SEARCH · Search NASA

Results for “sensor network design”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 361 records · Page 20

All sources NASA NTRS DOE OSTI

DOE OSTI2023

Digital Twin for Hydropower System Object Modeling: Alder Dam (FY2023)

Hydropower is the world's largest source of renewable electricity, and hydropower plants are distributed all over the world. Typical major components of a hydropower plant are the governor, excitation, generator, thrust bearing, hydraulic turbine, transformer, the main lead, metering and control, tailwater depression, and dissolved oxygen. For each component, various measures are taken. The measurements are acquired by various heterogeneous systems, including standalone sensors, programmable logic controllers (PLC), Supervisory control and data acquisition (SCADA), Internet of Things (IoT), and data acquisition and integration platforms such as OSI/PI. The measured data are often archived within the plant by a data management platform, and many institutions have cloud-based archive systems, such as Hydropower Research Institution (HRI), U.S. Army Corps of Engineers (USACE), and Columbia River Data Access in Real Time (DART). Object Modeling is a general framework for designing information systems. It focuses on objects, the actions they perform, and the messages they send to one another to cause those actions to be taken. The major differences among object modeling, network modeling, data modeling, and process modeling are that in the first we focus on the actions in response to information, objects which form the system, the actions they perform, and how they pass information to one another, while in the second we concentrate on where, when and how much information is moved, while in the third we focus on what information is moved and where it is moved, while in the last we focus on how it is moved and when it is moved. Object modeling was developed basically as a method to develop object-oriented systems and to support object-oriented programming. It describes the static structure of the system. The object Modeling Technique is easy to draw and use. That is why we choose object modeling to connect physical hydropower plants to Digital Twin. It recognizes the objects and the relationship between them. It identifies the attributes and functions of each class. Dynamic Modeling: It explains how objects respond to events. Functional Modeling indicates the processes executed in an object and how data changes when it moves to objects. It has been used in many applications like telecommunication, transportation, etc.

13 HYDRO ENERGY↗

NASA NTRS2020

PHALANX: Expendable Projectile Sensor Networks for Planetary Exploration

Technologies enabling long-term, wide-ranging measurement in hard-to-reach areas are a critical need for planetary science inquiry. Phenomena of interest include flows or variations in volatiles, gas composition or concentration, particulate density, or even simply temperature. Improved measurement of these processes enables understanding of exotic geologies and distributions or correlating indicators of trapped water or biological activity. However, such data is often needed in unsafe areas such as caves, lava tubes, or steep ravines not easily reached by current spacecraft and planetary robots. To address this capability gap, we have developed miniaturized, expendable sensors which can be ballistically lobbed from a robotic rover or static lander - or even dropped during a flyover. These projectiles can perform sensing during flight and after anchoring to terrain features. By augmenting exploration systems with these sensors, we can extend situational awareness, perform long-duration monitoring, and reduce utilization of primary mobility resources, all of which are crucial in surface missions. We call the integrated payload that includes a cold gas launcher, smart projectiles, planning software, network discovery, and science sensing: PHALANX. In this paper, we introduce the mission architecture for PHALANX and describe an exploration concept that pairs projectile sensors with a rover “mothership.” Science use cases explored include reconnaissance using ballistic cameras, volatiles detection, and building timelapse maps of temperature and illumination conditions. Strategies to autonomously coordinate constellations of deployed sensors to self-discover and localize with peer ranging (i.e. a “local GPS”) are summarized, thus providing communications infrastructure beyond-line-of-sight (BLOS) of the rover. Capabilities were demonstrated through both simulation and physical testing with a terrestrial prototype. The approach to developing a terrestrial prototype is discussed, including design of the launching mechanism, projectile optimization, micro-electronics fabrication, and sensor selection. Results from early testing and characterization of commercial-off-the-shelf (COTS) components are reported. Nodes were subjected to successful burn-in tests over 48 hours at full logging duty cycle. Integrated field tests were conducted in the Roverscape, a half-acre planetary analog environment at NASA Ames, where we tested up to 10 sensor nodes simultaneously coordinating with an exploration rover. Ranging accuracy has been demonstrated to be within +/-10cm over 20m using commodity radios when compared to high-resolution laser scanner ground truthing. Evolution of the design, including progressive miniaturization of the electronics and iterated modifications of the enclosure housing for streamlining and optimized radio performance are described. Finally, lessons learned to date, gaps toward eventual flight mission implementation, and continuing future development plans are discussed.

