Project Title: Intelligent Workflow for Collection Management
System Name: Smart Workflow for Collection Management (SWIM)
A research project funded by DARPA under the Advanced Intelligence, Surveillance and Reconnaissance Management (AIM) Program, contract number: DABT63-98-C-0069.
Principal Investigator: Dr. Pauline Berry, AI Center,
Principal Team Members:
Dr. Pauline Berry, Artificial Intelligence Center, SRI
Dr. Brian Drabble, CIRL, University of Oregon
On a daily basis, intelligence planners are faced with the task of coordinating multiple ISR assets to
provide as much information about the battlefield as possible and thus increase the effectiveness of the
forces deployed. However, the shrinking defense budget means that, fewer collection and analysis resources are available while the demands for more accurate and more timely information continue to increase.
The AIM program aims to address these issues by providing intelligence planners with tools for more
effective and efficient management of the assets and information in the system. These tools will address
the complete ISR management problem from the strategic development of ISR objectives to the management of
individual ISR assets. Each of these tools and the humans in the ISR domain are viewed as agents that
are capable of adding value to the emerging ISR plan. In order for these highly coupled agents to work
efficiently, they will have to be organized and managed, that is, a workflow plan that describes how the
problem will be solved and identifies the agents necessary to support it must be developed. This involves the development of intelligent workflow techniques that identify when an agent should be called and which tasks are appropriate for the agent to solve. Without intelligent workflow management the integration of information discovery, acquisition, exploitation and dissemination with multi-asset synchronization will be impossible
The workflow architecture developed will provide a revolutionary approach to intelligent workflow within
the collection management domain. To demonstrate the advantages of advanced workflow techniques in the
domain of ISR planning this project will research the following key areas:
The main focus of the research is the information flows between agents and the building of a model that describes the communication between the agents in the workflow plan. This will be used in conjunction with agent-capabilities models and advanced task-balancing techniques to develop an agenda-driven workflow architecture that will be ideally suited to the highly dynamic ISR domain.
- The information flows between agents, to identify the key information resources
- The capabilities of individual agents/agent clusters, to allow effective task delegation
- The algorithms necessary for reactive and flexible agent tasking, to allow rapid replanning and load balancing.
Inherent in each of the issues described above is the control and tasking of agents. This is especially significant in dynamic domains such as the ISR management system. In such a domain, the workflow plan being used to task the agents will be both situation and environment dependent, and as such the coordination of agents must be both flexible and reactive.
This project addresses a number of key issues that are directly relevant to the AIM program and a number of other DARPA programs, e.g. JFACC, ALP, ARPI, etc. The main focus of the project is to develop an intelligent approach to workflow within the collection management domain. We propose to research the information flows between agents and build a model of the resulting communication. In addition the project will investigate advanced task balancing techniques to enable an agenda-driven architecture to best serve agents in the highly dynamic domain of ISR
For AIM program members only.
AI Center Home Page
SRI International Home Page
Pauline M. Berry email@example.com
Last modified: $Date: 1998/05/15 22:45:39 $