USA flag logo/image

An Official Website of the United States Government

DOD/DARPA DOD STTR 2013.B 5

DoD 2013.B STTR Solicitation

Printer-friendly version
Agency: Department of Defense
Branch: Defense Advanced Research Projects Agency
Program/Year: STTR / 2013
Solicitation Number: 2013.B
Release Date: July 26, 2013
Open Date: August 26, 2013
Close Date: September 25, 2013
ST13B-001: Advanced Tools for Mammalian Genome Engineering
Description: OBJECTIVE: Improve the utility of Human Artificial Chromosomes (HACs) by developing new selectable metabolic markers for use in human cells, new high-fidelity methods for inserting DNA constructs of at least 50,000 base pairs (bp) in length into defined genomic loci, and new methodologies for facile intercellular genome transplantation. DESCRIPTION: The ability to deliver exogenous DNA to mammalian cell lines is a fundamental tool in the development of advanced therapeutics, vaccines, and cellular diagnostics, as well as for basic biological and biomedical research. Current approaches to genetic engineering of mammalian cells rely on gene transfer methods such as plasmids, adenovirus-, lentivirus-, and retrovirus-vectors, cDNA, and minigene constructs. While these tools do provide the basic ability to deliver DNA to mammalian cells, there are several shortcomings associated with these state-of-the-art techniques. These include random DNA insertion into the host genome, variation in stable integration sites between cell lines, variation in the copy number and expression level of DNA that is delivered, limitations on the number and size of DNA constructs that can be delivered, and immunological responses to foreign DNA. Coupled with the significant time that is required to obtain useable engineered cell lines, these factors severely limit the scale and scope of research that can be performed and the applications that can be pursued. One recently developed method for gene transfer that has the potential to address many of these shortcomings is the use of human artificial chromosomes (HACs). HACs possess several ideal properties, including very large DNA delivery capacities, stable, episomal maintenance within the cell, and lack of immunogenicity. Additionally, HACs can be designed to contain specific DNA sequences, such as integration sites, making them ideal for the creation of a completely engineerable platform. Although HACs show significant potential as a gene delivery vehicle, several technical hurdles remain that have prevented wide adoption of the technology. First, while HACs have the capacity to contain extremely large segments of DNA (potentially up to or surpassing 1,000,000 bp), currently molecular biology techniques are limiting in the amount of DNA that can be inserted into a DNA vector. It is typically difficult to insert more than 20,000 bp of DNA into a vector, negating much of the advantage that HACs possess as a delivery platform. Second, few selectable markers exist that are suitable for use in human cell lines, limiting the ability to screen for insertion or maintenance of the delivery platform. Third, methods utilized to transfer HACs between cell lines for vector delivery are extremely technically challenging, requiring highly specialized knowledge in order to be able to work with existing HAC vectors. This solicitation is focused on improving the utility of HACs as a DNA delivery platform by developing technologies to address several key technical hurdles associated with current HAC vectors. These includes development of new selectable metabolic markers suitable for use in human cell lines, new high-fidelity methods for inserting DNA constructs at least 50,000 bp in length into defined genomic loci, and new methodologies for facile intercellular genome transplantation. A successful technology will be able to integrate into existing HAC vectors and will be capable of being readily transitioned to academic, government, and commercial researchers, all of whom rely on the ability to deliver DNA to mammalian cell lines. PHASE I: Determine the technical feasibility of a new approach that focuses on addressing ONE of the following technical challenges: a) The development of at least 3 new selectable markers based on metabolism for use in delivery of a HAC. Each metabolic marker should allow for the identification and selection of cells that have been transfected with the HAC vector, and should be broadly useful across human cell lines with minimal or no genetic manipulation of the host cell line required. Appropriate metabolic markers should be identified, methods for genetic selection should be detailed, and an analysis of cost for selection and maintenance of cell lines using each marker should performed. b) The ability to stably integrate at least 50,000 bp of DNA into at least 10 different targeted genomic loci. Methods for insuring single copy integration at each site should be described and the ability to test for proper integration site targeting and specificity should be addressed. Performance goals of the new approach for maximum size limit of DNA that can be integrated and the percentage of correct cells that are achieved per transformation event should be established. c) The development of a methodology for the rapid and facile shuttling of chromosomes or chromosome-sized DNA (e.g. HACs) between at least 10 different cell lines. The possibility of truncation, rearrangement, or fragmentation during chromosome transfer should be addressed and risk mitigation strategies described as appropriate. Performance goals for the frequency of chromosome transfer should be established. For the selected challenge, develop an initial concept design and describe an approach for transitioning this technology from a laboratory benchtop to an established commercial protocol. PHASE I deliverables will include: a technical report detailing experiments and results supporting the feasibility of the approach, and defined milestones and metrics as appropriate for the selected