Dille, Michael↗

NASA NTRS1996

The NIST Real-Time Control System (RCS): A Reference Model Architecture for Computational Intelligence

The Real-time Control System (RCS) developed at NIST and elsewhere over the past two decades defines a reference model architecture for design and analysis of complex intelligent control systems. The RCS architecture consists of a hierarchically layered set of functional processing modules connected by a network of communication pathways. The primary distinguishing feature of the layers is the bandwidth of the control loops. The characteristic bandwidth of each level is determined by the spatial and temporal integration window of filters, the temporal frequency of signals and events, the spatial frequency of patterns, and the planning horizon and granularity of the planners that operate at each level. At each level, tasks are decomposed into sequential subtasks, to be performed by cooperating sets of subordinate agents. At each level, signals from sensors are filtered and correlated with spatial and temporal features that are relevant to the control function being implemented at that level.

Albus, James S.↗

DOE OSTI2024

Artificial Intelligence for Conjugated Polymers

Conjugated polymers have garnered significant attention due to their diverse applications in electronics, photonics, and energy storage. However, realizing their full potential poses a formidable challenge, as their design has historically relied on iterative adjustments and continuous inspiration from researchers. Traditional methods often struggle to efficiently navigate their vast chemical landscape. In this work, the application of artificial intelligence (AI), specifically machine learning (ML), needs to be discussed in the realm of conjugated polymers. Our paper emphasizes the importance of understanding the structure–property relationships of these polymers and how ML can facilitate property prediction and inverse-design. We delve into various chemical fingerprints, structural descriptors, and ML algorithms, showcasing their utility across a spectrum of applications, including simulations, glass transition temperature determination, photovoltaics, reorganization energy for charge transport, photocatalysts, and sensors. Finally, we give some outlooks in this filed and propose unexplored areas within the field that hold the potential to benefit from ML techniques.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2021

Advanced Computing, Data Science, and Artificial Intelligence Research Opportunities for Energy-Focused Transportation Science

The Energy Efficient Mobility Systems (EEMS) technology landscape is complex and rapidly evolving, which provides both tremendous opportunities and formidable challenges. Significant alterations to the mobility landscape are underway due to the advent of vehicle and infrastructure connectivity, autonomous driving, and rapid passenger- and freight-vehicle electrification. Advanced computing will play an increasingly important role in enabling the EEMS program to understand and identify the most important levers to improve the energy productivity of future integrated mobility systems. It is also driving new approaches to mobility and the research to unlock an affordable, efficient, safe, and accessible transportation future. Driving much of this change is the collection, analysis, and strategic use of massive amounts of diverse, complex data from infrastructure and vehicles with on-board sensors and data storage and transmission capabilities. Diverse and representative data are key to implementing approaches to maximize mobility energy productivity. While high-fidelity modeling of integrated transportation networks has strengthened our understanding of dynamic movement and behavior patterns, existing tools must be expanded beyond their current focus. This work necessitates data infrastructure investments (e.g., secure-streaming data platforms driven by ubiquitous sensors and video analytics) as well as investments in critical capabilities for large-scale automated analysis and organization using modern machine learning, statistics, and artificial intelligence. Other chief needs include agile, large-scale storage that can be quickly searched and queried for relevant data to support validation and model development, data-sharing agreements, and formatting standards for key data types. The future of public transit must be explored in greater detail, research must inform design, and opportunities must be identified for improving the mobility productivity of public transit in both urban and rural America.

