Search NASA⌕ Search

SEARCH · Search NASA

Results for “Stochastic Algorithms”

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 325 records · Page 18

An Aircraft Separation Algorithm with Feedback and Perturbation

A separation algorithm is a set of rules that tell aircraft how to maneuver in order to maintain a minimum distance between them. This paper investigates demonstrating that separation algorithms satisfy the FAA requirement for the occurrence of incidents by means of simulation. Any demonstration that a separation algorithm, or any other aspect of flight, satisfies the FAA requirement is a challenge because of the stringent nature of the requirement and the complexity of airspace operations. The paper begins with a probability and statistical analysis of both the FAA requirement and demonstrating meeting it by a Monte Carlo approach. It considers the geometry of maintaining separation when one plane must change its flight path. It then develops a simple feedback control law that guides the planes on their paths. The presence of feedback control permits the introduction of perturbations, and the stochastic nature of the chosen perturbation is examined. The simulation program is described. This paper is an early effort in the realistic demonstration of a stringent requirement. Much remains to be done.

White, Allan L.↗

System-Level Prognostics for the National Airspace

In the National Airspace System (NAS), safety is assured through a set of rules, regulations, and procedures to respond to unsafe events. However, safety stands to benefit immensely from the introduction of tools and methodologies from Prognostics and Health Management (PHM). PHM will enable the NAS to stochastically predict unsafe states within the NAS, enabling a proactive preventative response strategy, as opposed to a reactive mitigative one. However, current PHM methods do not directly apply to the NAS for several reasons: they typically apply only at the component level, are implemented in a centralized manner, and are focused only on predicting remaining useful life. In this paper, we extend the model-based prognostics approach to PHM in order to provide a framework that can be applied to the NAS. We offer a system-level approach that supports a distributed implementation, and provide algorithms to predict the probability of an unsafe state, either at a specific time or within a time interval, and to predict the time of an unsafe state. Experimental results in simulation demonstrate the new approach.

prognostics↗

Language Model Combination and Adaptation Using Weighted Finite State Transducers

In speech recognition systems language model (LMs) are often constructed by training and combining multiple n-gram models. They can be either used to represent different genres or tasks found in diverse text sources, or capture stochastic properties of different linguistic symbol sequences, for example, syllables and words. Unsupervised LM adaption may also be used to further improve robustness to varying styles or tasks. When using these techniques, extensive software changes are often required. In this paper an alternative and more general approach based on weighted finite state transducers (WFSTs) is investigated for LM combination and adaptation. As it is entirely based on well-defined WFST operations, minimum change to decoding tools is needed. A wide range of LM combination configurations can be flexibly supported. An efficient on-the-fly WFST decoding algorithm is also proposed. Significant error rate gains of 7.3% relative were obtained on a state-of-the-art broadcast audio recognition task using a history dependently adapted multi-level LM modelling both syllable and word sequences

Liu, X.↗

Scope of Gradient and Genetic Algorithms in Multivariable Function Optimization

Global optimization of a multivariable function - constrained by bounds specified on each variable and also unconstrained - is an important problem with several real world applications. Deterministic methods such as the gradient algorithms as well as the randomized methods such as the genetic algorithms may be employed to solve these problems. In fact, there are optimization problems where a genetic algorithm/an evolutionary approach is preferable at least from the quality (accuracy) of the results point of view. From cost (complexity) point of view, both gradient and genetic approaches are usually polynomial-time; there are no serious differences in this regard, i.e., the computational complexity point of view. However, for certain types of problems, such as those with unacceptably erroneous numerical partial derivatives and those with physically amplified analytical partial derivatives whose numerical evaluation involves undesirable errors and/or is messy, a genetic (stochastic) approach should be a better choice. We have presented here the pros and cons of both the approaches so that the concerned reader/user can decide which approach is most suited for the problem at hand. Also for the function which is known in a tabular form, instead of an analytical form, as is often the case in an experimental environment, we attempt to provide an insight into the approaches focusing our attention toward accuracy. Such an insight will help one to decide which method, out of several available methods, should be employed to obtain the best (least error) output. *

