Search NASA⌕ Search

SEARCH · Search NASA

Results for “Learning with errors”

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 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
NASA NTRS1995

Fast temporal neural learning using teacher forcing

A neural network is trained to output a time dependent target vector defined over a predetermined time interval in response to a time dependent input vector defined over the same time interval by applying corresponding elements of the error vector, or difference between the target vector and the actual neuron output vector, to the inputs of corresponding output neurons of the network as corrective feedback. This feedback decreases the error and quickens the learning process, so that a much smaller number of training cycles are required to complete the learning process. A conventional gradient descent algorithm is employed to update the neural network parameters at the end of the predetermined time interval. The foregoing process is repeated in repetitive cycles until the actual output vector corresponds to the target vector. In the preferred embodiment, as the overall error of the neural network output decreasing during successive training cycles, the portion of the error fed back to the output neurons is decreased accordingly, allowing the network to learn with greater freedom from teacher forcing as the network parameters converge to their optimum values. The invention may also be used to train a neural network with stationary training and target vectors.

Toomarian, Nikzad↗
NASA NTRS1993

A proposal of an architecture for the coordination level of intelligent machines

The issue of obtaining a practical, structured, and detailed description of an architecture for the Coordination Level of Center for Intelligent Robotic Systems for Sapce Exploration (CIRSSE) Testbed Intelligent Controller is addressed. Previous theoretical and implementation works were the departure point for the discussion. The document is organized as follows: after this introductory section, section 2 summarizes the overall view of the Intelligent Machine (IM) as a control system, proposing a performance measure on which to base its design. Section 3 addresses with some detail implementation issues. An hierarchic petri-net with feedback-based learning capabilities is proposed. Finally, section 4 is an attempt to address the feedback problem. Feedback is used for two functions: error recovery and reinforcement learning of the correct translations for the petri-net transitions.

Beard, Randall↗
NASA NTRS2006

Assessment of Postflight Locomotor Performance Utilizing a Test of Functional Mobility: Strategic and Adaptive Responses

Space flight induces adaptive modification in sensorimotor function, allowing crewmembers to operate in the unique microgravity environment. This adaptive state, however, is inappropriate for a terrestrial environment. During a re-adaptation period upon their return to Earth, crewmembers experience alterations in sensorimotor function, causing various disturbances in perception, spatial orientation, posture, gait, and eye-head coordination. Following long duration space flight, sensorimotor dysfunction would prevent or extend the time required to make an emergency egress from the vehicle; compromising crew safety and mission objectives. We are investigating two types of motor learning that may interact with each other and influence a crewmember's ability to re-adapt to Earth's gravity environment. In strategic learning, crewmembers make rapid modifications in their motor control strategy emphasizing error reduction. This type of learning may be critical during the first minutes and hours after landing. In adaptive learning, long-term plastic transformations occur, involving morphological changes and synaptic modification. In recent literature these two behavioral components have been associated with separate brain structures that control the execution of motor strategies: the strategic component was linked to the posterior parietal cortex and the adaptive component was linked to the cerebellum (Pisella, et al. 2004). The goal of this paper was to demonstrate the relative contributions of the strategic and adaptive components to the re-adaptation process in locomotor control after long duration space flight missions on the International Space Station (ISS). The Functional Mobility Test (FMT) was developed to assess crewmember s ability to ambulate postflight from an operational and functional perspective. Sixteen crewmembers were tested preflight (3 sessions) and postflight (days 1, 2, 4, 7, 25) following a long duration space flight (approx 6 months) on the ISS. We have further analyzed the FMT data to characterize strategic and adaptive components during the postflight readaptation period. Crewmembers walked at a preferred pace through an obstacle course set up on a base of 10 cm thick medium density foam (Sunmate Foam, Dynamic Systems, Inc., Leicester, NC). The 6.0m X 4.0m course consisted of several pylons made of foam; a Styrofoam barrier 46.0cm high that crewmembers stepped over; and a portal constructed of two Styrofoam blocks, each 31cm high, with a horizontal bar covered by foam and suspended from the ceiling which was adjusted to the height of the crewmember s shoulder. The portal required crewmembers to bend at the waist and step over a barrier simultaneously. All obstacles were lightweight, soft and easily knocked over. Crewmembers were instructed to walk through the course as quickly and as safely as possible without touching any of the objects on the course. This task was performed three times in the clockwise direction and three times in the counterclockwise direction that was randomly chosen. The dependent measures for each trial were: time to complete the course (seconds) and the number of obstacles touched or knocked down. For each crewmember, the time to complete each FMT trial from postflight days 1, 2, 4, 7 and 25 were further analyzed. A single logarithmic curve using a least squares calculation was fit through these data to produce a single comprehensive curve (macro). This macro curve composed of data spanning 25 days, illustrates the re-adaptive learning function over the longer time scale term. Additionally, logarithmic curves were fit to the 6 data trials within each individual post flight test day to produce 5 separate daily curves. These micro curves, produced from data obtained over the course of minutes, illustrates the strategic learning function exhibited over a relative shorter time scale. The macro curve for all subjects exhibited adaptive motor learning patterns over the 25 day period. Howev, 9/16 crewmembers exhibited significant strategic motor learning patterns in their micro curves, as defined by m > 1 in the equation of the line y=m*LN(x) +b. These data indicate that postflight recovery in locomotor function involves both strategic and adaptive mechanisms. Future countermeasures will be designed to enhance both recovery processes.

