David Atkinson

Retired, Not Applicable: Retired, Department Member

Followers

307

Following

Co-authors

Public Views

David J. Atkinson is retired. He was formerly Chief Research Scientist for Artificial Intelligence at Continental Automotive Group. He supervised AI for future mobility systems, advised Continental senior management and leaders globally, and conducted individual research in special AI topics of interest to the company.

Atkinson was previously a Senior Research Scientist in the Institute for Human and Machine Cognition. His research areas of interest include trustworthy autonomous systems, human-robot interaction, cognitive architectures and related topics. Most recently, he has been on assignment with the Air Force Office of Scientific Research in Tokyo, Japan as program manager for information sciences in Asia. He also directed a portfolio of research grants in artificial intelligence for AFOSR in the United States. Prior to joining IHMC, he worked 20 years at the California Institute of Technology (Caltech), Jet Propulsion Laboratory (JPL), the lead NASA Center responsible for robotic space exploration missions. At NASA Headquarters in Washington DC, he served as Deputy Program Executive of the Lunar Precursor Robotic Program which successfully carried out the LRO and LCROSS lunar missions. Prior to his assignment at NASA Headquarters, Atkinson was responsible for JPL's systems engineering contributions to NASA's Constellation Program to explore the Moon and Mars. He also led a cross-agency NASA team of computer scientists that provided engineering studies, requirements development and core engineering processes for Constellation. At JPL, his work spanned basic and applied research in artificial intelligence and robotics to senior executive management. He managed JPL's Information Technologies and Software Systems Division, with responsibility for basic research, development and operations of mission flight software and ground systems. Earlier, Atkinson helped create NASA's basic research programs in artificial intelligence and robotics, proposed and carried out innovative research resulting in numerous publications and inventions, and supervised research projects at JPL and partner institutions in these areas. During a brief sojourn from his career in space exploration, in 1995 Atkinson founded a successful startup which was acquired by IHS (NYSE: IHS) in 2008. PCNAlert.com became the leading provider to the electronics industry of Internet-based component event management data, solutions and applications that enable original equipment manufacturers (OEM) and electronic manufacturing service (EMS) providers to proactively monitor and manage the many components that comprise their end products. Atkinson received the Doctor of Technology (D. Tech) in Computer Systems Engineering from Chalmers University of Sweden in Gøteborg, Sweden and was named a Docent in 1992. He was awarded the Master of Science (M.S) and Master of Philosophy (M.Phil) degrees in Computer Science from Yale University (1982, 1983) and the Bachelor of Arts (B.A) in Psychology from the University of Michigan (1980).

http://orcid.org/0000-0001-5685-2542
Address: david dot atkinson at continental dot com

less

InterestsView All (14)

Uploads

Papers by David Atkinson

Robot Trustworthiness

Proceedings of the Tenth Annual ACM/IEEE International Conference on Human-Robot Interaction Extended Abstracts

ABSTRACT Well-justified human evaluations of autonomous robot trustworthiness require evidence fr... more ABSTRACT Well-justified human evaluations of autonomous robot trustworthiness require evidence from a variety of sources, including observation of robot behavior. Displays of affect by a robot that reflect important internal states not otherwise overtly visible could provide useful evidence for evaluation of robot agent trustworthiness. As an analogy, the human limbic system, sometimes described as an ancient sub-cognitive system, drives human display of affect in a manner that is largely independent of purposeful behavior arising from cognition. Such displays of affect and corresponding attributions of emotion provide important social information that aids understanding and prediction of human behavior. Could an &quot;artificial limbic system&quot; provide similar useful insight into a robot&#39;s internal state? The value of affect signals for evaluation of robot trustworthiness depends on three crucial factors that require investigation: 1) Correlation of affective signals to trust-related, measurable attributes of robot agent internal state, 2) Fidelity in portrayal of emotion by the robot agent such that affective signals evoke human anthropomorphic social recognition, and 3) Correct human interpretation of the affective signals for justifiable modulation of beliefs about the robot agent. This paper discusses these three factors as principles to guide robotic simulation of emotion for increasing human ability to make reasonable assessments of robot trustworthiness and appropriate reliance.