technical challenge. Also included with the PHASE I deliverables is a PHASE II plan for transitioning initial proof-of-concept experiments to protocols that are sufficiently robust and reproducible that they are viable as commercial technologies. The plan should include a detailed assessment of the potential path to commercialization, barriers to market entry, and collaborators or partners identified as early adopters for the new system. PHASE II: Finalize the experimental approach from PHASE I and initiate the development and production of the technology to address the selected technical challenge. Establish appropriate performance parameters through experimentation to determine the efficaciousness, robustness, and fidelity of the approach being pursued. Develop, demonstrate, and validate the reagents and protocols necessary to meet the key metrics as defined for the selected technical challenge. PHASE II deliverables include a prototype set of reagents, a detailed technical protocol sufficient to allow replication of results in an outside laboratory, and valid test data, appropriate for a commercial production path. PHASE III: The successful development of technologies for rapid introduction of large DNA vectors into human cell lines will enable the ability to engineer much more complex functionalities into human cell lines than are currently possible. This capability may support a number of DoD challenges, including the development of complex, multifunctional cell-based sensors for chem/biodefense applications, the simultaneous encoding of thousands of prophylactic or therapeutic antibodies for on-demand production of next-generation disease prevention and treatment, and the creation of complex cell lines that can be rapidly reconfigured to produce large volumes of a given vaccine. The biotechnology and pharmaceutical sectors are heavily reliant on the ability to rapidly manipulate and introduce DNA into human cell lines. The successful development of technologies that allow for improved human cell line generation has significant potential to rapidly transition to commercial use, enabling biologically based production of new protein-based therapeutics, new systems for vaccine development and production, and new platforms for small molecule drug screens that provide a more specific and sophisticated testing environment. Many of these applications are currently inaccessible due to the limitation of existing DNA delivery technologies and have the potential to be transformative if the technologies described herein are developed.
ST13B-002: Quantum Dot Mid-Wave Infrared Focal Plane Array
Description: OBJECTIVE: Develop a mid-wave infrared (MWIR) focal plan array (FPA) using quantum dots for next-generation night vision. DESCRIPTION: Historically, night vision has provided the United States Armed Forces with an asymmetric tactical advantage in combat operations. However, the tradeoffs of low sensitivity (microbolometers), high power consumption (active cooling), or specialized consumables (liquid-nitrogen cooled HgCdTe) are a major technological hurdle to achieving low-power, low-cost and portable thermal night vision imaging. Quantum dots have seen gradual improvements in reducing the band gap in recent years, making a highly-efficient and lower-cost detector material within the thermal infrared (IR) range potentially realizable [1]. High-efficiency of light detection from quantum dots results from the large extinction coefficient induced by quantum confinement [2], creating high-sensitivity without the need for external cooling, and thus reducing weight, size and power consumption. Current epitaxially-grown IR camera detectors cost>$10,000, while the projected cost of quantum dot-enabled detectors is $100, which could enable wider deployment of night vision technology to warfighters as well as low-power surveillance units. Beyond decreased detector cost, quantum dots may be amenable to facile fabrication techniques, such as spin coating, which could further decrease device costs [3]. Significant technical and market challenges exist in transitioning these recent, laboratory-quality quantum dots into device-ready materials that can be fabricated into focal plane arrays; material quality must be improved to enhance intrinsic and extrinsic quantum efficiency; manufacturing scalability and batch-to-batch consistency must be demonstrated; integrating wet solution processing methods with the control and readout structure; fabrication of a detector of comparable size to a commercial focal plane array. Proposers are free to formulate any approach that will contribute to the goal of making a low-cost, high-sensitivity MWIR FPA based on quantum dots. PHASE I: Develop a design plan to fabricate and incorporate quantum dots into a focal plane array. The device must conform in size, shape and power requirements to existing commercial infrared focal plane array. Characterize the expected performance of this focal plane array including spectral sensitivity, resolution and projected cost. Show the feasibility of one or more critical elements of this approach through a lab demonstration. PHASE I deliverables will include a design review simulating device performance and a report presenting plans for PHASE II. PHASE II: Construct and demonstrate the operation of a prototype quantum dot focal plane array validating the device performance outlined in PHASE I. The Transition Readiness Level to be reached is 5: Component and/or bread-board validation in relevant environment. PHASE III: High-sensitivity quantum dot-based solutions will enable widespread deployment of night vision sensors across many platforms including small UAVs, helmet-mounted sensors, night vision goggles, security cameras, guided missile platforms and personnel vehicles. These low cost sensors will help maintain a tactical nighttime operations advantage. Commercial applications include the development of low-cost infrared cameras for private security and automotive applications.
