Remote Coordination, Actuation, and Planning (ReCAP) Cooperative MultiAgent Architecture

Overview

Development stage

In development or acquisitionSource: a) Pre-deployment – The use case is in a development or acquisition status.

High-impact designation

Not high-impactSource: c) Not high-impact

Impact justification

Not reported

Start date

Not reported

Withheld from public reporting

NoSource: a) No

Mission

Topic area

Science

Operational functions

No function tag matched Derived by keyword rules; see methodology

Problem the AI is intended to solve

ReCAP provides an architecture for lightweight, efficient coordination of highly-capable agents in a comms-limited environment. By providing agents with high-level, lightweight information about the capabilities of other agents in the fleet, ReCAP can facilitate allocation of tasks that are dynamically discovered and assigned via an auction. Auctions are conducted for particular tasks, and agents formulate bids based on encoded subject matter expertise predicting how a given action will affect their operations. For example, one agent may request that two others bid on a sampling task. The two agents bidding may possess different instruments for sampling, one of them able to sample remotely, but at low fidelity, the other able to sample at high fidelity but requiring a physical sample be collected. The auctioneer agent will consider its own mission goals and award the task to what it deems to be the highest bid. Onboard single agents, ReCAP provides a planning/scheduling module and a behavior tree for real-time control. The planning/scheduling module includes the actions of other connected agents, so that any agent may form a plan sequence involving actions from multiple agents – this is what triggers an auction. Finally, the actions that can be executed onboard are handled by a behavior tree, which is more reactive and requires less frequent reconfiguration than a static planner.

Expected benefits

Compared to existing multi-agent control systems, ReCAP requires no training data, before or during operations, and does not require centralized control. It was conceived alongside a NASA push for extensible mission architectures, intended to provide technology that would allow missions to interoperate as new assets are added or encountered, and to enable on-demand coordination between otherwise separate missions. ReCAP enables this capability by basing coordination on high-level, lightweight communications, and its auctioning system prevents the need for interchange of complex state information between agents. ReCAP has been deployed to success in field, laboratory, and simulated use cases across terrestrial and space applications.

System outputs

Through communications interfaces, ReCAP outputs (as well as receiving) state and auction information from other agents. The behavior tree additionally outputs control commands at a level tunable to the given use case; ReCAP has been implemented with direct vehicle control, commanding velocity and position changes, as well as at a much higher level, giving commands to an existing onboard controller.

Technology

AI classification

Agentic AISource: Agentic AI: AI systems that perform tasks or make decisions autonomously with minimal human intervention.

System name(s)

Not reported

Custom-developed code

Not reported

Public source code

Not reported

Sourcing

How it was built

Not reported

Vendor (as reported)

Not reported

Vendors (standardized)

None on the standard list

A vendor is the supplier named by the agency. It does not identify the underlying model or AI technology.

Data and privacy

Involves PII

Not reported

Privacy Impact Assessment

Not reported

Authorization to Operate

Not reported

Demographic features

Not reported

Training and evaluation data

Not reported

Federal Data Catalog

Not reported

Governance

0 of 8 minimum-practice questions answered. A blank answer means the agency reported nothing; it does not mean the practice is absent.

Potential impacts and how they were identified

Not reported