SEARCH · Search NASA
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.
This report summarizes the research done during 1991/92 under the Center for Aeronautics and Space Information Science (CASIS) program. The topics covered are computer architecture, networking, and neural nets.
Soft computing is a general term for algorithms that learn from human knowledge and mimic human skills. Example of such algorithms are fuzzy inference systems and neural networks. Many applications, especially in control engineering, have demonstrated their appropriateness in building intelligent systems that are flexible and robust. Although recent research have shown that certain class of neuro-fuzzy controllers can be proven bounded and stable, they are implementation dependent and difficult to apply to the design and validation process. Many practitioners adopt the trial and error approach for system validation or resort to exhaustive testing using prototypes. In this paper, we describe our on-going research towards establishing necessary theoretic foundation as well as building practical tools for the verification and validation of soft-computing systems. A unified model for general neuro-fuzzy system is adopted. Classic non-linear system control theory and recent results of its applications to neuro-fuzzy systems are incorporated and applied to the unified model. It is hoped that general tools can be developed to help the designer to visualize and manipulate the regions of stability and boundedness, much the same way Bode plots and Root locus plots have helped conventional control design and validation.
Many nondestructive methods exist for the detection of localized material anomalies in an otherwise good composite structure. The problem arises when the material system as a whole has degraded during service or was improperly manufactured. Porosity and intra-ply microcracking are two such conditions that in unlined composite pressure vessels can be very troublesome to detect and when linked through the thickness can be critical to mission success. These leak paths may lead to loss of pressure/propellant, increased risk of explosion and possible cryo-pumping. Research sought nondestructive methods for quantifying porosity and microcracking in composite tankage. Both thermographic and resonance ultrasound methods have been utilized with artificial neural network and statistical approaches to analyze the data. Resonant ultrasound spectroscopy provides measurements, which are sensitive to fine details in the materials character, such as micro-cracking and porosity. Here, the higher frequency (shorter wavelength) components of the signal train provide more significant interaction with the defects causing the spectral characteristics to shift toward lower amplitudes at the higher frequencies. As the density of the defects increases more interactions occur and more drastic amplitude changes are observed. From a thermal perspective, the higher the defect density the lower the through thickness thermal diffusivity will be. Utilizing a point heat source, and thermographically recording the heat profile with time, diffusivity calculations can be made which in turn can be related to the relative quality of the material. Preliminary experiments to verify the measurable effect on the resonance spectrum of the ultrasonic data to detect microcracking and for porosity detection thermographically are presented. Methods involving supervised and unsupervised artificial neural networks as well as other clustering algorithms are developed for signal identification.
This report summarizes results in conduct of research sponsored by the 2006 Independent Research and Development (IR&D) program at Marshall Space Flight Center (MSFC) at Redstone Arsenal, Alabama. The focus of this IR&D is neural network (NN) technology provided by Imagination Engines, Incorporated (IEI) of St. Louis, Missouri. The technology already has many commercial, military, and governmental applications, and a rapidly growing list of other potential spin-offs. The goal for this IR&D is implementation and demonstration of the technology for autonomous robotic operations, first in software and ultimately in one or more hardware realizations. Testing is targeted specifically to the MSFC Flat Floor, but may also include other robotic platforms at MSFC, as time and funds permit. For the purpose of this report, the NN technology will be referred to by IEI's designation for a subset configuration of its patented technology suite: Self-Training Autonomous Neural Network Object (STANNO).
A key challenge in designing the new High Speed Civil Transport (HSCT) aircraft is determining a good match between the airframe and engine. Multidisciplinary design optimization can be used to solve the problem by adjusting parameters of both the engine and the airframe. Earlier, an example problem was presented of an HSCT aircraft with four mixed-flow turbofan engines and a baseline mission to carry 305 passengers 5000 nautical miles at a cruise speed of Mach 2.4. The problem was solved by coupling NASA Lewis Research Center's design optimization testbed (COMETBOARDS) with NASA Langley Research Center's Flight Optimization System (FLOPS). The computing time expended in solving the problem was substantial, and the instability of the FLOPS analyzer at certain design points caused difficulties. In an attempt to alleviate both of these limitations, we explored the use of two approximation concepts in the design optimization process. The two concepts, which are based on neural network and linear regression approximation, provide the reanalysis capability and design sensitivity analysis information required for the optimization process. The HSCT aircraft optimization problem was solved by using three alternate approaches; that is, the original FLOPS analyzer and two approximate (derived) analyzers. The approximate analyzers were calibrated and used in three different ranges of the design variables; narrow (interpolated), standard, and wide (extrapolated).
