MODIFICATION
A -- Defense Sciences Research and Technology
- Notice Date
- 2/28/2005
- Notice Type
- Modification
- NAICS
- 541710
— Research and Development in the Physical, Engineering, and Life Sciences
- Contracting Office
- Other Defense Agencies, Defense Advanced Research Projects Agency, Contracts Management Office, 3701 North Fairfax Drive, Arlington, VA, 22203-1714
- ZIP Code
- 22203-1714
- Solicitation Number
- BAA05-19
- Response Due
- 2/7/2006
- Archive Date
- 2/8/2006
- Description
- ROBUST UNCERTAINTY MANAGEMENT, SOL BAA05-19, Addendum 2, DUE 052305, POC: DR. CAREY SCHWARTZ, DARPA/DSO, Ph: (571) 218-4536. Email: baa05-19@darpa.mil, URL: www.darpa.mil/dso. Website Submission: http://www.sainc.com/dso0519/ The Defense Sciences Office is interested in new proposals for Robust Uncertainty Management (RUM). The RUM Program will emphasize the development of mathematics and algorithms that guarantees performance by design for complex large-scale nonlinear federated systems. The Department of Defense will often integrate a number of large, nonlinear subsystems to form a single system whose purpose is to perform a specific function. The total system is composed of subsystems that are highly interconnected, simultaneously feature weak and strong nonlinear couplings, and are subject to resource constraints. Typically the subsystems exploit different physical properties and mechanisms, and the subsystems will be governed by their own nonlinear dynamics across multiple temporal and spatial scales. By its nature, the system is federated, distributed, and subject to uncertainty arising from incomplete knowledge of initial conditions, reducible errors arising from sensors, irreducible errors derived from modeling the physical system, as well as exogenous disturbances with an unknown distribution. The primary issue is then to guarantee performance of the overall system and provide bounds on the performance of the system in the presence of epistemic and aleatory errors by design. Guaranteeing performance is to be understood to mean that in the presence of partial or incomplete knowledge of the parameters or model describing the system that the values of specific state variables of the federated system can be predicted and maintained within a specified bound. In particular the state variables of interest will be directly associated with the observable metrics of performance for the system. The goal of the RUM program is to develop new mathematical tools that enable the design of large federated nonlinear systems with guaranteed performance in the presence of epistemic, aleatory, a priori, and a posteriori errors. Intellectual effort should be directed toward a solution of the challenge problem that is described below. The solution must be achieved by developing and employing a suite of new mathematical tools that enable analysis of the system, as well as incorporation of the analysis into a design for the system that guarantees performance. These new methods shall be applied to either variation of the following cooperative surveillance challenge problem: Consider N airborne sensors restricted to level flight at fixed altitude H. Each sensor maneuvers in the altitude plane H and has a footprint on the ground of area A. The N sensors need to perform surveillance of an area L squared with L >> NA. It is further assumed that the sensors have a known constant probability of detection, pd and probability of false alarm, pfa against a single target class which is specified. The sensors can determine if they are over terrain for which their probability of detecting a target is zero (i.e., an IR or visible sensor operating over foliage). The challenge problems are for a fixed set of nonlinear dynamics between the surveillance sensors, what is the minimal connectivity between the sensors and the minimal time to guarantee surveillance of the area L squared at a specified Pd > pd and Pfa < pfa assuming a noisy communications channel while guaranteeing no collisions between the sensors. The second variation of the challenge problem assumes a fixed noisy communication network and the objective is to define the interactions among the sensors and the dynamics of the sensors themselves to achieve the minimal time for surveillance while guaranteeing detection performance, i.e., a specified Pd > pd and Pfa < pfa, with no collisions between assets. We would like to invite white papers and proposals that are far reaching in their implications, innovative, and ambitious in their goals and implementation. We encourage the submission of white papers and proposals that provide solutions to one or both of the challenge problem variations. It is anticipated that to provide these solutions, research efforts will have to address one or more of the following areas: 1. New mathematical methods that will enable the development of nonlinear multiscale spatiotemporal models of physical systems that exhibit nonlinear dynamics. We are specifically encouraging efforts that provide a methodology for the design of experiments in coupled nonlinear systems to support the development of models and which can characterize the effects of unmodeled dynamics upon the outputs of the model. We strongly discourage methods that do not exhibit linear scaling with the number of parameters in the model. 