization Design Algorithm for High-Frequency TWTs

Traveling-wave tubes (TWTs), such as the Ka-band (26-GHz) model recently developed for the Lunar Reconnaissance Orbiter, are essential as communication amplifiers in spacecraft for virtually all near- and deep-space missions. This innovation is a computational design algorithm that, for the first time, optimizes the efficiency and output power of a TWT while taking into account the effects of dimensional tolerance variations. Because they are primary power consumers and power generation is very expensive in space, much effort has been exerted over the last 30 years to increase the power efficiency of TWTs. However, at frequencies higher than about 60 GHz, efficiencies of TWTs are still quite low. A major reason is that at higher frequencies, dimensional tolerance variations from conventional micromachining techniques become relatively large with respect to the circuit dimensions. When this is the case, conventional design- optimization procedures, which ignore dimensional variations, provide inaccurate designs for which the actual amplifier performance substantially under-performs that of the design. Thus, this new, robust TWT optimization design algorithm was created to take account of and ameliorate the deleterious effects of dimensional variations and to increase efficiency, power, and yield of high-frequency TWTs. This design algorithm can help extend the use of TWTs into the terahertz frequency regime of 300-3000 GHz. Currently, these frequencies are under-utilized because of the lack of efficient amplifiers, thus this regime is known as the "terahertz gap." The development of an efficient terahertz TWT amplifier could enable breakthrough applications in space science molecular spectroscopy, remote sensing, nondestructive testing, high-resolution "through-the-wall" imaging, biomedical imaging, and detection of explosives and toxic biochemical agents.

Wilson, Jeffrey D.↗

Applications of the Dynamic N-Dimensional K-Vector

The n-dimensional k-vector (NDKV) is an appealing alternative to binary tress for resolving complex queries in large relational databases. The method has excelled in several applications involving static databases. The present paper extends the theory supporting the NDKV to handle dynamic databases, where the data is updated frequently. This includes deleting records, adding new entries, or editing existing elements. The merit of this new version of the NDKV, the dynamic n-dimensional k-vector (DNDKV), is that it is no longer necessary to recompute the entire k-vector (the main structure that indexes the data) every time a record changes. The algorithm updates the four constituents of the standard NDKV on the fly: the database, sorted database, index, and k-vector tables. As a result, the DNDKV becomes comparable in terms of capabilities and flexibility to stateof-the-art storage engines relying on structured query languages (SQL). The performance of the DNDKV is assessed by running typical read/write operations on a database that contains millions of pre-computed missions to celestial bodies. This database requires frequent updates whenever an orbit solution is refined or new bodies are discovered. The DNDKV is faster than rebuilding the k-vector tables completely, provided that the number of elements being added or removed is not excessively large. Direct runtime comparisons with MySQL suggest that the DNDKV is several times faster for reading but might be slower for writing and updating the database. One limit of the technique is the elements being added must be within the range of the current k-vector tables. If this is not the case, the technique cannot be used and the k-vector tables must be rebuilt from scratch.

Mortari, Daniele↗

Time-dependent viscous incompressible Navier-Stokes equations - The finite difference Galerkin formulation and streamfunction algorithms

Numerical techniques are developed to solve the Navier-Stokes equations for unsteady incompressible flow. The extension of the finite-difference Galerkin (FDG) method of Stephens et al. (1984) to the continuous-time case in two or three space dimensions is explained, and the numerical implementation of the method is discussed with particular attention to the staggered-MAC-grid primitive-variable discretization, the application of discrete mass balance to avoid problems inherent in FDG schemes, the direct interpretation of the FDG expansion variables as a discrete streamfunction, and a mass-balance approach to two-dimensional problems with throughflow or obstacles. Numerical results are presented graphically for the evolution of asymptotic steady flow in a driven cavity at Reynolds number 400, 1000, or 3200; good agreement with published experimental data is demonstrated, with accurate predictions of secondary-vortex formation from wall bubble recirculations at Reynolds number 1000.

Goodrich, John W.↗

An efficient parallel algorithm for the solution of a tridiagonal linear system of equations.

Tridiagonal linear systems of equations can be 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 computation 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.↗

Determining the 3-D structure and motion of objects using a scanning laser range sensor

In order for the EVAHR robot to autonomously track and grasp objects, its vision system must be able to determine the 3-D structure and motion of an object from a sequence of sensory images. This task is accomplished by the use of a laser radar range sensor which provides dense range maps of the scene. Unfortunately, the currently available laser radar range cameras use a sequential scanning approach which complicates image analysis. Although many algorithms have been developed for recognizing objects from range images, none are suited for use with single beam, scanning, time-of-flight sensors because all previous algorithms assume instantaneous acquisition of the entire image. This assumption is invalid since the EVAHR robot is equipped with a sequential scanning laser range sensor. If an object is moving while being imaged by the device, the apparent structure of the object can be significantly distorted due to the significant non-zero delay time between sampling each image pixel. If an estimate of the motion of the object can be determined, this distortion can be eliminated; but, this leads to the motion-structure paradox - most existing algorithms for 3-D motion estimation use the structure of objects to parameterize their motions. The goal of this research is to design a rigid-body motion recovery technique which overcomes this limitation. The method being developed is an iterative, linear, feature-based approach which uses the non-zero image acquisition time constraint to accurately recover the motion parameters from the distorted structure of the 3-D range maps. Once the motion parameters are determined, the structural distortion in the range images is corrected.