Computational intelligence (CI) is a rapidly evolving field that utilizes life imitating metaphors for guiding model building including, but not limited to neural networks, fuzzy logic, genetic algorithms, artificial life, and hybrid CI paradigms. Although the boundaries between artificial intelligence (AI) and CI are not distinct, their research communities are separate and distinct. CI researchers tend to focus on processing numerical data from sensors, while the AI community generally relies on symbolic computing to capture human knowledge. In both areas, there is a great deal of interest and activity in hybrid systems that can offset the limitations of individual methods, extend their capabilities, and create new capabilities. Examples of the benefits that can accrue from hybrid systems are contained.
The objective of the Gross Motion Control project is to study alternative control approaches that will provide payload invariant high speed trajectory tracking for nonrepetitive motions in free space. The research has concentrated on modifications to the model-based control structure. Development and evaluation is being actively pursued of both adaptive primary (inner loop) and robust secondary (output loop) controllers. In-house developments are compared and contrasted to the techniques proposed by other researchers. The case study for the evaluation is the first three links of a PUMA-560. Incorporating the principals of multiple model adaptive estimation, artificial neural networks, and Lyapunov theory into the model based paradigm has shown the potential for enhanced tracking. Secondary controllers based on Quantitative Feedback Theory, or augmented with auxiliary inputs, significantly improve the robustness to payload variations and unmodeled drive system dynamics. An overview is presented of the different concepts under investigation and a sample is provided of the latest experimental results.
We describe the PACE-MAPP algorithm that simultaneously retrieves aerosol and ocean optical parameters using multiangle and multi-channel polarimeter measurements from the SPEXone, Hyper-Angular Rainbow Polarimeter 2 (HARP2), and Ocean Color Instrument (OCI) instruments onboard the NASA Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) observing system PACE-MAPP is adapted from the Research Scanning Polarimeter (RSP) Microphysical Aerosol Properties from Polarimetry (RSP-MAPP) algorithm. A key feature of the MAPP family of algorithms is the use of a coupled vector radiative transfer model such that the atmosphere and ocean are always considered together as one system. Consequently, conservation of energy ensures that negative water-leaving radiances do not occur. PACE-MAPP uses optimal estimation to simultaneously characterize the optical and microphysical properties of aerosol and ocean constituents, find the optimal solution, and reliably account for the uncertainties of each parameter. This coupled approach, together with multiangle, multi-channel polarimeter measurements, will enable retrievals of aerosol and water properties across the Earth’s oceans. The PACE-MAPP algorithm provides aerosol and ocean products for both the open ocean and coastal areas and is designed to be accurate, modular, and efficient by using fast neural networks that replace the time-consuming vector radiative transfer calculations. We provide an overview of the PACE-MAPP framework and also describe its modular components including its aerosol and hydrosol models, ocean bio-optical models, and thin cirrus model.
This paper describes the development of a planned approach for Autonomous operation of an Unmanned Aerial Vehicle (UAV). A Hybrid approach will seek to provide Knowledge Generation through the application of Artificial Intelligence (AI) and Intelligent Agents (IA) for UAV control. The applications of several different types of AI techniques for flight are explored during this research effort. The research concentration is directed to the application of different AI methods within the UAV arena. By evaluating AI and biological system approaches. which include Expert Systems, Neural Networks. Intelligent Agents, Fuzzy Logic, and Complex Adaptive Systems, a new insight may be gained into the benefits of AI and CAS techniques applied to achieving true autonomous operation of these systems. Although flight systems were explored, the benefits should apply to many Unmanned Vehicles such as: Rovers. Ocean Explorers, Robots, and autonomous operation systems. A portion of the flight system is broken down into control agents that represent the intelligent agent approach used in AI. After the completion of a successful approach, a framework for applying an intelligent agent is presented. The initial results from simulation of a security agent for communication are presented.
We describe a neural network based robotic system. Unlike traditional robotic systems, our approach focussed on non-stationary problems. We indicate that self-organization capability is necessary for any system to operate successfully in a non-stationary environment. We suggest that self-organization should be based on an active exploration process. We investigated neural architectures having novelty sensitivity, selective attention, reinforcement learning, habit formation, flexible criteria categorization properties and analyzed the resulting behavior (consisting of an intelligent initiation of exploration) by computer simulations. While various computer vision researchers acknowledged recently the importance of active processes (Swain and Stricker, 1991), the proposed approaches within the new framework still suffer from a lack of self-organization (Aloimonos and Bandyopadhyay, 1987; Bajcsy, 1988). A self-organizing, neural network based robot (MAVIN) has been recently proposed (Baloch and Waxman, 1991). This robot has the capability of position, size rotation invariant pattern categorization, recognition and pavlovian conditioning. Our robot does not have initially invariant processing properties. The reason for this is the emphasis we put on active exploration. We maintain the point of view that such invariant properties emerge from an internalization of exploratory sensory-motor activity. Rather than coding the equilibria of such mental capabilities, we are seeking to capture its dynamics to understand on the one hand how the emergence of such invariances is possible and on the other hand the dynamics that lead to these invariances. The second point is crucial for an adaptive robot to acquire new invariances in non-stationary environments, as demonstrated by the inverting glass experiments of Helmholtz. We will introduce Pavlovian conditioning circuits in our future work for the precise objective of achieving the generation, coordination, and internalization of sequence of actions.
