Results for “Buildings and Connected Systems” — Page 9

NASA NTRS2013

Compute Element and Interface Box for the Hazard Detection System

The Autonomous Landing and Hazard Avoidance Technology (ALHAT) program is building a sensor that enables a spacecraft to evaluate autonomously a potential landing area to generate a list of hazardous and safe landing sites. It will also provide navigation inputs relative to those safe sites. The Hazard Detection System Compute Element (HDS-CE) box combines a field-programmable gate array (FPGA) board for sensor integration and timing, with a multicore computer board for processing. The FPGA does system-level timing and data aggregation, and acts as a go-between, removing the real-time requirements from the processor and labeling events with a high resolution time. The processor manages the behavior of the system, controls the instruments connected to the HDS-CE, and services the "heavy lifting" computational requirements for analyzing the potential landing spots.

Villalpando, Carlos Y.↗

NASA NTRSDate not supplied

Open Data Integration (ODIN): A Concurrent, Distributed Message-Based Architecture and Framework for Disaster Response

The Runtime for Airspace Concept Evaluation (RACE) is an open-source software architecture and framework to build configurable, highly concurrent and distributed message-based systems that offer scalable, low-latency performance on commodity hardware. RACE was used in commercial aviation applications to rapidly build systems that span several machines (including synchronized displays), interface existing hardware simulators and other live data feeds, and incorporate sophisticated visualization components such as NASA WorldWind. These RACE applications validated elements of the FAA’s System Wide Information Management (SWIM) Program, handling up to 1000 messages/sec from diverse sources (SFDPS, TFM-DATA, TAIS, ASDE-X, ITWS and local ADS) for 4,500 simultaneous flights tracked in the next-generation air transportation system’s digital backbone. We have since generalized RACE to support Open Data Integration (ODIN) applications outside aviation. Systems built with RACE/ODIN can be deployed in the field, on commodity hardware, and operate with limited or intermittent connectivity to the outside world. Our primary use case is a web-server with local/persistent data storage that runs within and only serves the stakeholder network (e.g. an incident command post). We are tailoring the RACE/ODIN system to support wildland fire management for the upcoming NASA Wildland Fire Safety Demonstration Series. RACE-ODIN is under consideration for application in the Scalable Traffic Management for Emergency Response Operations project, or STEReO, which aims to create a system that can be deployed during emergencies, to coordinate multiple elements of disaster response. Such data sources predominantly come from existing services on the internet (e.g. weather and satellite data, imported from so called "edge servers") but can also include dynamic (real-time) data from computer simulations and within the stakeholder network (such as aircraft and personnel tracking information). We will present the architecture and ODIN system demonstration incorporating local data from instrumented power-line towers, interpolated weather data and geospatial data from space-based platforms.

Joseph C Coughlan↗

NASA NTRSDate not supplied

Feasibility Study of Distributed Decision-Making on the Edge for Urban Air Mobility

The Concept of Operations for Urban Air Mobility (UAM) put forward by FAA, NASA, and several industry stakeholders acknowledges the diversity and complexity in UAM operations and, thereby, envisions a federated architecture for UAM management. In this architecture, the decision-making is distributed to a set of service providers who collectively manage the shared airspace usage by different stakeholders. This notionally brings autonomy closer to the UAM businesses and encourages to explore the feasibility of decision making on the very edge, which is the topic of the presented research. This paper reports research conducted on the hypothesis based on which the residual compute capability onboard smart unmanned aerial systems (UASs) is utilized to build situational awareness and resolve conflicts by passive and active coordination among multiple UASs, thereby implementing a layer of distributed autonomy in UAM. Key features of the edge-computing approach involve inter-UAS information exchange, independent assessment of own flight and environmental conditions, and estimation of other UASs’ flight preferences, incorporating machine learning techniques in the last two. Parallel computing on portable graphics processing unit (GPU) enables the machine learning workflow on the edge. A custom-built 3D simulator is used to evaluate the efficacy of the distributed decision-making on the edge. Each edge node, representing a smart UAS, connects to the simulator from a remote location and independently controls the behavior of the corresponding virtual asset in the simulator, analogous to participants in an online multi-player game. The presented edge-computing-based distributed decision-making framework is envisioned to pave the way for collective mobility of autonomous air vehicles in the future shared airspace, while allowing the inclusion of the business preferences of the UAS operators within allowed regulatory limits.

