| Sol No.: |
Navy SBIR FY2011.3 |
| Topic No.: |
N113-176 |
| Topic Title: |
CAT Learning Algorithm Workbench (CLAW) |
| Proposal No.: |
N113-176-0016 |
| Firm: |
Charles River Analytics Inc.
625 Mount Auburn Street
Cambridge, Massachusetts 02138 |
| Contact: |
Wayne Thornton |
| Phone: |
(617) 491-3474 |
| Web Site: |
www.cra.com |
| Abstract: |
Current countermeasure anti-torpedo (CAT) systems use explicit logic to direct intercepts resulting in an inability to adapt to the complexities of the stochastic marine environment. The CAT Learning Algorithm Workbench (CLAW) is an analytical research testbed capable of comparing the effectiveness of different machine learning approaches to optimize and automate anti-torpedo fire control and develop criteria concepts for discriminating among them. By applying recent developments in intelligent algorithms to existing simulations and models in the program of record using an instrumented test environment, investigators can identify the most promising designs for using adaptive learning in the Torpedo Warning System. The benefit of the approach is to harden battle group defenses against torpedo salvos by finding optimal fire control solutions and automating the launch decision process. |
| Benefits: |
The CLAW architecture is applicable to programs requiring testing and evaluation of adaptive learning algorithms to complex optimization problems. Government transition opportunities include the ACAT III program. Potential commercial licensing includes placement with prime contractors developing advanced fire control systems. CLAW will enhance the existing AgentWorksT product suite with algorithms and aspects resulting from this research. |