Warren, L. E.↗
NASA NTRSDate not supplied

Towards Autonomous Lunar Resource Excavation via Reinforcement Learning

To continue on a sustainable and flexible path, NASA needs to address the challenge of collecting and moving large amounts of regolith at the destination. NASA’s Regolith Advanced Surface Systems Operations Robot (RASSOR) is principally designed to mine and deliver regolith for In-Situ Resource Utilization (ISRU) processing. RASSOR’s design enables it to efficiently collect and deposit regolith, return collected material for processing, and myriad related ISRU activities. To reliably perform these operations on the lunar surface, RASSOR software and sensory systems need to be robust and maximize the information extracted from a reduced sensor payload. Herein, we present preliminary findings from the Intelligent Capabilities Enhanced RASSOR project. We created reduced-order simulation environments to develop autonomous trenching controllers via reinforcement learning and prototype state estimation architectures. The goal of reinforcement learning is for an agent to learn a policy (task strategy) through interactions with an environment. When the agent performs an action, a change occurs in environment state and a numerical reward is received which informs the agent whether the action performed was good or not. Since reinforcement learning algorithms learn through trial-and-error, a simulation is a desirable first environment for development and learning. We developed two simulations, the first is a 2D excavation simulation developed to facilitate parameter selection, and a 3D simulation developed using a game physics engine, to simulate simplified soil interactions and increase the fidelity of the dynamic models of the robotic agents. The development of this 3D simulation has enabled the training of additional sensing capabilities and research both at the granular mechanics and operations levels. We experimented with various virtual sensor payloads to identify a combination that enabled efficient excavation operation and learning. Our reward function is based on how much material is excavated per step. A penalty is also received for leaving the dig site and to smooth the acceleration of the drum arms. We implemented pseudo time-of-flight sensors to report distance from each drum to ground and the height above ground which was found to be more efficient than existing solutions. Our findings suggest that reinforcement learning for autonomous operations has learned viable trenching strategies within 3000 training episodes in our simplified 2D environment and helped identify desirable sensing capabilities, arrangements, and considerations such as the positioning of time-of-flight sensors. Future work includes expanding our simulation to more complex environments and scenarios, and transfer learning from simulation to RASSOR 2.0 hardware for deployment in the Regolith Test Bin at NASA's Kennedy Space Center.