Edge computing↗

NASA NTRSDate not supplied

Feasibility Study of Distributed Decision-Making on the Edge for Urban Air Mobility

The Concept of Operations for Urban Air Mobility (UAM) put forward by FAA, NASA, and several industry stakeholders acknowledges the diversity and complexity in UAM operations and, thereby, envisions a federated architecture for UAM management. In this architecture, the decision-making is distributed to a set of service providers who collectively manage the shared airspace usage by different stakeholders. This notionally brings autonomy closer to the UAM businesses and encourages to explore the feasibility of decision making on the very edge, which is the topic of the presented research. This paper reports research conducted on the hypothesis based on which the residual compute capability onboard smart unmanned aerial systems (UASs) is utilized to build situational awareness and resolve conflicts by passive and active coordination among multiple UASs, thereby implementing a layer of distributed autonomy in UAM. Key features of the edge-computing approach involve inter-UAS information exchange, independent assessment of own flight and environmental conditions, and estimation of other UASs’ flight preferences, incorporating machine learning techniques in the last two. Parallel computing on portable graphics processing unit (GPU) enables the machine learning workflow on the edge. A custom-built 3D simulator is used to evaluate the efficacy of the distributed decision-making on the edge. Each edge node, representing a smart UAS, connects to the simulator from a remote location and independently controls the behavior of the corresponding virtual asset in the simulator, analogous to participants in an online multi-player game. The presented edge-computing-based distributed decision-making framework is envisioned to pave the way for collective mobility of autonomous air vehicles in the future shared airspace, while allowing the inclusion of the business preferences of the UAS operators within allowed regulatory limits.

Edge computing↗

NASA NTRSDate not supplied

Open Data Integration (ODIN): A Concurrent, Distributed Message-Based Architecture and Framework for Disaster Response

The Runtime for Airspace Concept Evaluation (RACE) is an open-source software architecture and framework to build configurable, highly concurrent and distributed message-based systems that offer scalable, low-latency performance on commodity hardware. RACE was used in commercial aviation applications to rapidly build systems that span several machines (including synchronized displays), interface existing hardware simulators and other live data feeds, and incorporate sophisticated visualization components such as NASA WorldWind. These RACE applications validated elements of the FAA’s System Wide Information Management (SWIM) Program, handling up to 1000 messages/sec from diverse sources (SFDPS, TFM-DATA, TAIS, ASDE-X, ITWS and local ADS) for 4,500 simultaneous flights tracked in the next-generation air transportation system’s digital backbone. We have since generalized RACE to support Open Data Integration (ODIN) applications outside aviation. Systems built with RACE/ODIN can be deployed in the field, on commodity hardware, and operate with limited or intermittent connectivity to the outside world. Our primary use case is a web-server with local/persistent data storage that runs within and only serves the stakeholder network (e.g. an incident command post). We are tailoring the RACE/ODIN system to support wildland fire management for the upcoming NASA Wildland Fire Safety Demonstration Series. RACE-ODIN is under consideration for application in the Scalable Traffic Management for Emergency Response Operations project, or STEReO, which aims to create a system that can be deployed during emergencies, to coordinate multiple elements of disaster response. Such data sources predominantly come from existing services on the internet (e.g. weather and satellite data, imported from so called "edge servers") but can also include dynamic (real-time) data from computer simulations and within the stakeholder network (such as aircraft and personnel tracking information). We will present the architecture and ODIN system demonstration incorporating local data from instrumented power-line towers, interpolated weather data and geospatial data from space-based platforms.

