SEARCH · Engineering Papers
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.
In this presentation, we will describe how the system uses CIRs in terms of subscriptions, filters and priorities, and how client interests are propagated throughout the system, and we will describe the novel information flows that result from this CIR propagation.
We are using Web (Grid) service technology to demonstrate the assimilation of multiple distributed data sources (a typical data grid problem) into a major parallel high-performance computing earthquake forecasting code. Such a linkage of Geoinformatics with Geocomplexity demonstrates the value of the Solid Earth Research Virtual Observatory (SERVO) Grid concept, and advance Grid technology by building the first real-time large-scale data assimilation grid Here we develop the next steps for both the SERVO concept and the identified need for a Solid Earth problem-solving environment. We use a challenging motivating problem of importance to NASA namely integrating NASA space geodetic observations with numerical simulations of a changing earth.
This paper investigates the effects of axial liner displacement on modal attenuation and scattering. Tests are conducted with a uniform liner mounted in the NASA Langley Curved Duct Test Rig, with a multimode source at flow conditions of Mach 0.0 and 0.3. The impedance of the liner is educed from data acquired in the NASA Langley Grazing Flow Impedance Tube. This impedance is then used as input to the CDUCT-LaRC propagation code to enable comparisons of predicted and measured modal content upstream and downstream of the liner as the axial displacement is varied. Results show that the liner position has a clear effect on the sound field as the number of modes is increased from two to six. The comparison of measured and predicted modal content is very good when only a few modes are present. For frequencies in which more modes are present, this comparison is less favorable very near resonance, but improves for frequencies away from resonance.
In complex environments where the human operator is a supervisor, he must allocate his attention between different kinds of tasks for optimum overall performance. When a portion of the future reference trajectory for the control task is available for preview, scheduling various activities is possible. A model has been developed for this situation using dynamic programming to solve an optimal control problem. An experiment was conducted where subjects controlled an airplane symbol over a map, shown a fixed distance into the future. Discrete, non-control tasks were introduced as number entry tasks. Results from the model are compared with experimental results.
In complex environments where the human operator is a supervisor, he must allocate his attention between different kinds of tasks for satisfactory overall performance. When a portion of the future reference trajectory for a continuous control task is available for preview, scheduling various other discrete activities is possible. A model has been developed for this situation using dynamic programming to solve an optimal control problem. An experiment was conducted where subjects controlled an airplane symbol over a map, shown a fixed distance into the future. Discrete tasks were introduced as data entry tasks. Results showed that the model compared favorably with experimental results.
Air transportation is undergoing a rapid evolution globally with the introduction of Advanced Air Mobility (AAM) and with it comes novel challenges and opportunities for transforming aviation. As AAM operations introduce increasing heterogeneity in vehicle capabilities and density, increased levels of automation are likely necessary to achieve operational safety and efficiency goals. This paper focuses on one example where increased automation has been suggested. Autonomous operations will need contingency management systems that can monitor evolving risk across a span of interrelated (or interdependent) hazards and, if necessary, execute appropriate control interventions via supervised or automated decision making. Accommodating this complex environment may require automated functions (autonomy) that apply artificial intelligence (AI) techniques that can adapt and respond to a quickly changing environment. This paper explores the use of Deep Reinforcement Learning (DRL) which has shown promising performance in complex and high-dimensional environments where the objective can be constructed as a sequential decision-making problem. An extension of a prior formulation of the contingency management problem as a Markov Decision Process (MDP) is presented and uses a DRL framework to train agents that mitigate hazards present in the simulation environment. A comparison of these learning-based agents and classical techniques is presented in terms of their performance, verification difficulties, and development process.
Air transportation is undergoing a rapid evolution globally with the introduction of Advanced Air Mobility (AAM) and with it comes novel challenges and opportunities for transforming aviation. As AAM operations introduce increasing heterogeneity in vehicle capabilities and density, increased levels of automation are likely necessary to achieve operational safety and efficiency goals. This paper focuses on one example where increased automation has been suggested. Autonomous operations will need contingency management systems that can monitor evolving risk across a span of interrelated (or interdependent) hazards and, if necessary, execute appropriate control interventions via supervised or automated decision making. Accommodating this complex environment may require automated functions (autonomy) that apply artificial intelligence (AI) techniques that can adapt and respond to a quickly changing environment. This paper explores the use of Deep Reinforcement Learning (DRL) which has shown promising performance in complex and high-dimensional environments where the objective can be constructed as a sequential decision-making problem. An extension of a prior formulation of the contingency management problem as a Markov Decision Process (MDP) is presented and uses a DRL framework to train agents that mitigate hazards present in the simulation environment. A comparison of these learning-based agents and classical techniques is presented in terms of their performance, verification difficulties, and development process.