33 ADVANCED PROPULSION SYSTEMS↗

NASA NTRS2011

The NASA Soil Moisture Active Passive (SMAP) Mission - Algorithm and Cal/Val Activities and Synergies with SMOS and Other L-Band Missions

NASA's Soil Moisture Active Passive (SMAP) mission, planned for launch in late 2014, has as its key measurement objective the frequent, global mapping of near-surface soil moisture and its freeze-thaw state. SMAP soil moisture and freeze/thaw measurements at 10 km and 3 km resolutions respectively, would enable significantly improved estimates of water, energy and carbon transfers between the land and atmosphere. Soil moisture control of these fluxes is a key factor in the performance of atmospheric models used for weather forecasts and climate projections Soil moisture measurements are also of great importance in assessing floods and for monitoring drought. In addition, observations of soil moisture and freeze/thaw timing over the boreal latitudes can help reduce uncertainties in quantifying the global carbon balance. The SMAP measurement concept utilizes an L-band radar and radiometer sharing a rotating 6-meter mesh reflector antenna. The SMAP radiometer and radar flight hardware and ground processing designs are incorporating approaches to identify and mitigate potential terrestrial radio frequency interference (RFI). The radar and radiometer instruments are planned to operate in a 680 km polar orbit, viewing the surface at a constant 40-degree incidence angle with a 1000-km swath width, providing 3-day global coverage. Data from the instruments would yield global maps of soil moisture and freeze/thaw state to be provided at 10 km and 3 km resolutions respectively, every two to three days. Plans are to provide also a radiometer-only soil moisture product at 40-km spatial resolution. This product and the underlying brightness temperatures have characteristics similar to those provided by the Soil Moisture and Ocean Salinity (SMOS) mission. As a result, there are unique opportunities for common data product development and continuity between the two missions. SMAP also has commonalities with other satellite missions having L-band radiometer and/or radar sensors applicable to soil moisture measurement, such as Aquarius, SAO COM, and ALOS-2. The algorithms and data products for SMAP are being developed in the SMAP Science Data System (SDS) Testbed. The algorithms are developed and evaluated in the SDS Testbed using simulated SMAP observations as well as observational data from current airborne and spaceborne L-band sensors including SMOS. The SMAP project is developing a Calibration and Validation (Cal/Val) Plan that is designed to support algorithm development (pre-launch) and data product validation (post-launch). A key component of the Cal/Val Plan is the identification, characterization, and instrumentation of sites that can be used to calibrate and validate the sensor data (Level I) and derived geophysical products (Level 2 and higher). In this presentation we report on the development status of the SMAP data product algorithms, and the planning and implementation of the SMAP Cal/Val program. Several components of the SMAP algorithm development and Cal/Val plans have commonality with those of SMOS, and for this reason there are shared activities and resources that can be utilized between the missions, including in situ networks, ancillary data sets, and long-term monitoring sites.

Njoku, Eni↗

DOE OSTI2022

Data recovery via covert cognizance for unattended operational resilience

One of the important premises of unattended operation, a highly promoted characteristic of fission batteries and advanced microreactors, is the ability to automate the analysis of sensors data used in support of operational monitoring and control. Here, to meet this vision, this work proposes a new monitoring and data recovery paradigm to ensure resilience against data corruption which may be the result of malicious intrusion into the reactor operational network. This is paramount to ensure 100% availability under contingency scenarios such as cyberattacks. In support of this vision, earlier work has presented the concept of covert cognizance and demonstrated its mathematical ability to identify and embed cognizance parameters under the noise-dominated null space of the sensors data. This work extends this concept and applies it in real-time to demonstrate three key characteristics: zero-impact, zero-observability, and data recovery, where the first characteristic is to ensure no impact on operation, the second is immunity to discovery by pattern recognition techniques, and the third is to allow recovery of corrupt or falsified data. Recognizing that fission batteries are designed to operate under steady state most of the time, we elect to employ a small modular reactor model under transient operational conditions to demonstrate the operational resilience enabled by the covert cognizance paradigm. Specifically, the PI controller is augmented with the covert cognizance modules to develop self-awareness and enable automatic data recovery. The developed modules are expected to be equally applicable to a wide range of advanced reactor technologies relying on full or partial unattended control.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

