Sensor Network Optimization using Multi-Agent Negotiation (SNOMAN)
The US Missile Defense Agency (MDA) is developing new technologies to defend the US against hostile missile attacks. A key part of the effort is devoted to figuring out how to best use a variety of distributed sensors, including land-based and ship-based radar, to track incoming missiles. Because we won’t know when these missiles will be launched, or where they will be coming from, traditional sensing and tracking solutions won’t work in the short period of time available for an effective defense. Overcoming this challenge requires the development of a novel, adaptive, rapid approach to sensor management that can allocate sensors to track missiles even in dynamic, rapidly changing environments.
The Charles River Analytics Solution
To address this need for a dynamic sensor-management system, scientists and software engineers at Charles River Analytics are building on 200 years of economic theory, which describes how to rationally and optimally distribute goods and services within a market, and 65 years of game theory, which describes how interacting agents should behave to most efficiently achieve their goals. In the case of allocating radar sensors to track missiles, the sensors and the missiles are treated as buyers and sellers in a simulated market where they need to bid against each other to determine which radar will track which missile. As new missiles are launched, the sensors bid for this new resource, possibly selling their current goods to others that are willing to pay for them. The result of this effort is a ground-breaking approach to Sensor Network Optimization using Multi-Agent Negotiation (SNOMAN), which rapidly and dynamically allocates radar sensors to missiles for missile defense.
SNOMAN's interface displays missile and radar information to track threats, as seen in this image.
SNOMAN helps build an optimized resource allocation plan by creating buyer and seller services, attaching value to assets and tasks, and viewing and managing resource management plans.
Charles River’s approach offers significant improvements over existing methods to track incoming missiles in certain situations, such as when a missile defense system has limited sensor resources and must track multiple threats. As an added benefit, SNOMAN decentralizes how sensor networks are tasked, which makes it easier to deploy, while also reducing network bandwidth requirements.
Because this approach to optimizing the use of assets is more general than the specific missile defense problem described here, Charles River is applying the SNOMAN solution to other government command and control projects. SNOMAN can also be applied to comparable problems in the commercial sector, as well, since it can optimize any resource that can be described in terms of cost, time, or scope, such as in logistics management or resource management.
To read an article on SNOMAN featured in MDA's TechUpdate, click here.