This project is more like the culmination of three or four seperate projects. First, we’ve developed a simple mechanism to adjust the autonomy of an agent during run time, thus constraining the behavior in a desired manner. Next, we’ve written behavior bounding policies which reflect the appropriate set of goals/actions an agent may pursue. The policies not only enforce permitted behavior, but obligation and temporal constraints as well. Finally, we implement a simple default operator which the agent selects when there are no available actions to perform.
Using JOGL, we capture the policy (appropriate set of actions) defined for an agent and give the operator a visual representation of the behavior constraints. In addition, the operator may specify the agent’s level of autonomy, requiring (for example) that the agent suspend its goal until it either consults an expert system or rescinds control to the user.
In the future, we will allow the user to create/modify the policy by detecting events within the OpenGL window. The user may specify not only the level of autonomy, but the scope of available actions defined within this policy.