MokSAF imagemap


The MokSAF Agents
Two main agent types are embedded within the MokSAF interface. Each has been designed according to a different approach to intelligent machines.

The Path Planner Agent

The Path Planner Agent guides the human team members through the route-planning task and performs much of the task itself. This agent acts much like a "black box" or a "Greek Oracle." The agent creates the route using its knowledge of the physical terrain and an artificial intelligence planning algorithm that seeks to find the shortest path. The agent is aware of physical constraints only. Commanders must translate the intangible constraints into physical ones by drawing contrained areas on the map. Click the following thumbnail for examples of user-defined constraints, indicated by shaded boxes, that the Planner Agent must negotiate in designing an acceptable route:

Constraints Placed on the Path Planner Agent

Once the constraints have been drawn on the map, the Path Planner Agent will not draw a route through them.

While the Oracle approach has been criticized for being unable to handle unanticipated situations, we believe that the MokSAF system tempers this critique by providing a mechanism for specifiying social and other constraints on route planning.

The Critique Agent

The second agent analyzes the routes drawn or modified by the human team members and helps them to refine their plans. In this mode, the human and agent work jointly to plan a route to the rendevous point. The workload is distributed so that each component is working to its strengths. Thus, the commander, who has a privileged understanding of the intangible constraints and utilities associated with the mission, can direct the route around these constraints as desired. However, the commander may not be as knowledgeable about the terrain and so the agent can indicate where the path is sub-optimal due to violations of physical constraints. The commander draws the desired route and requests that the Critique Agent review the route for physical, social, or economic violations or to indicate where the path can be improved. Click the following thumbnail for a sample critique:

Sample Critique of Proposed Plan

Given an sub-optimal route, the commander can iteratively improve the plan until a satisfactory solution is reached.

We expect that human team members will employ some combination of both the Path Planner Agent and the Critique Agent.

The MokSAF Agents:

The Path Planner Agent

The Critique Agent


Home - Overview - Objectives - Agent System - Contact Us - Publications - People - Demo - Internal Site (Restricted Access)
_________________________________________________

We welcome your comments and suggestions.
Copyright © 1999

Carnegie Mellon University
Software Agents Group
Robotics Institute
   University of Pittsburgh
Dept of Information Science
& Telecommunications
Project Overview Research Objectives Description of MokSAF Agents Contact Information Publications Researchers and Support Agencies System Demonstration Internal Site MokSAF Home Page