NASA NTRS2011

Wireless Sensor Networks for Developmental and Flight Instrumentation

Wireless sensor networks (WSN) based on the IEEE 802.15.4 Personal Area Network and ZigBee Pro 2007 standards are finding increasing use in home automation and smart energy markets providing a framework for interoperable software. The Wireless Connections in Space Project, funded by the NASA Engineering and Safety Center, is developing technology, metrics and requirements for next-generation spacecraft avionics incorporating wireless data transport. The team from Stennis Space Center and Mobitrum Corporation, working under a NASA SBIR grant, has developed techniques for embedding plug-and-play software into ZigBee WSN prototypes implementing the IEEE 1451 Transducer Electronic Datasheet (TEDS) standard. The TEDS provides meta-information regarding sensors such as serial number, calibration curve and operational status. Incorporation of TEDS into wireless sensors leads directly to building application level software that can recognize sensors at run-time, dynamically instantiating sensors as they are added or removed. The Ames Research Center team has been experimenting with this technology building demonstration prototypes for on-board health monitoring. Innovations in technology, software and process can lead to dramatic improvements for managing sensor systems applied to Developmental and Flight Instrumentation (DFI) aboard aerospace vehicles. A brief overview of the plug-and-play ZigBee WSN technology is presented along with specific targets for application within the aerospace DFI market. The software architecture for the sensor nodes incorporating the TEDS information is described along with the functions of the Network Capable Gateway processor which bridges 802.15.4 PAN to the TCP/IP network. Client application software connects to the Gateway and is used to display TEDS information and real-time sensor data values updated every few seconds, incorporating error detection and logging to help measure performance and reliability in relevant target environments. Test results from our prototype WSN running the Mobitrum software system are summarized and the implications to the scalability and reliability for DFI applications are discussed. Our demonstration system, incorporating sensors for life support system and structural health monitoring is described along with test results obtained by running the demonstration prototype in relevant environments such as the Wireless Habitat Testbed at Johnson Space Center in Houston. An operations concept for improved sensor process flow from design to flight test is outlined specific to the areas of Environmental Control and Life Support System performance characterization and structural health monitoring of human-rated spacecraft. This operations concept will be used to highlight the areas where WSN technology, particularly plug-and-play software based on IEEE 1451, can improve the current process, resulting in significant reductions in the technical effort, overall cost and schedule for providing DFI capability for future spacecraft. RELEASED -

Alena, Richard↗

DOE OSTI2024

Smart Pixels: In-pixel AI for on-sensor data filtering

We present a smart pixel prototype readout integrated circuit (ROIC) designed in CMOS 28 nm bulk process, with in-pixel implementation of an artificial intelligence (AI) / machine learning (ML) based data filtering algorithm designed as proof-of-principle for a Phase III upgrade at the Large Hadron Collider (LHC) pixel detector. The first version of the ROIC consists of two matrices of 256 smart pixels, each 25$\times$25 $\mu$m$^2$ in size. Each pixel consists of a charge-sensitive preamplifier with leakage current compensation and three auto-zero comparators for a 2-bit flash-type ADC. The frontend is capable of synchronously digitizing the sensor charge within 25 ns. Measurement results show an equivalent noise charge (ENC) of $\sim$30e$^-$ and a total dispersion of $\sim$100e$^-$ The second version of the ROIC uses a fully connected two-layer neural network (NN) to process information from a cluster of 256 pixels to determine if the pattern corresponds to highly desirable high-momentum particle tracks for selection and readout. The digital NN is embedded in-between analog signal processing regions of the 256 pixels without increasing the pixel size and is implemented as fully combinatorial digital logic to minimize power consumption and eliminate clock distribution, and is active only in the presence of an input signal. The total power consumption of the neural network is $\sim$ 300 $\mu$W. The NN performs momentum classification based on the generated cluster patterns and even with a modest momentum threshold, it is capable of 54.4% - 75.4% total data rejection, opening the possibility of using the pixel information at 40MHz for the trigger. The total power consumption of analog and digital functions per pixel is $\sim$ 6 $\mu$W per pixel, which corresponds to $\sim$ 1 W/cm$^2$ staying within the experimental constraints.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2025