Shaykhian, Gholam Ali↗

Dual adaptive control: Design principles and applications

The design of an actively adaptive dual controller based on an approximation of the stochastic dynamic programming equation for a multi-step horizon is presented. A dual controller that can enhance identification of the system while controlling it at the same time is derived for multi-dimensional problems. This dual controller uses sensitivity functions of the expected future cost with respect to the parameter uncertainties. A passively adaptive cautious controller and the actively adaptive dual controller are examined. In many instances, the cautious controller is seen to turn off while the latter avoids the turn-off of the control and the slow convergence of the parameter estimates, characteristic of the cautious controller. The algorithms have been applied to a multi-variable static model which represents a simplified linear version of the relationship between the vibration output and the higher harmonic control input for a helicopter. Monte Carlo comparisons based on parametric and nonparametric statistical analysis indicate the superiority of the dual controller over the baseline controller.

Mookerjee, Purusottam↗

Adaptive, Distributed Control of Constrained Multi-Agent Systems

Product Distribution (PO) theory was recently developed as a broad framework for analyzing and optimizing distributed systems. Here we demonstrate its use for adaptive distributed control of Multi-Agent Systems (MASS), i.e., for distributed stochastic optimization using MAS s. First we review one motivation of PD theory, as the information-theoretic extension of conventional full-rationality game theory to the case of bounded rational agents. In this extension the equilibrium of the game is the optimizer of a Lagrangian of the (Probability dist&&on on the joint state of the agents. When the game in question is a team game with constraints, that equilibrium optimizes the expected value of the team game utility, subject to those constraints. One common way to find that equilibrium is to have each agent run a Reinforcement Learning (E) algorithm. PD theory reveals this to be a particular type of search algorithm for minimizing the Lagrangian. Typically that algorithm i s quite inefficient. A more principled alternative is to use a variant of Newton's method to minimize the Lagrangian. Here we compare this alternative to RL-based search in three sets of computer experiments. These are the N Queen s problem and bin-packing problem from the optimization literature, and the Bar problem from the distributed RL literature. Our results confirm that the PD-theory-based approach outperforms the RL-based scheme in all three domains.

Bieniawski, Stefan↗

Tradeoff methods in multiobjective insensitive design of airplane control systems

The latest results of an ongoing study of computer-aided design of airplane control systems are given. Constrained minimization algorithms are used, with the design objectives in the constraint vector. The concept of Pareto optimiality is briefly reviewed. It is shown how an experienced designer can use it to find designs which are well-balanced in all objectives. Then the problem of finding designs which are insensitive to uncertainty in system parameters are discussed, introducing a probabilistic vector definition of sensitivity which is consistent with the deterministic Pareto optimal problem. Insensitivity is important in any practical design, but it is particularly important in the design of feedback control systems, since it is considered to be the most important distinctive property of feedback control. Methods of tradeoff between deterministic and stochastic-insensitive (SI) design are described, and tradeoff design results are presented for the example of the a Shuttle lateral stability augmentation system. This example is used because careful studies have been made of the uncertainty in Shuttle aerodynamics. Finally, since accurate statistics of uncertain parameters are usually not available, the effects of crude statistical models on SI designs are examined.

Schy, A. A.↗

Estimation of spectral emissivity in the thermal infrared

A number of algorithms are available in the literature that attempt to remove most of the effects of temperature from thermal multispectral data where the final goal is to extract emissivity differences. Early approaches include adjacent spectral band ratioing, broad band radiance normalization and the use of one band where emissivities are generally high (e.g., 11 to 12 micrometers) to determine the temperature. More recent work has produced two techniques that use data averaging to extract temperature to leave a quantity related to emissivity changes. These two techniques have been investigated and compared and appear to provide reasonable results. The analysis presented in this paper develops a thermal IR multispectral temperature/emissivity estimation procedure based on formal estimation theory, Gaussian statistics, and a stochastic radiance signal model including the effects of both temperature and emissivity. The importance of this work is that this is an optimal estimation procedure which will provide minimum variance estimates of temperature and emissivity changes directly. Section 2 discusses optimal linear spectral emissivity estimation and Section 3 is a summary.