2. New methods that reduce the computational complexity of nonlinear models resulting in a nonlinear model with less complex dynamics and for which the impact of reducing the model can be quantified by guaranteeing the performance of the system as well as providing bounds on the performance of the system as a result of simplifying the model. 3. We are also interested in computational strategies that will minimize the computational costs associated with the evaluation of a nonlinear model while maintaining an appropriate precision in control or optimization applications. 4. Analysis methods and computational techniques that enable the propagation of uncertainty through coupled nonlinear systems. In particular these new methods must be capable of accounting for branching in the dynamics, accounting for non-Gaussian probability distributions associated with parameters of the models, sensor noise, and initial conditions. 5. New numerical methods for the estimate of the probability distribution function for the outputs of coupled nonlinear dynamical systems on multiple spatiotemporal scales that are faster and more computationally efficient than Monte-Carlo simulation or particle filter methods for a given non-Gaussian distribution of input parameters, model uncertainty, and reduction of the dynamics. 6. Methods to control and, if required, optimize the performance of large federated nonlinear systems with dynamics on multiple spatiotemporal scales. We strongly discourage white papers and proposals that rely on methods that attempt to linearize the dynamics and then apply established methods such as a Kalman Filter. We are also discouraging white papers and proposals that attempt to solve the challenge problems by application of known nonlinear control methods and that do not involve the development of new mathematical methods. RUM is envisioned as a three-phase program that leads to the solution of the challenge problem by developing new methods for the solution of coupled nonlinear dynamical systems in the face of uncertainty. Each phase of RUM will consist of a twelve-month period of performance. In the first twelve-month phase of RUM, the focus will be the development of new methods focused upon the challenge problem variants. At the conclusion of the first phase we expect to test the new methods by simplifying the challenge problem to a large coupled system that has linear dynamics and Gaussian distributions associated with the probability distribution functions of all uncertainties. At the conclusion of Phase 1 the efficacy of the new methods will be verified by comparing their predicted results against standard techniques for these problems. The second twelve-month phase may include a down-select of Phase 1 efforts. The down-selection criteria may include success in the verification process described above. The focus of Phase 2 will be in refining and demonstrating that the new methods developed for the challenge problem can be applied to the special case of a large linear system with non-Gaussian distributions of input parameters and produce results that are consistent with Monte Carlo or particle filtering methods but at reduced computational costs. The final phase will consist of 12 months and may include another down-select based partially upon success in verifying performance at the conclusion of Phase 2. During the third phase, the methods developed will be applied to the challenge problems leading to a design with performance guarantees in the presence of uncertainty, demonstration of design of experiments to produce complex nonlinear models, and the application of the design of experiments methodology to produce reduced complexity models whose effect upon the design can be assessed in the face of sensor errors and uncertainty in initial conditions. Authors of white papers and proposals may include an optional task in which the system design can be tested experimentally. WHITE PAPER REQUIREMENTS. White papers must lead to a solution of either variant of the challenge problem and the solution must be based upon the development and application of new mathematics. We strongly discourage white papers that focus on the design of new sensors, data collection efforts, linearization of the system, or application of known nonlinear control methodologies to address the challenge problem. We are also strongly discouraging white papers and proposals that focus on experimentally determining conditions that lead to a solution of the challenge problem. We encourage the formation of interdisciplinary teams integrated toward solutions to these challenge problems. It