Illuminating the Night: A Survey of Super-Resolution Methods for Nighttime Light Images

Nighttime Light (NTL) images provide critical insights into urbanization, disaster response, and energy consumption. The VIIRS Day/Night Band (DNB) sensor offers high-quality NTL imagery with daily revisit rates, but the available spatial resolution hinders fine-grained accurate analysis. Super-resolution techniques aim to increase the resolution of NTL images, enabling more detailed assessments of infrastructure, light pollution, economic activity, and power outages. However, existing state-of-the-art super-resolution methods designed for natural images struggle with the unique characteristics of NTL data. This work provides a comprehensive review of super-resolution methods across multiple image modalities, evaluates their effectiveness on VIIRS DNB data, and proposes a multi-modal super-resolution approach tailored to NTL imagery. The proposed approach integrates VIIRS DNB data with road networks and land use information to improve reconstruction accuracy and spatial detail. Code is available for this project at https://code.ornl.gov/viirs-sr/sr-demos.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

NASA NTRS2022

TruePAL – An AI Assistant for First Responder Safety

This paper presents the development of an AI assistant, Trusted and Explainable Artificial Intelligence for Saving Lives (TruePAL), to provide real-time warning of risks of potential crashes to the first responders. The TruePAL system employs an AI and deep learning technology for saving first responders and roadside crews lives in and around active traffic. A deep neural network (DNN) and a Non-Axiomatic Reasoning System (NARS) are implemented as an AI system. A mobile app with AI interface is developed to perform verbal communication with the first responders. The TruePAL team has developed an explainable AI approach by opening up the DNN blackbox to extract the activation filters of various features and parts of the targeted objects. The combination of DNN and NARS makes the TruePAL system explainable to the users. TruePAL ingests on-board cameras, radar, and other sensor signals, analyzes the environment and traffic patterns to generate timely warning to drivers and roadside crews to avoid crashes. The TruePAL team, in collaboration with the Miami/Dade Police Dept., has designed five use cases and multiple sub-scenarios in a CARLA driving simulator to test the capability of TruePAL in timely warning to the first responder drivers in potential crash scenarios. We have successfully demonstrated its capability of timely warning in over a dozen scenarios based on the use cases. The preliminary test simulation results show that TruePAL could provide the drivers and crew members advanced warning before a crash occurs.

Chow, Edward↗

DOE OSTI2024

Smart Pixels: In-pixel AI for on-sensor data filtering

We present a smart pixel prototype readout integrated circuit (ROIC) designed in CMOS 28 nm bulk process, with in-pixel implementation of an artificial intelligence (AI) / machine learning (ML) based data filtering algorithm designed as proof-of-principle for a Phase III upgrade at the Large Hadron Collider (LHC) pixel detector. The first version of the ROIC consists of two matrices of 256 smart pixels, each 25$\times$25 µm\textsuperscript{2} in size. Each pixel consists of a charge-sensitive preamplifier with leakage current compensation and three auto-zero comparators for a 2-bit flash-type ADC. The frontend is capable of synchronously digitizing the sensor charge within 25 ns. Measurement results show an equivalent noise charge (ENC) of $\sim$30e\textsuperscript{-} and a total dispersion of $\sim$100e\textsuperscript{-} The second version of the ROIC uses a fully connected two-layer neural network (NN) to process information from a cluster of 256 pixels to determine if the pattern corresponds to highly desirable high-momentum particle tracks for selection and readout. The digital NN is embedded in-between analog signal processing regions of the 256 pixels without increasing the pixel size and is implemented as fully combinatorial digital logic to minimize power consumption and eliminate clock distribution, and is active only in the presence of an input signal. The total power consumption of the neural network is $\sim$ 300 $\mu$W. The NN performs momentum classification based on the generated cluster patterns and even with a modest momentum threshold, it is capable of 54.4\% – 75.4\% total data rejection, opening the possibility of using the pixel information at 40MHz for the trigger. The total power consumption of analog and digital functions per pixel is $\sim$ 6 $\mu$W per pixel, which corresponds to $\sim$ 1 W/cm\textsuperscript{2} staying within the experimental constraints.