Guillaume P Brat↗

NASA NTRS1994

Potential aerospace applications of high temperature superconductors

The recent discovery of High Temperature Superconductors (HTS) with superconducting transition temperature, T(sub c), above the boiling point of liquid nitrogen has opened the door for using these materials in new and practical applications. These materials have zero resistance to electric current, have the capability of carrying large currents and as such have the potential to be used in high magnetic field applications. One of the space applications that can use superconductors is electromagnetic launch of payloads to low-earth-orbit. An electromagnetic gun-type launcher can be used in small payload systems that are launched at very high velocity, while sled-type magnetically levitated launcher can be used to launch larger payloads at smaller velocities. Both types of launchers are being studied by NASA and the aerospace industry. The use of superconductors will be essential in any of these types of launchers in order to produce the large magnetic fields required to obtain large thrust forces. Low Temperature Superconductor (LTS) technology is mature enough and can be easily integrated in such systems. As for the HTS, many leading companies are currently producing HTS coils and magnets that potentially can be mass-produced for these launchers. It seems that designing and building a small-scale electromagnetic launcher is the next logical step toward seriously considering this method for launching payloads into low-earth-orbit. A second potential application is the use of HTS to build sensitive portable devices for the use in Non Destructive Evaluation (NDE). Superconducting Quantum Interference Devices (SQUID's) are the most sensitive instruments for measuring changes in magnetic flux. By using HTS in SQUID's, one will be able to design a portable unit that uses liquid nitrogen or a cryocooler pump to explore the use of gradiometers or magnetometers to detect deep cracks or corrosion in structures. A third use is the replacement of Infra-Red (IR) sensor leads on Earth Orbit Systems (EOS) with HTS leads. IR detectors on these EOS missions are cooled to a 4.2K to improve their signal to noise ratio. They are connected to data acquisitions systems using manganin wires (low thermal conductors) to reduce the heat load on the cryogen. Replacing these wires with HTS leads will increase the lifetime of these missions by about 50 percent. This is a promising application that is ready for actual implementation on such systems. The analysis also show that an the number of IR detectors increase in larger EOS systems, substantial increase in the lifetime of each mission will be realized by using HTS leads instead of the manganin ones.

Selim, Raouf↗

NASA NTRSDate not supplied

In-situ Imaging of Pyrolyzing Aerospace Materials

Tracking morphological changes of materials during heating is crucial to understand its response in fire protection, biofuel production, thermal protection systems (TPS) for hypersonic flight. As materials are heated, they undergo physical and chemical changes due to water loss, stretching or shrinking, pyrolysis and chemical reactions in the ambient environment. The effects of these changes can have a profound impact on the material’s performance, indicated by changes in on the porosity and volume. While materials such as wood shrink as they pyrolyze and lose mass, others swell due to their inherent characteristics when exposed to heat [1]. This study focuses on experiments conducted at the Advanced Light Source (ALS) beamline 8.3.2, where in situ micro-computed tomography (µ-CT) is performed on materials as they are being pyrolyzed. Through in situ µ-CT, the change in total volume and porosity can be obtained in real-time, allowing for better understanding of the underlying thermophysical and chemical processes as a function of temperature. This study also focuses on the implementation of the Porous Microstructure Analysis software (PuMA) [2] to obtain thermal conductivity, permeability, and other properties of the material from the 3D tomographies. The information gained from these tomographies will supplement microscale model development of material morphological change and will aid macroscale modeling for high-temperature applications. For this study, Room Temperature Vulcanizing silicone (RTV) [3-5] is heated from room temperature to 1000°C using an infrared lamp heating system, and tomographies are continuously collected as the sample is heated. The tomographies are then segmented to obtain solid and void phases, from which estimates of pore size, porosity and total volume are extracted as a function of temperature. PuMA is deployed on the segmented tomographies to obtain thermal conductivity, permeability, and other properties as a function of temperature. Preliminary results show that RTV first intumesces (swells) as pyrolysis begins, due to build-up of pyrolysis gases in closed pores, and then shrinks significantly as more open pores are formed and the pyrolysis gases outgas. Pore network visualization of the tomographies using OpenPNM [6] showed the increase in pore connectivity with increase in temperature. Future work will focus on using PuMA to obtain macroscopic properties of RTV as a function of temperature.

Tomography↗

NASA NTRS1997

Parallel adaptive mesh refinement techniques for plasticity problems

The accurate modeling of the nonlinear properties of materials can be computationally expensive. Parallel computing offers an attractive way for solving such problems; however, the efficient use of these systems requires the vertical integration of a number of very different software components, we explore the solution of two- and three-dimensional, small-strain plasticity problems. We consider a finite-element formulation of the problem with adaptive refinement of an unstructured mesh to accurately model plastic transition zones. We present a framework for the parallel implementation of such complex algorithms. This framework, using libraries from the SUMAA3d project, allows a user to build a parallel finite-element application without writing any parallel code. To demonstrate the effectiveness of this approach on widely varying parallel architectures, we present experimental results from an IBM SP parallel computer and an ATM-connected network of Sun UltraSparc workstations. The results detail the parallel performance of the computational phases of the application during the process while the material is incrementally loaded.