ST13B-003: Multiferroic Materials for RF Applications
Description: OBJECTIVE: Demonstrate RF/microwave devices, components, and circuits based on multiferroic composite structures. Design discrete devices for radio and radar with a new tunability feature that adds to the performance over conventional RF/microwave components by leveraging the voltage-tunable frequency response of multiferroics. Demonstrate voltage tunable devices with performance equal to or better than state-of-the-art circuits. DESCRIPTION: Multiferroic composites demonstrate a unique ability to control their ferromagnetic resonance (FMR) by applying an electric field that causes a shift in their FMR frequency. Multiferroic materials consist of strain-coupled ferromagnetic and ferroelectric phases resulting in a magneto-electric coupling between the two materials. This magneto-electric coupling mechanism has an electrostatically-controllable magnetization, a feature that can enable an entirely new family of RF components and circuits where frequency and bandwidth are electrically tunable. RF/microwave multiferroic components such as tunable filters, duplexers, isolators, antennas and phase shifters are just a few examples of drop-in replacements, although entirely new devices, components and architectures are conceivable. Tunability is a feature added to the normal performance of an RF component. Multiferroic components with voltage tunability should perform at least as well as its non-multiferroic counterpart. In this way, tunability becomes a feature added to the RF/microwave designer"s toolbox. Frequency agile devices have been fabricated from many combinations of electrostrictive and magnetostrictive materials. However these devices have tended to be bulky, slow, consume excess energy, and perform within a narrow band of frequencies. For instance, BST-based tunable filters operating in the wireless 1.6-2.0GHz band have published results of 10dB return loss and 4dB insertion loss with a 25% tuning range using a rather large 0-200VDC tuning voltage. X-band (8-10GHz) BST tunable filters have demonstrated 15dB return loss, 8dB insertion loss, and a 23% tuning range using 0-90VDC. By comparison, commercially-available, off-the-shelf filters offer better than 3dB insertion loss and 20dB return loss. Multiferroic device performance must exceed current commercial capabilities to economically viable while providing new capabilities for more demanding Defense applications which may have to deal with countermeasures. Multiferroic solutions having a highly-tunable ferromagnetic resonance frequency should offer>70dB dynamic range in addition to meeting commercial component performance. Multiferroic phase shifters could conceivably have -180 to +180 degrees of phase shift tunable from 900 MHz to 6 GHz with an instantaneous bandwidth of 20 MHz. The availability of such components would dramatically transform the approach towards designing military RF/microwave radios and radars while also advancing the competitiveness of the U.S. electronics industry. Highly innovative and creative approaches, concepts and solutions based on multiferroic composites and their unique features are especially sought. It is highly desirable to develop manufacturing processes that enable commercial adoption of multiferroic devices by commercial vendors. A commercialization plan including adoption by RF/microwave component suppliers is encouraged. Recent advancements in the fabrication of high-quality ferrimagnetic-ferroelectric stacks have made this the time to commercialize multiferroic RF components. PHASE I: Demonstrate a proof-of-concept multiferroic-based voltage tunable device used as the basis of a building block for a radio or radar. The device should be optimized for realistic radio or radar applications (voltage tuning range, frequency tuning range, power handling, quality factor, temperature coefficient, etc.) and should provide maximum tunability in the military frequency band assignments of the communications or radar spectrum. In addition to achieving conventional state-of-the-art performance, the design should provide for frequency tunability using a low-voltage (~12 volt) DC power supply. The physical design and fabrication shall be based on multiferroic composites with accompanying modeling and analysis to design physical layout and prediction of the high-frequency performance. PHASE I deliverables will also include development of an initial concept design and RF models of the key aspects of the device. Elements of the multiferroic design may be bread-boarded and measured as validation of detailed analysis of the predicted RF performance. The project plan should define and develop key technological milestones such as performance modeling and simulation of the device. PHASE II: Fabricate and demonstrate an operating prototype of a multiferroics-based building block used in a radio or radar architecture. For example, the building block may be a tunable filter for a radio, a phase shifter for a radar set, or a similar 50 ohm component. It should build on the device developed in PHASE I, or use as its