Kryskowski, David↗

Health Monitoring for Airframe Structural Characterization

This study established requirements for structural health monitoring systems, identified and characterized a prototype structural sensor system, developed sensor interpretation algorithms, and demonstrated the sensor systems on operationally realistic test articles. Fiber-optic corrosion sensors (i.e., moisture and metal ion sensors) and low-cycle fatigue sensors (i.e., strain and acoustic emission sensors) were evaluated to validate their suitability for monitoring aging degradation; characterize the sensor performance in aircraft environments; and demonstrate placement processes and multiplexing schemes. In addition, a unique micromachined multimeasure and sensor concept was developed and demonstrated. The results show that structural degradation of aircraft materials could be effectively detected and characterized using available and emerging sensors. A key component of the structural health monitoring capability is the ability to interpret the information provided by sensor system in order to characterize the structural condition. Novel deterministic and stochastic fatigue damage development and growth models were developed for this program. These models enable real time characterization and assessment of structural fatigue damage.

Munns, Thomas E.↗

Martian Atmospheric Modeling of Scale Factors for MarsGRAM 2005 and the MAVEN Project

For spacecraft missions to Mars, especially the navigation of Martian orbiters and landers, an extensive knowledge of the Martian atmosphere is extremely important. The generally-accepted NASA standard for modeling (MarsGRAM), which was developed at Marshall Space Flight Center. MarsGRAM is useful for task such as aerobraking, performance analysis and operations planning for aerobraking, entry descent and landing, and aerocapture. Unfortunately, the densities for the Martian atmosphere in MarsGRAM are based on table look-up and not on an analytical algorithm. Also, these values can vary drastically from the densities actually experienced by the spacecraft. This does not have much of an impact on simple integrations but drastically affects its usefulness in other applications, especially those in navigation. For example, the navigation team for the Mars Atmosphere Volatile Environment (MAVEN) Project uses MarsGRAM to target the desired atmospheric density for the orbiter's pariapse passage, its closet approach to the planet. After the satellite's passage through pariapsis the computed density is compared to the MarsGRAM model and a scale factor is assigned to the model to account for the difference. Therefore, large variations in the atmosphere from the model can cause unexpected deviations from the spacecraft's planned trajectory. In order to account for this, an analytic stochastic model of the scale factor's behavior is desired. The development of this model will allow for the MAVEN navigation team to determine the probability of various Martian atmospheric variations and their effects on the spacecraft.

Martian atmosphere↗

A space-time stochastic model of rainfall for satellite remote-sensing studies

A model of the spatial and temporal distribution of rainfall is described that produces random spatial rainfall patterns with these characteristics: (1) the model is defined on a grid with each grid point representing the average rain rate over the surrounding grid box, (2) rain occurs at any one grid point, on average, a specified percentage of the time and has a lognormal probability distribution, (3) spatial correlation of the rainfall can be arbitrarily prescribed, and (4) time stepping is carried out so that large-scale features persist longer than small-scale features. Rain is generated in the model from the portion of a correlated Gaussian random field that exceeds a threshold. The portion of the field above the threshold is rescaled to have a lognormal probability distribution. Sample output of the model designed to mimic radar observations of rainfall during the Global Atmospheric Research Program Atlantic Tropical Experiment (GATE), is shown. The model is intended for use in evaluating sampling strategies for satellite remote-sensing of rainfall and for development of algorithms for converting radiant intensity received by an instrument from its field of view into rainfall amount.