Barry, W. J.↗

NASA NTRS2024

Pliocene Model Intercomparison Project Phase 3 (PlioMIP3) – Science Plan and Experimental Design

The Pliocene Model Intercomparison Project (PlioMIP) was initiated in 2008. Over two phases PlioMIP has helped co-ordinate the experimental design and publication strategy of the community, which has included an increasing number of climate models and modelling groups from around the world. It has engaged with palaeoenvironmental scientists to foster new data synthesis supporting the construction of new model boundary conditions, as well as to facilitate new data-model comparisons. The work has advanced our understanding of Pliocene climates and environments, enhanced our knowledge regarding the ability of complex climate and Earth System models to accurately simulate climate change, and helped to refine our estimates of how sensitive the climate system is to forcing conditions. In this community protocol paper, we outline the scientific plan for PlioMIP Phase 3 (PlioMIP3). This plan provides the required guidance to participating modelling groups from around the world to successfully set up and perform PlioMIP3 climate model experiments. The project is open to new participants from the scientific community (both from the climate modelling and geosciences communities). In PlioMIP3, we retain the PlioMIP2 Core experiments (Eoi 400 , E 280 ) and extend the Core requirements to include either an experiment focussed on the Early Pliocene or an alternative Late Pliocene simulation (or both). These additions (a) allow a comparison of Early and Late Pliocene warm intervals and help build research connections and synergy with the MioMIP (Miocene Model Intercomparison Project - also known as DeepMIP-Miocene) and PlioMioVAR projects (Pliocene-Miocene Variability Working Group), and (b) create an alternative time slice simulation for 3.205 Ma (MIS KM5c) through removal of some of the largest palaeogeographic differences introduced between PlioMIP1 and 2 resulting in minimal land-sea mask variations from the modern. In addition, we present ten optional experiments designed to enhance our assessment of climate sensitivity and to explore the uncertainty in greenhouse gas-related forcing. For the first time, we introduce orbital sensitivity experiments into the science plan, as well as simulations incorporating dynamic vegetation-climate feedbacks and an experiment designed to examine the potential significance of East Antarctic Ice Sheet boundary condition uncertainty. These changes enhance palaeo-to-future scientific connections and enable an exploration of the significance of palaeogeographic uncertainties on climate simulations.

Climate modelling↗

NASA NTRS2019

Geometry Systems for Lattice-Based Reconfigurable Space Structures

We describe analytical methods for the design of the discrete elements of ultralight lattice structures. This modular, building block strategy allows for relatively simple element manufacturing, as well as relatively simple robotic assembly of low mass density structures on orbit, with potential for disassembly and reassembly into highly varying and large structures. This method also results in a structure that is easily navigable by relatively small mobile robots. The geometry of the cell can allow for high packing efficiency to minimize wasted payload volume while maximizing structural performance and constructability. We describe the effect of geometry choices on the final system mechanical properties and automated robotic constructability of a final system. Geometric properties considered include number of attachments per voxel, number of attachments per coefficient of volume, and effects of vertex, edge, and face connectivity of the unit cell. Mechanical properties considered include strength scaling, modulus scaling, and packing efficiency of the lattice. Automated constructibility metrics include volume allowance for an end-effector, strut clearance angle for an end-effector, and packing efficiency. These metrics were applied to six lattice unit cell geometries: cube, cuboctahedron, octahedron, octet, rhombic dodecahedron, and truncated octahedron. A case study is presented to determine the most suitable lattice system for a specific set of strength and modulus scaling requirements while optimizing for ease of robotic assembly.

lattice geometry↗

NASA NTRSDate not supplied

Geometry Systems for Lattice-Based Reconfigurable Space Structures

We describe analytic methods for the design of the discrete elements of ultralight lattice structures. This modular building block strategy allows for relatively simple element manufacturing as well as relatively simple robotic assembly of low mass-density structures on orbit, with potential for disassembly and reassembly into highly varying and large structures. This method also results in a structure that is easily navigable by relatively small, mobile robots. The geometry of the cell can allow for high packing efficiency to minimize wasted payload volume while maximizing structural performance and constructability. We describe the effect of geometry choices on the mechanical properties and automated robotic constructability of a final system. Geometric properties considered include number of attachments per voxel, number of attachments per coefficient of volume, and effects of vertex, edge, and face connectivity of the unit cell. Mechanical properties considered include strength scaling, modulus scaling, and packing efficiency of the lattice. Automated constructibility metrics include volume allowance for an end-effector, strut clearance angle for an end-effector, and packing efficiency. These metrics were applied to six lattice unit cell geometries: cube, cuboctahedron, octahedron, octet, rhombic dodecahedron, and truncated octahedron. A case study is presented to determine the most suitable lattice system for a specific set of strength and modulus scaling requirements while optimizing for ease of robotic assembly.