foundation the materials processing developed in PHASE I. A detailed plan of action to design, fabricate and assemble all the necessary RF components should be provided and followed by testing and characterization of the RF building block. The results of the testing should be used to update the multiferroic design, modeling and simulation tools. The RF performance parameters to be expected from multiferroic materials should be established through experiments performed on the prototype. Circuits of interest include, but are not limited to phase shifters (-180 to +180 degrees, tunable in the band 900 MHz to 6 GHz) and filters with high Dynamic Range and broad tunability (>70 dB dynamic range, 50% of center frequency). Broad band piezoelectric amplifiers (Reference 14) with composite microstructures (reference 15) with efficiency of greater than 60%; and compact B-field antennas are also components of interest for this topic. For devices utilizing the tunable response of multiferroics for other proposed devices/circuits (i.e.: broad band amplifiers and compact antennas) comparable quantitative metrics must be stated. At the conclusion of PHASE II effort, the building block prototype should be at a Technology Readiness Level 5 (TRL-5) or above. PHASE III: Millitary applications of this technology include radios and radar systems. Voltage tunable inductors allow design of frequency agile circuits while maintaining constant impedance. Specific radar applications include ground penetrating applications and low frequency arrays for airborne use. The technology represents radical innovation for the wireless communication industry as well as radar systems. The U.S. commercial electronics community could benefit from a multiferroics fabrication process that is robust enough to be adopted for manufacturing. This would ultimately lead to monolithic integration of multiferroic materials with conventional silicon semiconductor processes. The multiferroic devices, components, circuits, and architectures proposed should have a development path leading to commercial adoption by mainstream RF component suppliers such as Mini- Circuits, Digi-Key, Newark, and similar established distributors. Such a development path would clearly show that multiferroics technology is ready for commercial electronics use.
ST13B-004: Data-Parallel Analytics on Graphics Processing Units (GPUs)
Description: OBJECTIVE: Explore the space of data-centric problems and algorithms that lend themselves to high-performance implementation on GPUs; develop a high-level language for easy programming of GPUs; and develop a product that can support real-time, quantitative analysis of a wide variety of data using the cost and energy efficient compute capabilities of GPUs and other relevant many core architectures. DESCRIPTION: DOD is interested in exploiting trends in commercial technology for improved data analytics to provide improved real-time situation awareness capabilities essential to effective war fighting or for disaster response. Commercial trends indicate that CPU clock"s speeds have been essentially flat for the last decade. Continued speedups are only possible through many-core and distributed computing. General Purpose Graphics Processing Units (GPGPUs) are high-performance, general purpose, data-parallel processing architectures. Today"s GPUs offer Teraflops speeds over thousands of stream cores [1]. Even in high end supercomputers such as Titan [2], much of the Petaflop speed is derived by GPUs hosted in its many compute nodes. Even mainstream CPUs feature increasingly powerful GPU components [11-12]. GPUs have been highly successful in gaming technology and have become nearly universal in mobile devices, tablets, and laptops in their role as graphics processors and GIS platforms. However, high development costs, limited interoperability with existing software ecosystems, and limited access to expert programmers have restricted the use of GPUs by a wider audience. Further, algorithms and data with non-local memory access patterns, such as graph processing, require special treatment to create efficient algorithms with coalesced memory access patterns on the GPU. Industry has recognized the need for scalable, easy-to-use data-centric platforms as key drivers for innovation. This has led to a variety of scalable, open-source, and data-centric architectures. Tools such as Map/Reduce and Pregel (graph processing) enable scalability, but are not always efficient. Other approaches to scalable data-centric architectures offering high-level abstractions include programmable pipelines and databases. While expressing a desired analytic maybe easier in a given framework, there is often a best-in-class algorithm that could be orders of magnitude more efficient than a general purpose approach, for a given underlying hardware. Further, different algorithms may be required for GPU and CPU hardware. An opportunity exists to explore the relationship between these data-centric abstractions, and ways in which optimal programs could be derived. A number of efforts are underway exploring how GPUs can be integrated into