Bell, Thomas L.↗

Autonomous frequency domain identification: Theory and experiment

The analysis, design, and on-orbit tuning of robust controllers require more information about the plant than simply a nominal estimate of the plant transfer function. Information is also required concerning the uncertainty in the nominal estimate, or more generally, the identification of a model set within which the true plant is known to lie. The identification methodology that was developed and experimentally demonstrated makes use of a simple but useful characterization of the model uncertainty based on the output error. This is a characterization of the additive uncertainty in the plant model, which has found considerable use in many robust control analysis and synthesis techniques. The identification process is initiated by a stochastic input u which is applied to the plant p giving rise to the output. Spectral estimation (h = P sub uy/P sub uu) is used as an estimate of p and the model order is estimated using the produce moment matrix (PMM) method. A parametric model unit direction vector p is then determined by curve fitting the spectral estimate to a rational transfer function. The additive uncertainty delta sub m = p - unit direction vector p is then estimated by the cross spectral estimate delta = P sub ue/P sub uu where e = y - unit direction vectory y is the output error, and unit direction vector y = unit direction vector pu is the computed output of the parametric model subjected to the actual input u. The experimental results demonstrate the curve fitting algorithm produces the reduced-order plant model which minimizes the additive uncertainty. The nominal transfer function estimate unit direction vector p and the estimate delta of the additive uncertainty delta sub m are subsequently available to be used for optimization of robust controller performance and stability.

Yam, Yeung↗

Filtering Image Records Using Wavelets and the Zakai Equation

Consider the problem of detecting and localizing a faint object moving In an "essentially stationary" background, using a sequence of two-dimensional low-SNR images of the scene. A natural approach consists of "digitizing" each snapshot into a discrete set of observations, sufficiently (perhaps not exactly) matched to the object In question, then tracking the object using an appropriate stochastic filter. The tracking would be expected to make up for the low signal-to-noise ratio, this allowing one to "coherently" process successive images in order to beat down the noise and localize the object. Thus, "tracking" here does not refer to the usual notion of detecting then tracking: rather, we track in order to detect The problem then becomes one of choosing the appropriate image representation as well as the optimal (and necessarily nonlinear filter. We propose exact and approximate solutions using wavelets and the Zakai equation. The smoothness of the wavelets used is required in the derivation of the evolution equation for the conditional density giving the filter, and their orthogonality makes it possible to carry out actual computations of the Ito- and change-of-gauge-terms in the algorithm effectively.

Haddad, Ziad S.↗

Estimating Random Errors Due to Shot Noise in Backscatter Lidar Observations

In this paper, we discuss the estimation of random errors due to shot noise in backscatter lidar observations that use either photomultiplier tube (PMT) or avalanche photodiode (APD) detectors. The statistical characteristics of photodetection are reviewed, and photon count distributions of solar background signals and laser backscatter signals are examined using airborne lidar observations at 532 nm using a photon-counting mode APD. Both distributions appear to be Poisson, indicating that the arrival at the photodetector of photons for these signals is a Poisson stochastic process. For Poisson-distributed signals, a proportional, one-to-one relationship is known to exist between the mean of a distribution and its variance. Although the multiplied photocurrent no longer follows a strict Poisson distribution in analog-mode APD and PMT detectors, the proportionality still exists between the mean and the variance of the multiplied photocurrent. We make use of this relationship by introducing the noise scale factor (NSF), which quantifies the constant of proportionality that exists between the root-mean-square of the random noise in a measurement and the square root of the mean signal. Using the NSF to estimate random errors in lidar measurements due to shot noise provides a significant advantage over the conventional error estimation techniques, in that with the NSF uncertainties can be reliably calculated from/for a single data sample. Methods for evaluating the NSF are presented. Algorithms to compute the NSF are developed for the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) lidar and tested using data from the Lidar In-space Technology Experiment (LITE). OCIS Codes:

Liu, Zhaoyan↗

POD Model Reconstruction for Gray-Box Fault Detection

Proper orthogonal decomposition (POD) is the mathematical basis of a method of constructing low-order mathematical models for the "gray-box" fault-detection algorithm that is a component of a diagnostic system known as beacon-based exception analysis for multi-missions (BEAM). POD has been successfully applied in reducing computational complexity by generating simple models that can be used for control and simulation for complex systems such as fluid flows. In the present application to BEAM, POD brings the same benefits to automated diagnosis. BEAM is a method of real-time or offline, automated diagnosis of a complex dynamic system.The gray-box approach makes it possible to utilize incomplete or approximate knowledge of the dynamics of the system that one seeks to diagnose. In the gray-box approach, a deterministic model of the system is used to filter a time series of system sensor data to remove the deterministic components of the time series from further examination. What is left after the filtering operation is a time series of residual quantities that represent the unknown (or at least unmodeled) aspects of the behavior of the system. Stochastic modeling techniques are then applied to the residual time series. The procedure for detecting abnormal behavior of the system then becomes one of looking for statistical differences between the residual time series and the predictions of the stochastic model.