digital materials↗

NASA NTRS1992

Autonomous Space Processor for Orbital Debris (ASPOD)

A project in the Advanced Design Program at the University of Arizona is described. The project is named the Autonomous Space Processor for Orbital Debris (ASPOD) and is a Universities Space Research Association (USRA) sponsored design project. The development of ASPOD and the students' abilities in designing and building a prototype spacecraft are the ultimate goals of this project. This year's focus entailed the development of a secondary robotic arm and end-effector to work in tandem with an existent arm in the removal of orbital debris. The new arm features the introduction of composite materials and a linear drive system, thus producing a light-weight and more accurate prototype. The main characteristic of the end-effector design is that it incorporates all of the motors and gearing internally, thus not subjecting them to the harsh space environment. Furthermore, the arm and the end-effector are automated by a control system with positional feedback. This system is composed of magnetic and optical encoders connected to a 486 PC via two servo-motor controller cards. Programming a series of basic routines and sub-routines allowed the ASPOD prototype to become more autonomous. The new system is expected to perform specified tasks with a positional accuracy of 0.5 cm.

Ramohalli, Kumar↗

NASA NTRS1992

Autonomous space processor for orbital debris advanced design project in support of solar system exploration

This paper is regarding a project in the Advanced Design Program at the University of Arizona. The project is named the Autonomous Space Processor for Orbital Debris (ASPOD) and is a NASA/Universities Space Research Association (USRA) sponsored design project. The development of ASPOD and the students' abilities in designing and building a prototype spacecraft are the ultimate goals of this project. This year's focus entailed the development of a secondary robotic arm and end-effector to work in tandem with an existent arm in the removal of orbital debris. The new arm features the introduction of composite materials and a linear drive system, thus producing a light-weight and more accurate prototype. The main characteristic of the end-effector design is that it incorporates all of the motors and gearing internally, thus not subjecting them to the harsh space environment. Furthermore, the arm and the end-effector are automated by a control system with positional feedback. This system is composed of magnetic and optical encoders connected to a 486 PC via two servo-motor controller cards. Programming a series of basic routines and sub-routines has allowed the ASPOD prototype to become more autonomous. The new system is expected to perform specified tasks with a positional accuracy of 0.5 cm.

Ramohalli, Kumar↗

NASA NTRS2004

Viewing ISS Data in Real Time via the Internet

EZStream is a computer program that enables authorized users at diverse terrestrial locations to view, in real time, data generated by scientific payloads aboard the International Space Station (ISS). The only computation/communication resource needed for use of EZStream is a computer equipped with standard Web-browser software and a connection to the Internet. EZStream runs in conjunction with the TReK software, described in a prior NASA Tech Briefs article, that coordinates multiple streams of data for the ground communication system of the ISS. EZStream includes server components that interact with TReK within the ISS ground communication system and client components that reside in the users' remote computers. Once an authorized client has logged in, a server component of EZStream pulls the requested data from a TReK application-program interface and sends the data to the client. Future EZStream enhancements will include (1) extensions that enable the server to receive and process arbitrary data streams on its own and (2) a Web-based graphical-user-interface-building subprogram that enables a client who lacks programming expertise to create customized display Web pages.

Myers, Gerry↗

NASA NTRS2024

Hawaiʻi Urban Development: Mapping Urban Infrastructure and Linkages to Wastewater Impacts across Hawaiʻi

Coastal ecosystems and human health across the Hawaiʻian islands are under threat from cesspools and other unpermitted wastewater infrastructure. Cesspools, or home sewage systems that hold wastewater effluent in porous concrete containers underground, discharge raw, untreated sewage directly into waterways. In fact, it is estimated that approximately 88,000 cesspools release 55 million gallons of raw sewage into Hawaiʻi’s waterways each day. Our partners, The Nature Conservancy, Hawaiʻi Regional Office (TNC) and University of Hawaiʻi, Mānoa (UHM), are heavily involved with the implementation and impact of Act 125, which was passed in 2017 by Hawaiʻi State Legislature to help combat these effects. This requires the conversion of cesspools to be upgraded and either converted to approved onsite disposal systems or connected to a sewer system so that the wastewater may be treated and to avoid leaching into the groundwater. Thus, identifying and locating cesspool and other unpermitted wastewater infrastructure is key to state and partner goals. Utilizing NASA Earth observations including Landsat 7 Enhanced Thematic Mapper Plus (ETM+), Landsat 8 Operational Land Imager (OLI), and Landsat 9 OLI-2 imagery, as well as pre-classified Sentinel-2 Multispectral Instrument (MSI) imagery in Esri’s Living Atlas, our team showcased the various perspectives that wastewater pollution research can use to assist in identifying and tracking urban development and concrete condition. This process was demonstrated by utilizing supervised object-based classification models, calculating various urban indices, and finding building size estimates to calculate effluent levels. Additionally, providing our partners with a GIS Methodology that can be built on to reach their future needs.