rassor↗
NASA NTRS1997

Using APEX to Model Anticipated Human Error: Analysis of a GPS Navigational Aid

The interface development process can be dramatically improved by predicting design facilitated human error at an early stage in the design process. The approach we advocate is to SIMULATE the behavior of a human agent carrying out tasks with a well-specified user interface, ANALYZE the simulation for instances of human error, and then REFINE the interface or protocol to minimize predicted error. This approach, incorporated into the APEX modeling architecture, differs from past approaches to human simulation in Its emphasis on error rather than e.g. learning rate or speed of response. The APEX model consists of two major components: (1) a powerful action selection component capable of simulating behavior in complex, multiple-task environments; and (2) a resource architecture which constrains cognitive, perceptual, and motor capabilities to within empirically demonstrated limits. The model mimics human errors arising from interactions between limited human resources and elements of the computer interface whose design falls to anticipate those limits. We analyze the design of a hand-held Global Positioning System (GPS) device used for radical and navigational decisions in small yacht recalls. The analysis demonstrates how human system modeling can be an effective design aid, helping to accelerate the process of refining a product (or procedure).

VanSelst, Mark↗
NASA NTRS1993

Learning class descriptions from a data base of spectral reflectance of soil samples

Consideration is given to a program developed to learn class descriptions from positive and negative training examples of spectral reflectance data of bare soils. It is a combination of 'learning by example' and the generate-and-test paradigm and is designed to provide a robust learning environment that can handle error-prone data. The program was tested by having it learn class descriptions of various categories of organic carbon content, iron oxide content, and particle size distribution in soils. These class descriptions were then used to classify an array of targets. The program found the sequence of relationships between bands that contained the most important information to distinguish the classes. Physical explanations for the class descriptions obtained are presented.

Kimes, D. S.↗
NASA NTRS2003

Neural Flight Control System

The Neural Flight Control System (NFCS) was developed to address the need for control systems that can be produced and tested at lower cost, easily adapted to prototype vehicles and for flight systems that can accommodate damaged control surfaces or changes to aircraft stability and control characteristics resulting from failures or accidents. NFCS utilizes on a neural network-based flight control algorithm which automatically compensates for a broad spectrum of unanticipated damage or failures of an aircraft in flight. Pilot stick and rudder pedal inputs are fed into a reference model which produces pitch, roll and yaw rate commands. The reference model frequencies and gains can be set to provide handling quality characteristics suitable for the aircraft of interest. The rate commands are used in conjunction with estimates of the aircraft s stability and control (S&C) derivatives by a simplified Dynamic Inverse controller to produce virtual elevator, aileron and rudder commands. These virtual surface deflection commands are optimally distributed across the aircraft s available control surfaces using linear programming theory. Sensor data is compared with the reference model rate commands to produce an error signal. A Proportional/Integral (PI) error controller "winds up" on the error signal and adds an augmented command to the reference model output with the effect of zeroing the error signal. In order to provide more consistent handling qualities for the pilot, neural networks learn the behavior of the error controller and add in the augmented command before the integrator winds up. In the case of damage sufficient to affect the handling qualities of the aircraft, an Adaptive Critic is utilized to reduce the reference model frequencies and gains to stay within a flyable envelope of the aircraft.

Gundy-Burlet, Karen↗
NASA NTRS1992

Learning class descriptions from a data base of spectral reflectance with multiple view angles

A learning program has been developed which combines 'learning by example' with the generate-and-test paradigm to furnish a robust learning environment capable of handling error-prone data. The problem is shown to be capable of learning class descriptions from positive and negative training examples of spectral and directional reflectance data taken from soil and vegetation. The program, which used AI techniques to automate very tedious processes, found the sequence of relationships that contained the most important information which could distinguish the classes.

Kimes, Daniel S.↗
NASA NTRS1989

Learning control for slewing of a flexible panel

This paper studies the applicability of a discrete-time learning control method to the slewing control of a large flexible panel. Among the issues discussed are feasibility of the desired trajectories and their specification schemes, learning control by linear feedback, limitation of the actuator device output, and robustness to parameter changes or modeling errors associated with the chosen learning control design. To demonstrate the effectiveness of learning control, the system is designed with a proportional controller of the base angle only. Then application of the learning control is shown to make the system learn quickly to achieve the desired slewing without residual vibration at the end of the maneuver. Simulation results are reported and discussed.