general purpose programming languages and environments [5-9]. However, these approaches fail to systematically accommodate the core algorithms that bring out the best performance of the underlying hardware and put programmers in the position of attempting to optimize for a complex technology outside of their expertise. This STTR seeks research on exploring and characterizing the space of data-centric problems, and analytics and algorithms that lend themselves to high-performance implementation on GPUs. In addition, research is sought in developing a high-level language for easy programming of GPUs. Challenges include design of code from the outset to primarily run on the GPU (e.g., prefix scan, compact, allocate, hash tables [3]); identifying high-level data-centric abstractions and development of a language to express them; and, finally, translating user programs to execute efficiently on underlying hardware with runtime workload. Research must address choice of configurations to optimize for latency, throughput, or tradeoff completeness for characterized accuracy. Also of interest is the division of computation between device and cloud when deploying in environments with limited connectivity, bandwidth, or power. Finally, the STTR seeks development of a product that can support real-time, quantitative analysis of a wide variety of data using the cost and energy efficient compute capabilities of GPUs and other relevant many core architectures. An example data-centric problem is the space of graphs where GPUs are not traditionally thought to be relevant due to the data dependent scheduling of threads and difficulties in identifying redundant work without sacrificing throughput. However, recent research [4] has shown that graph traversal can be parallelized efficiently on the GPU. Earlier research [11] has demonstrated that Map/Reduce may be realized efficiently on GPUs, but extensions are required to handle some problems efficiently. Effort is required to identify tradeoffs among different highly parallel architectures to understand how work can be efficiently scheduled on different architectures, to identify key primitives for data-parallel hardware and abstractions that will increase the reusability of data parallel algorithms, and to translate high level user programs to efficient, low-level and hardware specific algorithms. The proposed approaches should provide significant advantages over current technologies: 1) technology enabling a widely available, low-cost, data-centric computing infrastructure to allow widespread use of data analytics in small and large enterprises, 2) ability to write new analytics easily to run efficiently on highly parallel and distributed systems, and 3) ability to easily incorporate new hardware and new algorithms into existing analytics. PHASE I: Develop an initial concept design and model key elements for a data-centric processing architecture leveraging GPUs and many-core computing platforms. Identify a set of data-centric problems that are likely to be efficient on GPUs, and implement a set of algorithms characterizing their structure and its relation to efficiency gains on a given underlying hardware. PHASE I deliverables will include a technical report and brief describing the plan of approach and key technological achievements for the development of a prototype system. PHASE II: Develop a high-level language for GPU programming, and implement translation of user programs to execute efficiently against the underlying hardware and the actual runtime workload. Construct and demonstrate the operation of a prototype in an operationally relevant environment. In parallel to this development, develop, test, and demonstrate validity and generalizability of GPU accelerated analytics for multiple applications. Required PHASE II deliverables will include the prototype and examples of GPU accelerated applications based on the prototype, and a technical report and brief describing 1) the system design and test results, 2) sample applications, 3) and feasibility of use in future commercial and/or military applications. PHASE III: A portable device offering low-cost and high-performance on a wide range of data-centric analytics would benefit the military broadly. Soldiers with access to improved analytics on the battlefield would have improved real-time situation awareness capabilities essential to effective war fighting or evolving disaster response scenarios. Situated analytics could synthesize data from a unit and the operational theatre, delivering answers and analytics that bear on the immediate problem of the war fighter. GPU accelerated analytics could improve the capabilities of reach back information and analysis systems, providing faster response times and handling larger workloads. The increased compute per square foot, compute per pound, and compute per watt would benefit resource-constrained environments. Many commercial entities would have interest in a low-cost, high-performance, easily customized, scalable analytics. Potential marketplace applications exist in marketing, gaming, medicine, and many related fields.