The continued operational environment complexity faced by the Department of Defense, despite a restricted resource environment, is a mandate for greater adaptability and availability in joint training. To address these constraints, this paper proposes a model for the potential integration of adaptability training, virtual world capabilities and immersive training into the wider Joint Live Virtual and Constructive (JLVC) Federation, supported by human, social, cultural and behavior modeling, and measurement and assessment. By fusing those capabilities and modeling and simulation enhancements into the JLVC federation, it will create a force who is more apt to arrive at and implement correct decisions, and more able to appropriately seize initiative in the field. The model would allow for the testing and training of capabilities and TTPs that cannot be reasonably explored to their logical conclusions in a 'live' environment, as well as enhance training fidelity for all echelons and tasks.
Space-based systems are developing into critical infrastructure required to support the quality of life on Earth. Hence, spacecraft reliability is a serious issue that is complicated by exposure to the space environment. Complex mission designs along with rapidly evolving technologies have outpaced efforts to accommodate detrimental space environment impacts on systems. Hazardous space environments, the effects on systems, and the accommodation of the effects are described with a focus on the need to predict space environments.
There are two main algorithmic approaches to autonomous driving systems: (1) An end-to-end system in which a single deep neural network learns to map sensory input directly into appropriate warning and driving responses. (2) A mediated hybrid recognition system in which a system is created by combining independent modules that detect each semantic feature. While some researchers believe that deep learning can solve any problem, others believe that a more engineered and symbolic approach is needed to cope with complex environments with less data. Deep learning alone has achieved state-of-the-art results in many areas, from complex gameplay to predicting protein structures. In particular, in image classification and recognition, deep learning models have achieved accuracies as high as humans. But sometimes it can be very difficult to debug if the deep learning model doesn't work. Deep learning models can be vulnerable and are very sensitive to changes in data distribution. Generalization can be problematic. It's usually hard to prove why it works or doesn't. Deep learning models can also be vulnerable to adversarial attacks. Here, we combine deep learning-based object recognition and tracking with an adaptive neurosymbolic network agent, called the Non-Axiomatic Reasoning System (NARS), that can adapt to its environment by building concepts based on perceptual sequences. We achieved an improved intersection-over-union (IOU) object recognition performance of 0.65 in the adaptive retraining model compared to IOU 0.31 in the COCO data pre-trained model. We improved the object detection limits using RADAR sensors in a simulated environment, and demonstrated the weaving car detection capability by combining deep learning-based object detection and tracking with a neurosymbolic model.
The dominant technology for developing AI applications is to work in a multi-mechanism, integrated, knowledge-based system (KBS) development environment. Unfortunately, systems developed in such environments are inappropriate for delivering many applications - most importantly, they carry the baggage of the entire Lisp environment and are not written in conventional languages. One resolution of this problem would be to compile applications from complex environments to conventional languages. Here the first efforts to develop a system for compiling KBS developed in KEE to Ada (trademark). This system is called KATYDID, for KEE/Ada Translation Yields Development Into Delivery. KATYDID includes early prototypes of a run-time KEE core (object-structure) library module for Ada, and translation mechanisms for knowledge structures, rules, and Lisp code to Ada. Using these tools, part of a simple expert system was compiled (not quite automatically) to run in a purely Ada environment. This experience has given us various insights on Ada as an artificial intelligence programming language, potential solutions of some of the engineering difficulties encountered in early work, and inspiration on future system development.
The emergence and development of advanced technologies and vehicle types has created a growing demand for the introduction of new forms of flight operations. These new and increasingly complex operational paradigms such as Advanced and Urban Air Mobility (AAM/UAM) present regulatory authorities and the aviation community with several design and implementation challenges – particularly for highly autonomous vehicles. An overarching and daunting task is finding methods to integrate these emerging operations without compromising safety or disrupting traditional airspace operations. A shift toward a more predictive risk mitigation capability becomes critical to meet this challenge. This paper focus on the development and testing of a prognostic service aimed at estimating the quality of Global Navigation Satellite System (GNSS) performance for an autonomous aircraft in complex environments. The intent of this function is to proactively reduce a flight operations risk of exposure to states that may induce poor or unacceptable navigation system performance by factoring in estimates of GNSS quality into pre-flight and/or in-flight route planning. Methodologies for producing quality estimates are specified and results are provided for selected simulation and flight test cases.
