N101-022 TITLE: Antenna Placement Optimization on Large, Airborne, Naval Platforms

TECHNOLOGY AREAS: Sensors, Electronics, Battlespace

ACQUISITION PROGRAM: PMA-290, Maritime Patrol and Reconnaissance Aircraft; PMA-265, Super Hornet

RESTRICTION ON PERFORMANCE BY FOREIGN CITIZENS (i.e., those holding non-U.S. Passports): This topic is "ITAR Restricted." The information and materials provided pursuant to or resulting from this topic are restricted under the International Traffic in Arms Regulations (ITAR), 22 CFR Parts 120 - 130, which control the export of defense-related material and services, including the export of sensitive technical data. Foreign Citizens may perform work under an award resulting from this topic only if they hold the "Permanent Resident Card", or are designated as "Protected Individuals" as defined by 8 U.S.C. 1324b(a)(3). If a proposal for this topic contains participation by a foreign citizen who is not in one of the above two categories, the proposal will be rejected.

OBJECTIVE: Port highly developed, high-frequency, serial antenna analysis codes to latest technology computer clusters in order to significantly reduce time in analyzing on-platform antenna performance and antenna-to-antenna interaction.

DESCRIPTION: Modern naval aircraft can be large in dimensions and may carry a large number of antennas. A good example is the Navy�s P-8A Poseidon aircraft [1], a Boeing 737 that is roughly 40 meters long and has a wingspan of about 34 meters. This aircraft carries over 100 antenna systems. For many of these systems, the surface area of the platform is in the tens of thousands of square wavelengths. In this case, the use of full-wave solvers to assess the on-platform performance of an antenna or the interaction between two antennas is impractical, both in terms of computing resources required and length of execution time. The next best choice is to use a high-frequency code. Although not as accurate as full-wave codes, high-frequency codes require modest computer resources and are faster than full-wave codes. In a serial mode, however, even these codes can take substantial time to execute depending on platform size and complexity. This is especially true when considering the on-platform coupling between two antennas if there is a large number of two-antenna combinations. If we are to optimize antenna performance and minimize its interaction with a number of other antennas, then the most cost-effective way to proceed is to port serial, high-frequency codes to clusters of parallel computers. This will improve execution time by orders of magnitude, thus reducing idle time and lost momentum in the workplace.

High-frequency codes are ideally suited to parallelization. The hardware for such an effort can be a traditional central processing unit (CPU) cluster [2] or a graphics processing unit (GPU) cluster [3]. We are favoring the GPU solution both because of the Flops/dollar advantage and because of the recent introduction of compute unified device architecture (CUDA) [4], a language that greatly facilitates programming a GPU. Researchers are already using GPU clusters for a variety of problems [5] and GPU-based hardware is already in the marketplace [6]. We are also interested in CPU clusters since we already own one. With the above in mind, we are seeking innovative solutions for porting high-frequency computational electromagnetic codes to both CPU and GPU-based parallel environments for the purpose of greatly accelerating their performance. These codes must have the capability of assessing antenna performance on large and complex platforms; they also must be able to handle in-situ coupling between antennas; additionally, it is highly desirable that they have a radar cross-section (RCS) calculation capability. Small businesses must clearly demonstrate the capabilities of their high-frequency code in their proposal. They should also have an understanding of GPUs and CUDA and be prepared to work in both a CPU and a GPU environment. Previous experience in programming GPUs is highly desirable. Teaming between electromagnetics and computer experts is also encouraged.

PHASE I: Develop a detailed description of the algorithms from an existing high-frequency solver that would need to be modified to run on a CPU and a GPU-based parallel computing architecture. Identify existing algorithms that may be problematic in transferring to a parallel environment and suggest modifications. Identify existing algorithms that can be improved upon to provide better answers, modify accordingly and test. Perform a study to estimate whether porting the code to both types of environments is feasible within the Phase II timeframe. Develop specifications for a GPU cluster and perform a market search for cluster. Develop a Phase II implementation plan for a CPU and a GPU cluster. Identify other hardware acceleration techniques that could potentially be developed during the Phase II effort.

PHASE II: Purchase test-size GPU cluster identified in Phase I. Use it and existing NAVAIR CPU cluster to port the algorithms identified in Phase I. Validate successful implementation of the parallelization through timing and accuracy studies on electrically very large problems. Ensure that the resulting algorithms are scalable with increasing number of processors. Deliver, install, and provide training for the parallelized high-frequency solver to NAVAIR along with thorough documentation. If NAVAIR is interested in other hardware acceleration techniques identified during Phase I, implement prototype capabilities during the Phase II effort.

PHASE III: Deliver, install, and provide training for the parallelized high-frequency solver to NAVAIR along with thorough documentation.

PRIVATE SECTOR COMMERCIAL POTENTIAL/DUAL-USE APPLICATIONS: The technology developed under this topic can be used in the commercial communications industry, including antenna design and placement, platform integration, electromagnetic compatibility (EMC) and electromagnetic interference (EMI).

REFERENCES:
1. http://www.boeing.com/defense-space/military/p8a/index.html

2. http://en.wikipedia.org/wiki/Computer_cluster

3. http://www.gpgpu.org/

4. http://www.nvidia.com/object/cuda_home.html

5. http://www.cs.sunysb.edu/~vislab/projects/urbansecurity/GPUcluster_SC2004.pdf

6. http://www.amax.com/TeslaPSC-1.asp?gclid=CNKc6M-Qh5kCFQwNGgodn0iemg

KEYWORDS: Antenna Simulations; Computer Clusters; High-Frequency Electromagnetics; Computer Gpus; Hardware Acceleration; Electrically Large Platforms

** TOPIC AUTHOR (TPOC) **

DoD Notice:   Between November 12 and December 9, 2009, you may talk directly with the Topic Authors to ask technical questions about the topics. Their contact information is listed above. For reasons of competitive fairness, direct communication between proposers and topic authors is not allowed starting December 10, 2009, when DoD begins accepting proposals for this solicitation.

However, proposers may still submit written questions about solicitation topics through the DoD's SBIR/STTR Interactive Topic Information System (SITIS), in which the questioner and respondent remain anonymous and all questions and answers are posted electronically for general viewing until the solicitation closes. All proposers are advised to monitor SITIS (10.1 Q&A) during the solicitation period for questions and answers, and other significant information, relevant to the SBIR 10.1 topic under which they are proposing.

If you have general questions about DoD SBIR program, please contact the DoD SBIR Help Desk at (866) 724-7457 or email weblink.