Nandhakumar, N.↗

A method for using a time interval counter to measure frequency stability

It is shown how a commercial time interval counter can be used to measure the relative stability of two signals that are offset in frequency and mixed down to a beat note of about 1 Hz. To avoid the dead-time problem, the counter is set up to read the time interval between each beat note upcrossing and the next pulse of a 10 Hz reference pulse train. The actual upcrossing times are recovered by a simple algorithm whose outputs can be used for computing residuals and Allan variance. A noise floor-test yielded a delta f/f Allan deviation of 1.3 times 10 to the minus 9th power/tau relative to the beat frequency.

Greenhall, C. A.↗

Evaluation of improved algorithms for nonlinear transient structural heat transfer

The paper presents the evaluation of algorithms having potential for reducing computer time to integrate large systems of equations governing nonlinear transient heat transfer in complex structures. A series of test problems is being developed to test two algorithms. The Shuttle Orbiter frame model having 190 nodes was the subject of an analytical/experimental study. Plans are to perform calculations for the complete wing model having about 4000 nodes as well as a regional (plug) model having about 200 nodes.

Adelman, H. M.↗

Determining F-factor using ground-based Doppler radar: Validation and results

Using a two-dimensional linear least-squares method applied to Doppler radar data, we test the viability of determining F-factor remotely. The ultimate application of such an algorithm will be supplying real-time F-factor maps, derived from ground-based Doppler radars to air traffic control personnel and pilots. Data from NASA deployments to the MIT/Lincoln Lab TDWR testbed radar in Orlando in 1991 and 1992 along with NASA deployments to the NCAR TDWR testbed radar in Denver are examined. Preliminary analyses show that the two-dimensional method correlates reasonably well with in situ measurements. Several effects, independent of the method used, act to reduce the correlation to less than one. These include time differences between radar and aircraft data, vertical misalignment between the aircraft and the radar beam, different spatial resolution scales between aircraft and radar data, inhomogeneous radar beam filling, noise in radar data that eludes filtering, and phase lag between time and space due to low pass filtering of the aircraft data. In the final assessment, it appears that a shear-based F-factor algorithm is preferable to the currently implemented TDWR algorithms which lack any local shear estimates.

Elmore, K. L.↗

Real-time image enhancement

Pipelined system with "vision" algorithm is implemented on LSI chip that processes input digital image data to produce image-edge map. System contains 3 input adder, difference and absolute value cells, and adder and comparator. Data store for 1 to 2 ms, and are easily transmitted or isolated; design has reduced package count and number of interconnections for increased reliability. Applications include locating objects on moving belt, deep-sea and coal mining, and control of robotic rovers.

Wong, V. S.↗

Some design guidelines for discrete-time adaptive controllers

There have been many algorithms proposed for adaptive control which will provide globally asymptotically stable controllers if some stringent conditions on the plant are met. The conditions on the plant cannot be met in practice as all plants will contain high frequency unmolded dynamics therefore, blind implementation of the published algorithms can lead to disastrous results. This paper uses a linearization analysis of a non-linear adaptive controller to demonstrate analytically design guidelines which aleviate some of the problems associated with adaptive control in the presence of unmodeled dynamics.

Rohrs, C. E.↗

Roundoff error effects on spatial lattice algorithm

The floating-point roundoff error effect under finite word length limitations is analyzed for the time updates of reflection coefficients in the spatial lattice algorithm. It is shown that recursive computation is superior to direct computation under finite word length limitations. Moreover, the forgetting factor, which is conventionally used to smooth the time variations of the inputs, is also a crucial parameter in the consideration of the system stability and adaptability under finite word length constraints.

An, S. H.↗

A method for using a time interval counter to measure frequency stability

It is shown how a commercial time interval counter can be used to measure the relative stability of two signals that are offset in frequency and mixed down to a beat note of about 1 Hz. To avoid the dead-time problem, the counter is set up to read the time interval between each beat note upcrossing and the next pulse of a 10 Hz reference pulse train. The actual upcrossing times are recovered by a simple algorithm whose outputs can be used for computing residuals and Allan variance. A noise floor-test yielded a delta f-f Allan deviation of 1.3 times 10 to the minus 9th power/tau relative to the beat frequency.

Greenhall, Charles A.↗