Future NASA missions will require successful integration of the human with highly complex systems. Highly complex systems are likely to involve humans, automation, and some level of robotic assistance. The complex environments will require successful integration of the human with automation, with robots, and with human-automation-robot teams to accomplish mission critical goals. Many challenges exist for the human performing in these types of operational environments with these kinds of systems. Systems must be designed to optimally integrate various levels of inputs and outputs based on the roles and responsibilities of the human, the automation, and the robots; from direct manual control, shared human-robotic control, or no active human control (i.e. human supervisory control). It is assumed that the human will remain involved at some level. Technologies that vary based on contextual demands and on operator characteristics (workload, situation awareness) will be needed when the human integrates into these systems. Predictive models that estimate the impact of the technologies on the system performance and the on the human operator are also needed to meet the challenges associated with such future complex human-automation-robot systems in extreme environments.
Simultaneous Localization and Mapping (SLAM) is a promising technique that provides localization information and precise mapping of the physical environment without having much prior knowledge of the surroundings. SLAM may have a vital role in aeronautics and aerospace, where vehicles and aircraft must operate in complex environments with traditional localization services that may be degraded or unavailable. This paper compares several pre-canned 3D SLAM algorithms based on vision and LiDAR, namely ORB-SLAM, ORB-SLAM2, LOAM, A-LOAM, and F-LOAM on NASA UAS (Unmanned Aircraft System) flight test data. The NASA ARC UAS flight test demonstrates preliminary SLAM algorithm results, which serve as a stepping stone to simulated AAM (Advanced Air Mobility) concepts. Conducting AFRC UAS flight test for simulated AAM approach and landing with SLAM algorithms provides an Alternative Precision Navigation and Timing solution based on distributed landmarks and fiducials in the landing zone. These algorithms use the telemetry data as ground truth for a baseline comparison. The criteria of the performance comparison include robustness, accuracy, re-localization, response to environmental changes, and real-time effectiveness, which are currently qualitative but to be quantitative in the future.
Simultaneous Localization and Mapping (SLAM) is a promising technique that provides localization information and precise mapping of the physical environment without having much prior knowledge of the surroundings. SLAM may have a vital role in aeronautics and aerospace, where vehicles and aircraft must operate in complex environments with traditional localization services that may be degraded or unavailable. This paper compares several pre-canned 3D SLAM algorithms based on vision and LiDAR, namely ORB-SLAM, ORB-SLAM2, LOAM, A-LOAM, and F-LOAM on NASA UAS (Unmanned Aircraft System) flight test data. The NASA ARC UAS flight test demonstrates preliminary SLAM algorithm results, which serve as a stepping stone to simulated AAM (Advanced Air Mobility) concepts. Conducting AFRC UAS flight test for simulated AAM approach and landing with SLAM algorithms provides an Alternative Precision Navigation and Timing solution based on distributed landmarks and fiducials in the landing zone. These algorithms use the telemetry data as ground truth for a baseline comparison. The criteria of the performance comparison include robustness, accuracy, re-localization, response to environmental changes, and real-time effectiveness, which are currently qualitative but to be quantitative in the future.
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.
HERMIES-III is an autonomous robot comprised of a seven degree-of-freedom (DOF) manipulator designed for human scale tasks, a laser range finder, a sonar array, an omni-directional wheel-driven chassis, multiple cameras, and a dual computer system containing a 16-node hypercube expandable to 128 nodes. The current experimental program involves performance of human-scale tasks (e.g., valve manipulation, use of tools), integration of a dexterous manipulator and platform motion in geometrically complex environments, and effective use of multiple cooperating robots (HERMIES-IIB and HERMIES-III). The environment in which the robots operate has been designed to include multiple valves, pipes, meters, obstacles on the floor, valves occluded from view, and multiple paths of differing navigation complexity. The ongoing research program supports the development of autonomous capability for HERMIES-IIB and III to perform complex navigation and manipulation under time constraints, while dealing with imprecise sensory information.
NASA's future vision for interplanetary networking includes a lunar network, Cube Satellite (CubeSat) constellations, and deep space robotic missions, comprising what could be viewed as a network of networks. Delay-tolerant networking (DTN) architecture and protocols provide a standard network layer among these varying scenarios and mitigate many challenges of the space environment, such as long delays, unplanned service interruptions, and asymmetric links. The Cognitive Space Gateway (CSG) is a routing method in a DTN architecture that uses spiking neural networks as the learning element to optimize routing decisions in a complex environment. This work aims to further develop cognitive networking technologies in several critical areas, including DTN, the CSG algorithm, CubeSat swarm topologies, and cloud services. To test the algorithm in a realistic scenario, the emulated network topology is based on a CubeSat swarm. The swarm may function as a mesh of nodes or as a hub-and-spoke network. An emulation environment will be built upon a commercial cloud service, such as Amazon Web Services (AWS) Elastic Compute Cloud. The cloud environment may enable a flexible, lower maintenance approach versus a multi-hop network based in a physical laboratory. The cloud platform will provide a secure environment allowing for collaboration among government and academic entities.