Phan, M.↗
NASA NTRS1992

Indirect decentralized learning control

The new field of learning control develops controllers that learn to improve their performance at executing a given task, based on experience performing this specific task. In a previous work, the authors presented a theory of indirect learning control based on use of indirect adaptive control concepts employing simultaneous identification and control. This paper develops improved indirect learning control algorithms, and studies the use of such controllers in decentralized systems. The original motivation of the learning control field was learning in robots doing repetitive tasks such as on an assembly line. This paper starts with decentralized discrete time systems, and progresses to the robot application, modeling the robot as a time varying linear system in the neighborhood of the nominal trajectory, and using the usual robot controllers that are decentralized, treating each link as if it is independent of any coupling with other links. The basic result of the paper is to show that stability of the indirect learning controllers for all subsystems when the coupling between subsystems is turned off, assures convergence to zero tracking error of the decentralized indirect learning control of the coupled system, provided that the sample time in the digital learning controller is sufficiently short.

Longman, Richard W.↗
NASA NTRS1988

A mathematical theory of learning control for linear discrete multivariable systems

When tracking control systems are used in repetitive operations such as robots in various manufacturing processes, the controller will make the same errors repeatedly. Here consideration is given to learning controllers that look at the tracking errors in each repetition of the process and adjust the control to decrease these errors in the next repetition. A general formalism is developed for learning control of discrete-time (time-varying or time-invariant) linear multivariable systems. Methods of specifying a desired trajectory (such that the trajectory can actually be performed by the discrete system) are discussed, and learning controllers are developed. Stability criteria are obtained which are relatively easy to use to insure convergence of the learning process, and proper gain settings are discussed in light of measurement noise and system uncertainties.

Phan, Minh↗
NASA NTRS1972

A robot conditioned reflex system modeled after the cerebellum.

Reduction of a theory of cerebellar function to computer software for the control of a mechanical manipulator. This reduction is achieved by considering the cerebellum, along with the higher-level brain centers which control it, as a type of finite-state machine with input entering the cerebellum via mossy fibers from the periphery and output from the cerebellum occurring via Purkinje cells. It is hypothesized that the cerebellum learns by an error-correction system similar to Perceptron training algorithms. An electromechanical model of the cerebellum is then developed for the control of a mechanical arm. The problem of modeling the granular layer which selects the set of parallel fibers which are active at any instant of time is considered, and a relevance matrix is constructed to model the relative degree of influence which mossy fibers from the various joints have on the sets of granule cells unique to each joint.

Albus, J. S.↗
NASA NTRS1993

Pattern classifier for health monitoring of helicopter gearboxes

The application of a newly developed diagnostic method to a helicopter gearbox is demonstrated. This method is a pattern classifier which uses a multi-valued influence matrix (MVIM) as its diagnostic model. The method benefits from a fast learning algorithm, based on error feedback, that enables it to estimate gearbox health from a small set of measurement-fault data. The MVIM method can also assess the diagnosability of the system and variability of the fault signatures as the basis to improve fault signatures. This method was tested on vibration signals reflecting various faults in an OH-58A main rotor transmission gearbox. The vibration signals were then digitized and processed by a vibration signal analyzer to enhance and extract various features of the vibration data. The parameters obtained from this analyzer were utilized to train and test the performance of the MVIM method in both detection and diagnosis. The results indicate that the MVIM method provided excellent detection results when the full range of faults effects on the measurements were included in training, and it had a correct diagnostic rate of 95 percent when the faults were included in training.

Chin, Hsinyung↗
NASA NTRS1990

Electronic Neural Networks

Viewgraphs on electronic neural networks for space station are presented. Topics covered include: electronic neural networks; electronic implementations; VLSI/thin film hybrid hardware for neurocomputing; computations with analog parallel processing; features of neuroprocessors; applications of neuroprocessors; neural network hardware for terrain trafficability determination; a dedicated processor for path planning; neural network system interface; neural network for robotic control; error backpropagation algorithm for learning; resource allocation matrix; global optimization neuroprocessor; and electrically programmable read only thin-film synaptic array.