Oscillation of electrical activity has been found in many nervous systems, from invertebrates to vertebrates including man. There exists experimental evidence of very simple circuits with the capability of oscillation. Neurons with intrinsic oscillation have been found and also neural circuits where oscillation is a property of the network. These two types of oscillations coexist in many instances. It is nowadays hypothesized that behind synchronization and oscillation there is a system of coupled oscillators responsible for activities that range from locomotion and feature binding in vision to control of sleep and circadian rhythms. The huge knowledge that has been acquired on oscillators from the times of Lord Rayleigh has made the simulation of neural oscillators a very active endeavor. This has been enhanced with more recent physiological findings about small neural circuits by means of intracellular and extracellular recordings as well as imaging methods. The future of this interdisciplinary field looks very promising; some researchers are going into quantum mechanics with the idea of trying to provide a quantum description of the brain. In this work we describe some simulations using neuron models by means of which we form simple neural networks that have the capability of oscillation. We analyze the oscillatory activity with root locus method, cross-correlation histograms, and phase planes. In the more complicated neural network models there is the possibility of chaotic oscillatory activity and we study that by means of Lyapunov exponents. The companion paper shows an example of that kind.
Environmental management, impact assessment, research and monitoring are multidisciplinary activities which are ideally suited to incorporate a multi-media approach to environmental problem solving. Geographic information systems (GIS), simulation models, neural networks and expert-system software are some of the advancing technologies being used for data management, query, analysis and display. At the 140,000 acre John F. Kennedy Space Center, the Advanced Software Technology group has been supporting development and implementation of a program that integrates these and other rapidly evolving hardware and software capabilities into a comprehensive Mapping, Analysis and Planning System (MAPS) based in a workstation/local are network environment. An expert-system shell is being developed to link the various databases to guide users through the numerous stages of a facility siting and environmental assessment. The expert-system shell approach is appealing for its ease of data access by management-level decision makers while maintaining the involvement of the data specialists. This, as well as increased efficiency and accuracy in data analysis and report preparation, can benefit any organization involved in natural resources management.
Researchers began studying automated agents that learn to perform multiple-step tasks early in the history of artificial intelligence (Samuel, 1963; Samuel, 1967; Waterman, 1970; Fikes, Hart & Nilsonn, 1972). Multiple-step tasks are tasks that can only be solved via a sequence of decisions, such as control problems, robotics problems, classic problem-solving, and game-playing. The objective of agents attempting to learn such tasks is to use the resources they have available in order to become more proficient at the tasks. In particular, each agent attempts to develop a good policy, a mapping from states to actions, that allows it to select actions that optimize a measure of its performance on the task; for example, reducing the number of steps necessary to complete the task successfully. Our study focuses on reinforcement learning, a set of learning techniques where the learner performs trial-and-error experiments in the task and adapts its policy based on the outcome of those experiments. Much of the work in reinforcement learning has focused on a particular, simple representation, where every problem state is represented explicitly in a table, and associated with each state are the actions that can be chosen in that state. A major advantage of this table lookup representation is that one can prove that certain reinforcement learning techniques will develop an optimal policy for the current task. The drawback is that the representation limits the application of reinforcement learning to multiple-step tasks with relatively small state-spaces. There has been a little theoretical work that proves that convergence to optimal solutions can be obtained when using generalization structures, but the structures are quite simple. The theory says little about complex structures, such as multi-layer, feedforward artificial neural networks (Rumelhart & McClelland, 1986), but empirical results indicate that the use of reinforcement learning with such structures is promising. These empirical results make no theoretical claims, nor compare the policies produced to optimal policies. A goal of our work is to be able to make the comparison between an optimal policy and one stored in an artificial neural network. A difficulty of performing such a study is finding a multiple-step task that is small enough that one can find an optimal policy using table lookup, yet large enough that, for practical purposes, an artificial neural network is really required. We have identified a limited form of the game OTHELLO as satisfying these requirements. The work we report here is in the very preliminary stages of research, but this paper provides background for the problem being studied and a description of our initial approach to examining the problem. In the remainder of this paper, we first describe reinforcement learning in more detail. Next, we present the game OTHELLO. Finally we argue that a restricted form of the game meets the requirements of our study, and describe our preliminary approach to finding an optimal solution to the problem.
