Carroll Thronesbery

Intelligent control of life support for space missions

Future manned space operations will include a greater use of automation than we currently see. For example, semiautonomous robots and software agents will perform difficult tasks while operating unattended most of the time. As these automated agents become more prevalent, human contact with them will occur more often and become more routine, so designing these automated agents according to the principles of human-centered computing is important. We describe two cases of semiautonomous control software developed and fielded in test environments at the NASA Johnson Space Center. This software operated continuously at the JSC and interacted closely with humans for months at a time.

Intelligent control of a water-recovery system: three years in the trenches

This article discusses our experience building and running an intelligent control system during a three-year period for a National Aeronautics and Space Administration advanced life support (ALS) system. The system under test was known as the Integrated Water-Recovery System (IWRS). We used the 3T intelligent control architecture to produce software that operated autonomously, 24 hours a day, 7 days a week, for 16 months. The article details our development approach, the successes and failures of the system, and our lessons learned. We conclude with a summary of spinoff benefits to the AI community and areas of AI research that can be useful for future ALS systems.

Supporting group interaction among humans and autonomous agents

An important aspect of interaction among groups of humans and software agents is supporting collaboration among these heterogeneous agents while they operate remotely and communicate asynchronously. We are developing an architecture that supports multiple humans interacting with multiple automated control agents in such a manner. We are evaluating this architecture with a group consisting of the crew of a space-based vehicle and the automated software agents controlling the vehicle systems. Such agent interaction is modelled as a loosely co-ordinated group because this model minimizes agent commitment to group goals and constraints while addressing a significant portion of crew and control agent group behaviours. In this paper, we give background on human interaction with space-based automation. We identify related research in multi-agent autonomous architectures and single agent human-computer interaction systems, we describe our architecture design for human-software agent groups and we identify research issues in loosely co-ordinated human-software groups.

Multi-agent diagnosis and control of an air revitalization system for life support in space

An architecture of inter-operating agents has been developed to provide control and fault management for advanced life support systems in space. In this multi-agent architecture, cooperating autonomous software agents coordinate with human agents, to provide support in novel fault management situations. This architecture combines the Livingstone model-based mode identification and reconfiguration (MIR) system with elements of the 3T architecture for autonomous flexible command and control. The MIR software agent performs model-based state identification and fault diagnosis. MIR also identifies novel recovery configurations and the set of commands required to accomplish the recovery. The 3T procedural executive and the human operator use the diagnoses and recovery recommendations, and provide command sequencing. Human interface extensions have been developed to support human monitoring and control of both 3T and MIR data and activities. This architecture has been exercised for control and fault management of an oxygen production system for air revitalization in space. The software operates in a dynamic simulation testbed.

An environment for distributed collaboration among humans and software agents

This paper describes an implemented software prototype for the Distributed Collaboration and Interaction (DCI) system, which helps humans to act as an integrated part of a multi-agent system. Human interaction with agents who act autonomously most of the time, such as a process control agent in a refinery, has received little attention compared to human interaction with agents who provide a direct service to humans, such as information retrieval. This paper describes how liaison agents within the DCI system can support human interaction with other agents that are not, by design, human-centric but must be supervised by, or coordinated with, humans. The DCI system provides a step toward future seamless integration of humans and software agents into a cohesive multi-agent system.

Human Interaction Challenges for Intelligent Environmental Control Software

Human interaction challenges for intelligent environmental control software involve finding the right balance between automation and the flexibility for human involvement. Automation is needed to relieve people from the tedium of maintaining a vigilant watch over low-level sensor data and controlling each life support hardware item individually. Flexibility is needed to deal effectively with anomalies and novel situations. This paper discusses strategies for supporting management by exception and shows how those strategies were implemented in support of the automated control system for product gas transfer during the Phase III Test of the Lunar/Mars Life Support Test Program.

Linguistic text mining for problem reports

This paper describes a linguistic text mining tool for analyzing problem reports in aerospace engineering and safety organizations. The Semantic Trend Analysis Tool (STAT) helps analysts find and review recurrences, similarities and trends in problem reports. The tool is being used to analyze engineering discrepancy reports at NASA Johnson Space Center. The tool has been augmented with a statistical natural language parser that also resolves parsing gaps and identifies verb arguments and adjuncts. The tool uses an aerospace ontology augmented with features of taxonomies and thesauruses. The ontology defines hierarchies of problem types, equipment types and function types. STAT uses the output of the parser and the aerospace ontology to identify words and phrases in problem report descriptions that refer to types of hazards, equipment damage, performance deviations or functional impairments. Tool performance has been evaluated on 120 problem descriptions from problem reports, with encouraging results.

Tools to support human factors and systems engineering interactions during early analysis

We describe an approach and existing software tool support for effective interactions between human factors engineers and systems engineers in early analysis activities during system acquisition. We examine the tasks performed during this stage, emphasizing those tasks where system engineers and human engineers interact. The concept of operations (ConOps) document is an important product during this phase, and particular attention is paid to its influences on subsequent acquisition activities. Understanding this influence helps ConOps authors describe a complete system concept that guides subsequent acquisition activities. We identify commonly used system engineering and human engineering tools and examine how they can support the specific tasks associated with system definition. We identify possible gaps in the support of these tasks, the largest of which appears to be creating the ConOps document itself. Finally, we outline the goals of our future empirical investigations of tools to support system concept definition

Situation views: getting started handling anomalies

Event-oriented recognition and display software has been developed to assist human supervisors of automated control systems in maintaining situation awareness and getting started handling anomalies in those systems. The recognition software encapsulates data sets describing related events (i.e., a situation), for review by the human supervisor. The display software supports quick overviews plus details on demand, as well as reminding the operator of definitions for the events in the situation and the expected values of associated parameters. This work is part of a larger effort at Johnson Space Center to develop intelligent aids for use by crew and flight controllers during mission operations.

Integrated Display for Supervisory Control of Space Operations

In recent years, intelligent software has been applied for monitoring and controlling space systems to reduce the workload of the crew and ground support personnel. The application of intelligent software to system control changes the human role in space operations. Instead of directly performing control tasks, humans now often supervise automated software performing control tasks. We have designed an integrated overview display supporting supervisory monitoring and control tasks. This design provides a view of control information integrated over time, supporting a supervisor in following a thread of operational information extending from the past into the future. It associates control activities with the resulting consequences of control on the environment and system configuration. And it provides a uniform way to access and display a heterogeneous set of archived information characterizing system control.