Thakoor, Anil↗
NASA NTRS2001

Resource Selection Using Execution and Queue Wait Time Predictions

We developed techniques to predict application execution times for instance-based learning with an average error of 33% of average run time. We developed techniques to predict queue wait times that included a simulation of scheduling algorithms and execution time predictions. We implemented these techniques for the NAS Origin cluster.

Smith, Warren↗
NASA NTRS1997

Convergence Analysis of a Cascade Architecture Neural Network

In this paper, we present a mathematical foundation, including a convergence analysis, for cascading architecture neural networks. From this, a mathematical foundation for the casade correlation learning algorithm can also be found. Furthermore, it becomes apparent that the cascade correlation scheme is a special case of an efficient hardware learning algorithm called Cascade Error Projection.

Neural Network↗
NASA NTRS2008

Instrument Display Visual Angles for Conventional Aircraft and the MQ-9 Ground Control Station

Aircraft instrument panels should be designed such that primary displays are in optimal viewing location to minimize pilot perception and response time. Human Factors engineers define three zones (i.e. cones ) of visual location: 1) "Easy Eye Movement" (foveal vision); 2) "Maximum Eye Movement" (peripheral vision with saccades), and 3) "Head Movement (head movement required). Instrument display visual angles were measured to determine how well conventional aircraft (T-34, T-38, F- 15B, F-16XL, F/A-18A, U-2D, ER-2, King Air, G-III, B-52H, DC-10, B747-SCA) and the MQ-9 ground control station (GCS) complied with these standards, and how they compared with each other. Selected instrument parameters included: attitude, pitch, bank, power, airspeed, altitude, vertical speed, heading, turn rate, slip/skid, AOA, flight path, latitude, longitude, course, bearing, range and time. Vertical and horizontal visual angles for each component were measured from the pilot s eye position in each system. The vertical visual angles of displays in conventional aircraft lay within the cone of "Easy Eye Movement" for all but three of the parameters measured, and almost all of the horizontal visual angles fell within this range. All conventional vertical and horizontal visual angles lay within the cone of Maximum Eye Movement. However, most instrument vertical visual angles of the MQ-9 GCS lay outside the cone of Easy Eye Movement, though all were within the cone of Maximum Eye Movement. All the horizontal visual angles for the MQ-9 GCS were within the cone of "Easy Eye Movement". Most instrument displays in conventional aircraft lay within the cone of Easy Eye Movement, though mission-critical instruments sometimes displaced less important instruments outside this area. Many of the MQ-9 GCS systems lay outside this area. Specific training for MQ-9 pilots may be needed to avoid increased response time and potential error during flight. The learning objectives include: 1) Know three physiologic cones of eye/head movement; 2) Understand how instrument displays comply with these design principles in conventional aircraft and an uninhabited aerial vehicle system. Which of the following is NOT a recognized physiologic principle of instrument display design? Cone of Easy Eye Movement 2) Cone of Binocular Eye Movement 3) Cone of Maximum Eye Movement 4) Cone of Head Movement 5) None of the above. Answer: # 2) Cone of Binocular Eye Movement

Kamine, Tovy Haber↗
NASA NTRS2018

JSC OCFO Cloud Integrated Budget Analytical Toolbox

The JSC OCFO analyst's job function relies heavily on analytics to provide every directorate customer with quality service in managing their respective budgets. This technology, configured on the Salesforce cloud platform, enhances the JSC OCFO customer user interface, provides state of the art data analytics and statistics, reduces IT costs, increases productivity, eliminates duplicate entry and errors, and utilizes machine learning and artificial intelligence to process day to day transactional tasks automatically without analyst manipulation. The unique configuration of the COTS technology allows for highly advanced analysis and reporting of government budget data. This provides the ability for complex budgets to be fully understood in detail by non-budget organizations and eliminates time consuming analysis on the OCFO analyst's behalf.

Robertson, Arden↗