A system of software implements an extended version of an approach, denoted shared activity coordination (SHAC), to the interleaving of planning and the exchange of plan information among organizations devoted to different missions that normally communicate infrequently except that they need to collaborate on joint activities and/or the use of shared resources. SHAC enables the planning and scheduling systems of the organizations to coordinate by resolving conflicts while optimizing local planning solutions. The present software provides a framework for modeling and executing communication protocols for SHAC. Shared ac- tivities are represented in each interacting planning system to establish consensus on joint activities or to inform the other systems of consumption of a common resource or a change in a shared state. The representations of shared activities are extended to include information on (1) the role(s) of each participant, (2) permissions (defined as specifications of which participant controls what aspects of shared activities and scheduling thereof), and (3) constraints on the parameters of shared activities. Also defined in the software are protocols for changing roles, permissions, and constraints during the course of coordination and execution.
This program was written by Bradley Clement and Anthony Barrett of Caltech for NASA 's Jet Propulsion Laboratory. For further information, access the Technical Support Package (TSP) free on-line at www.techbriefs.com/tsp under the Software category.
This software is available for commercial licensing. Please contact Don Hart of the California Institute of Technology at (818)393- 3425. Refer to NPO-40438
This Brief includes a Technical Support Package (TSP).
Controlling Distributed Planning
(reference NPO-40438) is currently available for download from the TSP library.
Don't have an account?
Overview
The document discusses a decision-theoretic approach to enhancing communication and coordination among multiple agents in a collaborative framework, particularly in the context of space missions, such as those involving Mars rovers and orbiters. It highlights the challenges of achieving consensus among agents that can only communicate intermittently and under varying constraints, which is a common scenario in multiagent systems.
The authors introduce the SHAC (Shared Activity Coordination) framework, which serves as a planner-independent continual coordination algorithm. This framework is designed to facilitate the design and evaluation of role-based coordination mechanisms. The document outlines the capabilities of SHAC and provides examples of higher-level mechanisms that can be built upon its foundational capabilities.
A significant focus is placed on the concept of consensus protocols, which are essential for agents to agree on shared activities and resource usage. The document categorizes these protocols based on data flow and computation time, illustrating examples such as voting and auction protocols. It emphasizes the need for agents to establish consensus within a defined "consensus window," a period during which negotiations must occur to ensure agreement before execution.
The authors also discuss the importance of communication properties that enable consensus and the overhead involved in establishing it. They present an algorithm for determining the time required to reach consensus given specific communication constraints, which is crucial for planning in environments where agents have limited communication opportunities.
Additionally, the document touches on delegation mechanisms, where one agent assigns tasks to others, and the argumentation methods that agents can use to propose and counter-propose activities. These methods are vital for resolving conflicts and ensuring that all agents can contribute to the planning process effectively.
In conclusion, the document emphasizes the need for ongoing research to evaluate different communication protocols and their benefits across various multiagent domains. The insights gained from coordinating simulated spacecraft in unexpected scenarios will inform future developments in distributed planning and coordination, ultimately enhancing the effectiveness of collaborative efforts in complex environments like space exploration.
