Building a Goal Reasoning Architecture for Cognitive Active Sonar
Abstract
The focus of the proposed effort is further development of models for incorporating high-level decision making, or cognitive processing, in active sonar systems. Active sonar systems, used for search and track tasks, are typically monitored by trained human operators. These operators use their training and experience to infer the state of the environment from the processed data presented to them and to make recommendations about high-level actions. Typically, parameters and processes within these systems are largely fixed; while the operator may be able to make some changes, e.g. to transmit waveform, most system parameters are not easily tunable. Hence, parameters are not tuned to best achieve the goals of the system. Even if system parameters were tunable by the operator, the task of determining values to meet one or more goals in a dynamic environment would present an impractical cognitive load to the operator. The idea behind cognitive sonar is to integrate intelligence within the sonar system, allowing the system to dynamically tune parameters to best meet current goals, as well as to translate high-level tasks given by an operator to low-level actions within the sonar system. We propose to use recent results in goal reasoning systems, developed within the artificial intelligence community, as the framework developing intelligent active sonar. In prior work, we have created a cognitive sonar architecture that implements prediction and real-time decision making. We propose to continue the development of frameworks that support real-time parameter adaptation to address higher-level goals. Our primary objective is to develop goal management capabilities within these frameworks, allowing the sonar system to arbitrate among competing system goals when making decisions about how to tune parameters and allocate resources. Integrating intelligence in the sonar system reduces the burden on system operators and has the potential to greatly improve how quickly and how well systems can meet surveillance and tracking goals. In addition, cognitive sonar can reduce operator burden by allowing the operator to task the system or query it for particular information. If successful, the proposed effort will produce a goal reasoning architecture for intelligent active sonar that is capable of addressing multiple simultaneous goals. Development of a cognitive sonar system will advance the Navy s ability to reliably conduct surveillance and tracking by allowing sonar operators to interact with the system via high-level commands, thereby reducing the time required to identify a threat.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 19, 2018
- Source ID
- N000141812825
Entities
People
- Jill Nelson
Organizations
- George Mason University
- Office of Naval Research
- United States Navy