Parpillon, Benjamin↗

NASA NTRS2013

Bulk Moisture and Salinity Sensor

Measurement and feedback control of nutrient solutions in plant root zones is critical to the development of healthy plants in both terrestrial and reduced-gravity environments. In addition to the water content, the amount of fertilizer in the nutrient solution is important to plant health. This typically requires a separate set of sensors to accomplish. A combination bulk moisture and salinity sensor has been designed, built, and tested with different nutrient solutions in several substrates. The substrates include glass beads, a clay-like substrate, and a nutrient-enriched substrate with the presence of plant roots. By measuring two key parameters, the sensor is able to monitor both the volumetric water content and salinity of the nutrient solution in bulk media. Many commercially available moisture sensors are point sensors, making localized measurements over a small volume at the point of insertion. Consequently, they are more prone to suffer from interferences with air bubbles, contact area of media, and root growth. This makes it difficult to get an accurate representation of true moisture content and distribution in the bulk media. Additionally, a network of point sensors is required, increasing the cabling, data acquisition, and calibration requirements. measure the dielectric properties of a material in the annular space of the vessel. Because the pore water in the media often has high salinity, a method to measure the media moisture content and salinity simultaneously was devised. Characterization of the frequency response for capacitance and conductance across the electrodes was completed for 2-mm glass bead media, 1- to 2-mm Turface (a clay like media), and 1- to 2-mm fertilized Turface with the presence of root mass. These measurements were then used to find empirical relationships among capacitance (C), the dissipation factor (D), the volumetric water content, and the pore water salinity.

Nurge, Mark↗

NASA NTRS2011

Circuit for Communication Over Power Lines

Many distributed systems share common sensors and instruments along with a common power line supplying current to the system. A communication technique and circuit has been developed that allows for the simple inclusion of an instrument, sensor, or actuator node within any system containing a common power bus. Wherever power is available, a node can be added, which can then draw power for itself, its associated sensors, and actuators from the power bus all while communicating with other nodes on the power bus. The technique modulates a DC power bus through capacitive coupling using on-off keying (OOK), and receives and demodulates the signal from the DC power bus through the same capacitive coupling. The circuit acts as serial modem for the physical power line communication. The circuit and technique can be made of commercially available components or included in an application specific integrated circuit (ASIC) design, which allows for the circuit to be included in current designs with additional circuitry or embedded into new designs. This device and technique moves computational, sensing, and actuation abilities closer to the source, and allows for the networking of multiple similar nodes to each other and to a central processor. This technique also allows for reconfigurable systems by adding or removing nodes at any time. It can do so using nothing more than the in situ power wiring of the system.

Krasowski, Michael J.↗

DOE OSTI2024

"Traffic Control via Connected and Automated Vehicles: An Open-Road Field Experiment with 100 CAVs"

The CIRCLES project aims to reduce instabilities in traffic flow, which are naturally occurring phenomena due to human driving behavior. These "phantom jams" or "stop-and-go waves,"are a significant source of wasted energy. Toward this goal, the CIRCLES project designed a control system referred to as the MegaController by the CIRCLES team, that could be deployed in real traffic. Our field experiment leveraged a heterogeneous fleet of 100 longitudinally-controlled vehicles as Lagrangian traffic actuators, each of which ran a controller with the architecture described in this paper. The MegaController is a hierarchical control architecture, which consists of two main layers. The upper layer is called Speed Planner, and is a centralized optimal control algorithm. It assigns speed targets to the vehicles, conveyed through the LTE cellular network. The lower layer is a control layer, running on each vehicle. It performs local actuation by overriding the stock adaptive cruise controller, using the stock on-board sensors. The Speed Planner ingests live data feeds provided by third parties, as well as data from our own control vehicles, and uses both to perform the speed assignment. The architecture of the speed planner allows for modular use of standard control techniques, such as optimal control, model predictive control, kernel methods and others, including Deep RL, model predictive control and explicit controllers. Depending on the vehicle architecture, all onboard sensing data can be accessed by the local controllers, or only some. Control inputs vary across different automakers, with inputs ranging from torque or acceleration requests for some cars, and electronic selection of ACC set points in others. The proposed architecture allows for the combination of all possible settings proposed above. Most configurations were tested throughout the ramp up to the MegaVandertest.