is essential that the preparation of white papers include the following areas: 1. A clear statement of the envisioned utility of the proposed research and development. We are looking for revolutionary applications and goals that could be enabled if the proposed work is completed successfully. The vision for the program should be long term and may exceed the period of performance while the goals and deliverables should reflect the actual period of performance; 2. A concise statement of the research challenges, approach, and potential anticipated solutions to the challenges that will be addressed. This should include explicit timelines for which progress toward the goals can be determined and measured. Intermediate milestones of approximately 12-month periods with demonstrable metrics of performance must be included for the proposed work; 3. A cost estimation for resources required for the proposed timeline; and 4. The white paper should consider phases of development as the challenges are met. White papers must be received by 1600 ET March 22, 2005. Please put the phrase ?Robust Uncertainty Management? in the title of the white paper. If the authors of white papers choose not to submit electronically, U.S. mail may be used. White papers will not be accepted by way of facsimile transmissions. Authors of white papers will be notified by April 22, 2005, if a full proposal will be requested. Full proposals must be submitted no later than 1600 ET May 23, 2005. To facilitate the submission of white papers, a website http://www.sainc.com/dso0519/ has been set up. For more detailed instructions on submitting white papers, please refer to the instructions for BAA05-19 found at the website http://www.darpa.mil/baa/baa05-19pt1.html. Not withstanding the disposition of white papers, DARPA will accept full proposals for this addendum. PROPOSAL REQUIREMENTS. Each proposal should: 1) explicitly address tests, demonstrations, and other research activities planned in the area(s) of interest described above; 2) include specific and quantitative scientific and/or technical objectives for each scientific/technical area of interest, for each phase of the program, addressed in the proposal that clearly demonstrate the research is on track for meeting the ultimate program goals; 3) include clearly delineated intellectual property arrangements and transition paths; 4) include identification and assessment of critical scientific and/or technical barriers to the program objectives and plausible approaches to develop solutions or overcome their limiting effects, and 5) a single viewgraph that concisely describes the; a) technical approach and research activities that are planned in the area(s) of interest; b) specific technical and quantifiable scientific objectives for each phase of the program; c) identification and assessment of critical scientific and/or technical barriers to the program objectives; and a; d) descriptive graphic. The viewgraph must be submitted electronically in either Adobe pdf or Microsoft PowerPoint format. Upon award, specific deliverables and appropriate level demonstrations of the science and/or technology elements will be required periodically and a final demonstration of the deliverable system is required at the end of the program. Proposed Phase 1 efforts should not exceed 12 months. Phase 2 and Phase 3 efforts should be planned for 12-month periods of performance, respectively. If multiple awards are made, down-selection may occur annually based on technical progress and achievements. Proposals with cost share should clearly identify the specific tasks to be cost shared in the technical proposal and separately break out the corresponding costs in the cost proposal. The number of awards will be dependent on the suitability of proposals received and availability of funds. Full proposals shall consist of two volumes: technical and cost. The technical and cost volumes shall conform to the guidelines in DARPA (DSO) BAA 05-19 of February 8, 2005. To receive consideration under this addendum PROPOSALS ARE DUE NO LATER THAN 1600 ET May 23, 2005, to the address shown below. Proposals received after that date will be considered under the open BAA but not this addendum. EVALUATION OF PROPOSALS. Evaluation of the proposals will be in accordance with BAA05-19. For general administrative questions, please refer to the original FedBizOpps announcement, BAA 05-19 of February 8, 2005. GENERAL INFORMATION: In all correspondence, reference BAA05-19, Addendum 2. Technical Point of Contact. Dr. Carey Schwartz, DARPA/DSO; Phone: (571) 218-4536; Fax: (571) 218-4553. Original Point of Contact Brett Giroir, Deputy Director, Defense Sciences Office, Phone 571-218-4224, Fax 571-218-4553, Email bgiroir@darpa.mil.
- Record
- SN00759848-W 20050302/050228212434 (fbodaily.com)
- Source
-
FedBizOpps.gov Link to This Notice
(may not be valid after Archive Date)
| FSG Index | This Issue's Index | Today's FBO Daily Index Page |