Park, Han↗

Towards Comprehensive Variation Models for Designing Vehicle Monitoring Systems

When designing vehicle vibration monitoring systems for aerospace devices, it is common to use well-established models of vibration features to determine whether failures or defects exist. Most of the algorithms used for failure detection rely on these models to detect significant changes in a flight environment. In actual practice, however, most vehicle vibration monitoring systems are corrupted by high rates of false alarms and missed detections. This crucial roadblock makes their implementation in real vehicles (e.g., helicopter transmissions and aircraft engines) difficult, making their operation costly and unreliable. Research conducted at the NASA Ames Research Center has determined that a major reason for the high rates of false alarms and missed detections is the numerous sources of statistical variations that are not taken into account in the modeling assumptions. In this paper, we address one such source of variations, namely, those caused during the design and manufacturing of rotating machinery components that make up aerospace systems. We present a novel way of modeling the vibration response by including design variations via probabilistic methods. Using such models, we develop a methodology to account for design and manufacturing variations, and explore the changes in the vibration response to determine its stochastic nature. We explore the potential of the methodology using a nonlinear cam-follower model, where the spring stiffness values are assumed to follow a normal distribution. The results demonstrate initial feasibility of the method, showing great promise in developing a general methodology for designing more accurate aerospace vehicle monitoring systems.

McAdams, Daniel A.↗

Probabilistic Path Planning of Montgolfier Balloons in Strong, Uncertain Wind Fields

Lighter-than-air vehicles such as hot-air balloons have been proposed for exploring Saturn s moon Titan, as well as other bodies with significant atmospheres. For these vehicles to navigate effectively, it is critical to incorporate the effects of surrounding wind fields, especially as these winds will likely be strong relative to the control authority of the vehicle. Predictive models of these wind fields are available, and previous research has considered problems of planning paths subject to these predicted forces. However, such previous work has considered the wind fields as known a priori, whereas in practical applications, the actual wind vector field is not known exactly and may deviate significantly from the wind velocities estimated by the model. A probabilistic 3D path-planning algorithm was developed for balloons to use uncertain wind models to generate time-efficient paths. The nominal goal of the algorithm is to determine what altitude and what horizontal actuation, if any is available on the vehicle, to use to reach a particular goal location in the least expected time, utilizing advantageous winds. The solution also enables one to quickly evaluate the expected time-to-goal from any other location and to avoid regions of large uncertainty. This method is designed for balloons in wind fields but may be generalized for any buoyant vehicle operating in a vector field. To prepare the planning problem, the uncertainty in the wind field is modeled. Then, the problem of reaching a particular goal location is formulated as a Markov decision process (MDP) using a discretized space approach. Solving the MDP provides a policy of what actuation option (how much buoyancy change and, if applicable, horizontal actuation) should be selected at any given location to minimize the expected time-to-goal. The results provide expected time-to-goal values from any given location on the globe in addition to the action policy. This stochastic approach can also provide insights not accessible by deterministic methods; for example, one can evaluate variability and risk associated with different scenarios, rather than only viewing the expected outcome.

Wolf, Michael↗

Growth Control and Disease Mechanisms in Computational Embryogeny

This paper presents novel approach to applying growth control and diseases mechanisms in computational embryogeny. Our method, which mimics fundamental processes from biology, enables individuals to reach maturity in a controlled process through a stochastic environment. Three different mechanisms were implemented; disease mechanisms, gene suppression, and thermodynamic balancing. This approach was integrated as part of a structural evolutionary model. The model evolved continuum 3-D structures which support an external load. By using these mechanisms we were able to evolve individuals that reached a fixed size limit through the growth process. The growth process was an integral part of the complete development process. The size of the individuals was determined purely by the evolutionary process where different individuals matured to different sizes. Individuals which evolved with these characteristics have been found to be very robust for supporting a wide range of external loads.

Genetic algorithm↗