Lee, Jonathan↗

DOE OSTI2025

Traffic Control via Connected and Automated Vehicles (CAVs): An Open-Road Field Experiment with 100 CAVs

The CIRCLES project aims to reduce instabilities in traffic flow, which are naturally occurring phenomena due to human driving behavior. Also called “phantom jams” or “stop-and-go waves,” these instabilities are a significant source of wasted energy. Toward this goal, the CIRCLES project designed a control system, referred to as the MegaController by the CIRCLES team, that could be deployed in real traffic. Our field experiment, the MegaVanderTest (MVT), leveraged a heterogeneous fleet of 100 longitudinally controlled vehicles as Lagrangian traffic actuators, each of which ran a controller with the architecture described in this article. The MegaController is a hierarchical control architecture that consists of two main layers. The upper layer is called the Speed Planner and is a centralized optimal control algorithm. It assigns speed targets to the vehicles, conveyed through the LTE cellular network. The lower layer is a control layer, running on each vehicle. It performs local actuation by overriding the stock adaptive cruise controller, using the stock onboard sensors. The Speed Planner ingests live data feeds provided by third parties as well as data from our own control vehicles and uses both to perform the speed assignment. The architecture of the Speed Planner allows for the modular use of standard control techniques, such as optimal control, model predictive control (MPC), kernel methods, and others. The architecture of the local controller allows for the flexible implementation of local controllers. Corresponding techniques include deep reinforcement learning (RL), MPC, and explicit controllers. Depending on the vehicle architecture, all onboard sensing data can be accessed by the local controllers or only some. Likewise, control inputs vary across different automakers, with inputs ranging from torque or acceleration requests for some cars to electronic selection of adaptive cruise control (ACC) setpoints in others. The proposed architecture technically allows for the combination of all possible settings proposed previously, that is {Speed Planner algorithms} × {local Vehicle Controller algorithms} × {full or partial sensing} × {torque or speed control}. As a result, most configurations were tested throughout the ramp up to the MegaVandertest (MVT).

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2025

Even Higher-Level Synthesis: An Exploration of AI Hardware Accelerators using HLS4ML