urban classification↗

NASA NTRS1984

Long term cryogen storage in space - A feasibility study

The conduction of future space missions will require the long term storage of cryogens in space. Such missions will be performed in connection with the employment of the long life version of the Shuttle Infrared Telescope Facility (SIRTF), the Large Deployable Reflector (LDR), Gravity Probe B (GPB), Orbital Transfer Vehicles (OTVs), and the Space Station. The feasibility of building a long lifetime helium Dewar for space applications has been studied. A feasible concept was to be developed for a Shuttle-launchable 3-yr lifetime Dewar. The developed concept was to be used for the identification of key technology development areas. The desired long lifetime requires the minimization of the parasitic heat loads. It has been found that the majority of the parasitic heat load can be provided by the fiberglass-epoxy tension band supports. A new type of support system, the Passive Orbital Disconnect Strut (PODS), was developed to reduce the considered contribution to the heat load.

Kittel, P.↗

NASA NTRS1997

Providing the Persistent Data Storage in a Software Engineering Environment Using Java/COBRA and a DBMS

An investigation was undertaken to build the software foundation for the WHERE (Web-based Hyper-text Environment for Requirements Engineering) project. The TCM (Toolkit for Conceptual Modeling) was chosen as the foundation software for the WHERE project which aims to provide an environment for facilitating collaboration among geographically distributed people involved in the Requirements Engineering process. The TCM is a collection of diagram and table editors and has been implemented in the C++ programming language. The C++ implementation of the TCM was translated into Java in order to allow the editors to be used for building various functionality of the WHERE project; the WHERE project intends to use the Web as its communication back- bone. One of the limitations of the translated software (TcmJava), which militated against its use in the WHERE project, was persistent data management mechanisms which it inherited from the original TCM; it was designed to be used in standalone applications. Before TcmJava editors could be used as a part of the multi-user, geographically distributed applications of the WHERE project, a persistent storage mechanism must be built which would allow data communication over the Internet, using the capabilities of the Web. An approach involving features of Java, CORBA (Common Object Request Broker), the Web, a middle-ware (Java Relational Binding (JRB)), and a database server was used to build the persistent data management infrastructure for the WHERE project. The developed infrastructure allows a TcmJava editor to be downloaded and run from a network host by using a JDK 1.1 (Java Developer's Kit) compatible Web-browser. The aforementioned editor establishes connection with a server by using the ORB (Object Request Broker) software and stores/retrieves data in/from the server. The server consists of a CORBA object or objects depending upon whether the data is to be made persistent on a single server or multiple servers. The CORBA object providing the persistent data server is implemented using the Java progranu-ning language. It uses the JRB to store/retrieve data in/from a relational database server. The persistent data management system provides transaction and user management facilities which allow multi-user, distributed access to the stored data in a secure manner.

Dhaliwal, Swarn S.↗

NASA NTRS1992

Learning and optimization with cascaded VLSI neural network building-block chips

To demonstrate the versatility of the building-block approach, two neural network applications were implemented on cascaded analog VLSI chips. Weights were implemented using 7-b multiplying digital-to-analog converter (MDAC) synapse circuits, with 31 x 32 and 32 x 32 synapses per chip. A novel learning algorithm compatible with analog VLSI was applied to the two-input parity problem. The algorithm combines dynamically evolving architecture with limited gradient-descent backpropagation for efficient and versatile supervised learning. To implement the learning algorithm in hardware, synapse circuits were paralleled for additional quantization levels. The hardware-in-the-loop learning system allocated 2-5 hidden neurons for parity problems. Also, a 7 x 7 assignment problem was mapped onto a cascaded 64-neuron fully connected feedback network. In 100 randomly selected problems, the network found optimal or good solutions in most cases, with settling times in the range of 7-100 microseconds.

Duong, T.↗