Time filter

Source Type

La Jolla, CA, United States

Agency: Department of Defense | Branch: Air Force | Program: SBIR | Phase: Phase I | Award Amount: 99.92K | Year: 2004

Current course of action (COA) planning is predominantly a manual ad hoc process where the mission planner seeks to maximize the immediate effects of his COAs while reacting to the opposition's known COAs, and possibly countering the enemy's immediate anticipated response. Long-term moves and countermoves are not considered, nor are objectives achieved in a calibrated quantitative manner. There is a need to search for optimal COAs by considering multiple friendly COAs in light of multiple potential enemy COAs (eCOAs) and "dovetailing" the effects of each side's actions into a chain of moves and countermoves, actions and reactions, that defines the overall sequence of unfolding events in combat. This Phase I project will develop and test software that uses an upgraded JANUS simulation to address this need within the context of UAV (or other vehicles) mission planning in light of enemy defenses. The Phase II effort will treat scenarios involving scores to hundreds of units on multiple sides, and be demonstrated within an EBO context. The software will be applicable to all branches of the military, and to the software entertainment industry. Natural Selection, Inc.'s sister company, Digenetics, Inc., is already well positioned to effect a successful commercial transition.

Agency: National Science Foundation | Branch: | Program: SBIR | Phase: Phase II | Award Amount: 405.90K | Year: 2005

This Small Business Innovation Research (SBIR) Phase II project will develop machine learning tools for RNA gene detection. Prior Phase I research resulted in the successful development of artificial neural networks for the discrimination of functional RNA (fRNA) coding regions from non-coding regions in four model eukaryotes. The Phase II project will focus on (1) refinement of best evolved neural networks for 10 key eukaryotes capable of discriminating fRNA coding from non-coding sequence information, (2) experimental verification of predicted fRNA coding regions in human and mouse, (3) development of machine learning algorithms capable of discriminating between eukaryotic fRNA subtypes, (4) extension of the approach to include machine learning tools capable of discriminating between fRNA subtypes and to evaluate this potential for additional functionality, and (5) development of a user-friendly graphical user interface (GUI) for the product. The commercial application of this project will be to identify a new class of targets for drug design and discovery for the pharmaceutical industry. The educational aspects of the proposed work will be to assist in dissemination of knowledge about the importance of fRNAs to the next generation of scientists.

Agency: Department of Defense | Branch: Air Force | Program: SBIR | Phase: Phase II | Award Amount: 748.85K | Year: 2005

There is a critical need for combat simulations that incorporate intelligently interactive agents, both for mission planning and training. Advances in simulation now offer the opportunity to choose simulation as an effective and cost-saving means for training; however, to be most effective, the agents in the simulation (e.g., an OPFOR) must act believably, in light of a quantified mission/purpose, and choose their courses of action (COAs) based on the likely opposing force’s response (eCOAs), extended as a chain of moves and countermoves. Phase I has demonstrated an innovative COA-eCOA process that uses evolutionary algorithms to optimize COA-eCOA decisions. Experiments involving UCAVs and mobile SAMs have documented the run-time performance of the process. Missions are described in a Valuated State Space® and normalizing function, which provides a framework that can incorporate specific target prioritization, timeliness of sensing and attacks, effects-based operations, and so forth. The work to date has set the stage for the Phase II that will construct the necessary hardware architecture, software development and optimizations, experimentation, and that can be brought to fruition in both government and civilian applications, including mission planning and training for all branches of the military, and entertainment software in the private sector.

Agency: Department of Defense | Branch: Air Force | Program: STTR | Phase: Phase I | Award Amount: 99.94K | Year: 2005

A key to successful command and control is to understand the enemy's intent, particularly in light of incomplete and perhaps inaccurate information regarding social and cultural norms. It is inappropriate to project our own goals and aspirations onto the enemy. Understanding the enemy's intent will become less a matter of understanding the thinking of higher command and more a matter of inferring the adversary's intent based on a priori beliefs regarding their objectives and observed data reflecting the actual decisions in real settings. A novel combination of two technologies, evolutionary computation and the Valuated State Spacer Approach used to quantifying purpose, holds the promise of a general procedure for inferring the enemy's purpose in combat settings ranging from the campaign-level to the level of the individual. The capability described in this proposal will generate an automatic method for optimizing models of the adversary's intent, structured in a hierarchic form. Natural Selection, Inc. will team with its academic partner UCSD on the Phase I effort. Success will allow for more effective course-of-action (COA) determination when viewing the sequence of actions and reactions because the adversary's mission will be better understood, allowing for more accurate predictions of his future behavior.

Agency: Department of Defense | Branch: Air Force | Program: SBIR | Phase: Phase II | Award Amount: 716.25K | Year: 2003

The development of autonomous intelligently interactive software agents is crucial to the success of autonomous control for unmanned vehicle systems. Autonomous systems are scheduled for use over one-third of all future military combat, hence developingsoftware to control these vehicles in a cooperative, intelligent manner is of paramount importance. The objective of this research project is to continue development and fully test evolutionary agent-based software systems (ABS) that can be directlyimplemented for control of fielded unmanned aerial vehicles (UAV) systems. In Phase I, use and broad applicability of these evolutionary agent-based systems was investigated. These evolutionary agents are capable of efficiently generating behaviors for avariety of autonomous systems. Phase II will pursue full implementation and advanced development of evolutionary agents with specific application within a UAV testbed. This same technology has broad commercial applicability, with uses in hazardous wastecontrol, fire-fighting, as well as in a variety of surveillance scenarios. Software developed should be transferable into existing, prototype and testbed autonomous UAV controllers.

Discover hidden collaborations