With the rise of artificial intelligence, the popularization of deep learning, and a constantly evolving industry, the demand for flexible and efficient tools has never been greater. As algorithms grow more complex, their runtime and energy consumption increase exponentially. Customized hardware accelerators, long used for specific mathematical operations, remain essential for managing modern applications' computational and power demands. Hardware accelerators can speed up complex computations by orders of magnitude, but their manual design and verification processes are often challenging and time-consuming. High-Level Synthesis (HLS) provides a solution by transforming high-level algorithm descriptions, typically written in C++ or SystemC, into synthesizable RTL suitable for hardware implementation. This approach reduces development time for RTL engineers while offering flexibility beyond what traditional handwritten RTL can provide. We extended this capability to the machine-learning domain with the open-source framework hls4ml, which allows neural networks trained in Python frameworks like Tensorflow or PyTorch to be synthesized into efficient hardware representations for the traditional FPGA and ASIC flows. This breakthrough addresses the growing need for reduced design turnaround and easy verification of ML hardware accelerators with low latency and power efficiency constraints. During this tutorial, we will demonstrate how Python complements HLS by simplifying the ML design process, bridging the gap between software and hardware development. Attendees will explore how we translate neural networks modeled in Python into fixed-point C++ models suitable for HLS workflows. We will dive into strategies like Value-Range Analysis and Quantization-Aware Training, which optimize these designs for deployment and evaluate their accuracy, power consumption, and energy efficiency. To exemplify these concepts, experts from Fermilab will share their experiences applying this technology to high-energy physics experiments, where real-time, low-latency processing is critical. Over the years, Fermilab engineers have demonstrated how deep neural networks, optimized for hardware using hls4ml, can meet the stringent requirements of trigger systems at the CERN Large Hadron Collider. These systems rely on rapid decision-making to process immense data volumes while retaining only the most relevant events for further analysis. The application of hls4ml has also been extended to innovative technologies like smart pixel arrays. These smart pixels integrate ML inference capabilities directly into sensor devices, enabling localized data processing at the pixel level. This approach drastically reduces the need to transmit raw data to external processing units, significantly decreasing power consumption and latency. By embedding neural networks within the pixel architecture, the smart pixels can identify and prioritize relevant data in real time, providing a highly efficient solution for edge computing in scenarios such as particle detectors and imaging systems. Fermilab's work highlights the potential of hardware-accelerated ML in scenarios where both speed and power efficiency are mission-critical. Through this tutorial, attendees will gain valuable insights into the challenges and solutions of deploying ML in hardware. Understanding how HLS and hls4ml streamline the development of neural network-based hardware accelerators is fundamental for the industry's future. Participants will learn how these technologies are shaping the future of AI and scientific computing.

Di Guglielmo, Giuseppe [Fermilab]↗

NASA NTRS2008

Ground-based Network and Supersite Measurements for Studying Aerosol Properties and Aerosol-Cloud Interactions

From radiometric principles, it is expected that the retrieved properties of extensive aerosols and clouds from reflected/emitted measurements by satellite (and/or aircraft) should be consistent with those retrieved from transmitted/emitted radiance observed at the surface. Although space-borne remote sensing observations contain large spatial domain, they are often plagued by contamination of surface signatures. Thus, ground-based in-situ and remote-sensing measurements, where signals come directly from atmospheric constituents, the sun, and the Earth-atmosphere interactions, provide additional information content for comparisons that confirm quantitatively the usefulness of the integrated surface, aircraft, and satellite datasets. The development and deployment of AERONET (AErosol RObotic NETwork) sunphotometer network and SMART-COMMIT (Surface-sensing Measurements for Atmospheric Radiative Transfer - Chemical, Optical & Microphysical Measurements of In-situ Troposphere) mobile supersite are aimed for the optimal utilization of collocated ground-based observations as constraints to yield higher fidelity satellite retrievals and to determine any sampling bias due to target conditions. To characterize the regional natural and anthropogenic aerosols, AERONET is an internationally federated network of unique sunphotometry that contains more than 250 permanent sites worldwide. Since 1993, there are more than 480 million aerosol optical depth observations and about 15 sites have continuous records longer than 10 years for annual/seasonal trend analyses. To quantify the energetics of the surface-atmosphere system and the atmospheric processes, SMART-COMMIT instrument into three categories: flux radiometer, radiance sensor and in-situ probe. Through participation in many satellite remote-sensing/retrieval and validation projects over eight years, SMART-COMMIT have gradually refine( and been proven vital for field deployment. In this paper, we will demonstrate the capability of AERONET SMART-COMMIT in current Asian Monsoon Year-2008 campaigns that are designed and being executed to study the compelling variability in temporal scale of both anthropogenic and natural aerosols (e.g., airborne dust, smoke, mega-city pollutant). Feedback mechanisms between aerosol radiative effects and monsoon dynamics have been recently proposed, however there is a lack of consensus on whether aerosol forcing would be more likely to enhance or reduce the strength of the monsoon circulation. We envision robust approaches which well-collocated ground-based measurements and space-borne observations will greatly advance our understanding of absorbing aerosols (e.g., "Global Dimming" vs. "Elevated Heat-Pump" effects) on aerosol cloud water cycle interactions.

Tsay, Si-Chee↗