Research in cognitive science suggests that much of cognition involves the rapid manipulation of complex data structures. However, it is very unclear how this could be realized in neural networks or connectionist systems. A core question is: how could the interconnectivity of items in an abstract-level data structure be neurally encoded? The answer appeals mainly to positional relationships between activity patterns within neural arrays, rather than directly to neural connections in the traditional way. The new method was initially devised to account for abstract symbolic data structures, but it also supports cognitively useful spatial analogue, image-like representations. As the neural model is based on massive, uniform, parallel computations over 2D arrays, the massively parallel processor is a convenient tool for simulation work, although there are complications in using the machine to the fullest advantage. An MPP Pascal simulation program for a small pilot version of the model is running.
The present work is intended to give technologists, research scientists, and mathematicians a graduate-level overview of the field of neurocomputing. After exploring the relationship of this field to general neuroscience, attention is given to neural network building blocks, the self-adaptation equations of learning laws, the data-transformation structures of associative networks, and the multilayer data-transformation structures of mapping networks. Also treated are the neurocomputing frontiers of spatiotemporal, stochastic, and hierarchical networks, 'neurosoftware', the creation of neural network-based computers, and neurocomputing applications in sensor processing, control, and data analysis.
The primary goal of the Adaptive Vision Laboratory Research project was to develop advanced computer vision systems for automatic target recognition. The approach used in this effort combined several machine learning paradigms including evolutionary learning algorithms, neural networks, and adaptive clustering techniques to develop the E-MOR.PH system. This system is capable of generating pattern recognition systems to solve a wide variety of complex recognition tasks. A series of simulation experiments were conducted using E-MORPH to solve problems in OCR, military target recognition, industrial inspection, and medical image analysis. The bulk of the funds provided through this grant were used to purchase computer hardware and software to support these computationally intensive simulations. The payoff from this effort is the reduced need for human involvement in the design and implementation of recognition systems. We have shown that the techniques used in E-MORPH are generic and readily transition to other problem domains. Specifically, E-MORPH is multi-phase evolutionary leaming system that evolves cooperative sets of features detectors and combines their response using an adaptive classifier to form a complete pattern recognition system. The system can operate on binary or grayscale images. In our most recent experiments, we used multi-resolution images that are formed by applying a Gabor wavelet transform to a set of grayscale input images. To begin the leaming process, candidate chips are extracted from the multi-resolution images to form a training set and a test set. A population of detector sets is randomly initialized to start the evolutionary process. Using a combination of evolutionary programming and genetic algorithms, the feature detectors are enhanced to solve a recognition problem. The design of E-MORPH and recognition results for a complex problem in medical image analysis are described at the end of this report. The specific task involves the identification of vertebrae in x-ray images of human spinal columns. This problem is extremely challenging because the individual vertebra exhibit variation in shape, scale, orientation, and contrast. E-MORPH generated several accurate recognition systems to solve this task. This dual use of this ATR technology clearly demonstrates the flexibility and power of our approach.
This article describes methods for online model-based diagnosis of subsystems of the advanced life support system (ALS). The diagnosis methodology is tailored to detect, isolate, and identify faults in components of the system quickly so that fault-adaptive control techniques can be applied to maintain system operation without interruption. We describe the components of our hybrid modeling scheme and the diagnosis methodology, and then demonstrate the effectiveness of this methodology by building a detailed model of the reverse osmosis (RO) system of the water recovery system (WRS) of the ALS. This model is validated with real data collected from an experimental testbed at NASA JSC. A number of diagnosis experiments run on simulated faulty data are presented and the results are discussed.
The work described in this research program originated with the choice by NASA of an ion mobility spectrometer for air quality monitoring on-board the international spacestation. Though the gas chromatograph-ion mobility spectrometer analyzer known as VOA met or exceeded expectations, limitations in the basic understanding of response and the utilization of foundational principles into usable technology was considered unacceptable. In this research program, a comprehensive model for the origins of mobility spectra was proposed, tested and verified. The principles considered responsible for the appearance of mobility spectra have now been elucidated through this project. This understanding has been applied in automated identification of mobility spectra using neural networks and routine procedures for this now exist. Finally, the limitation on linear range has been shown to be a technical limitation and not a fundamental limitation so that a hardware component was crafted to extend the linear range of a mobility spectrometer by 10X. This project has led to one Ph.D. dissertation and one MS thesis. In addition, over ten public presentations at professional meetings and six journal publications have resulted from this program of research. The findings are so plentiful that total analysis of the findings may require four to six years or more. The findings confirm that the decision to use VOA was sound and that the chemical and physical principles of mobility spectrometry are both understandable and predictable.