Final Report: The Role of Benevolence in Trust of Autonomous Systems

This report is furnished to the Air Force Office of Scientific Research as the final deliverable ... more This report is furnished to the Air Force Office of Scientific Research as the final deliverable for AFOSR Grant FA9550-‐12-‐1-‐0097, “The Role of Benevolence in Trust of Autonomous Systems.” This report provides a summary of the research and related activities performed with the support of this grant and key results, including pre-‐print copies of the peer-‐reviewed publications and related material. Without reliable and robust methods for assessing the trustworthiness of intelligent, autonomous systems, the issue of trust has become one of the most significant obstacles to broad use of autonomy technology by DoD and other agencies. However, the impact of the research described in this report supports creation of computational methods that create a bridge to future engineering of trustworthy autonomous systems. The core objectives of this research were (1) to operationalize the quality of “benevolence” and understand how it contributes to well-‐calibrated trust of, and reliance upon, autonomous systems, and (2) to investigate portrayal of trust-‐related attributes in the human-‐machine interface. Significant headway was achieved on key topics, including some notable results. Key accomplishments discussed in this report include: (1) the formulation of benevolence as a complex “belief structure” with antecedent beliefs having important semantic, temporal, causal and other interrelationships; (2) the mapping of a portion of this belief structure to measurable internal states of autonomous systems, thereby potentially creating new opportunities for assessment of trustworthiness of such systems; (3) the obtaining of empirical evidence in support of the proposition that previous psychological concepts of interpersonal human trust are applicable to trust in autonomous systems, including the role of personality and situation in modulating the role and importance of certain beliefs; (4) the creation of a theory of a “Human Social Interface” which, when expressed in systems engineering terms, provides guidance for machine portrayal of trust-‐related qualities in human-‐machine social interaction; (5) the design and implementation of a software prototype based on the Human Social Interface theory that provides a basis for future experimentation and evaluation. This project resulted in eight peer-‐reviewed publications and sixteen presentations in scientific venues, meetings with distinguished visitors, and other in support of technology transition opportunities within DoD and to industry. Many new research questions were generated and there remains considerable work to do to fully understand the role of benevolence with respect to intelligent autonomous systems. Overall, the theoretical foundation for trustworthiness of autonomous systems is immature and remains an important area of focus for multiple disciplines.

The Role of Benevolence in Trust of Autonomous Systems

: This report provides a summary of the research and related activities performed. The issue of t... more : This report provides a summary of the research and related activities performed. The issue of trust is one of the most significant obstacles to broad use of autonomy technology by DoD and other agencies. The impact of this research supports creation of computational methods that create a bridge to engineering of trustworthy autonomous systems. Objectives of this research were (1) to operationalize the quality of benevolence and understand how it contributes to well-calibrated trust of, and reliance upon, autonomous systems, (2) to investigate portrayal of trust-related attributes in the humanmachine interface. Accomplishments include: (1) the formulation of benevolence as a complex belief structure with antecedent beliefs having semantic, temporal, causal and other interrelationships; (2) the mapping of a portion of this belief structure to measurable internal states of autonomous systems; (3) empirical evidence in support of the applicability of psychological concepts of interper...

Download

Autonomous Navigation and Control of a Mars Rover

IFAC Proceedings Volumes, 1989

Download

Telerobot Task Planning and Reasoning: Introduction to JPL Artificial Intelligence Research

The goal of building robots with increasing amounts of autonomy from direct human control or supe... more The goal of building robots with increasing amounts of autonomy from direct human control or supervision is a long term goal of the Telerobot project. Among the primary missions for such robots are tasks in the areas of onsrbit satellite servicing, inspection, and various assembly operations associated with the space station. A research and development program with a series of demonstrations of increasing robot autonomy has been planned. This paper discusses some of the technical, methodological, and logistical issues of producing the artificial intelligence capabilities required by the long-term (1993 through 20OO-era) demonstrations of these robot systems. A substantial research effort must begin now if the demonstration objectives are to be met on schedule. It is no ccincidence that many of the same objectives for Telerobot artificial intelligence research are common to both the Telerobot and System Autonomy programs. Many of the capabilities and associated research req.&ed for autonomy transcend the particular application The planned series of Telembot demonstrations provides a relatively clear progression of increasing sophistication in the artificial intelligence capabilities required. As a directed research and develapment program, these demonstrations provide considerable applications pull in the kind of research issues which may be addressed. However, applications pull alone can easily become overly restrictive to new research ideas and also unrealistic about existing o r projected technological capability. Basic research, on the other hand, provides t5e technology push towards new ideas. Ultimately, this push results in demonstrations and application of more powerful capabilities. Technology push by itself runs the risk of being distractable and undirected towards economically realistic and desirable application goals. Clearly, a mixture of h t h applications pull and technology push is necessary. Achieving a sensible balance between these two opposing forces will be major crikria of on-going and overall success of the Telerobot project. Due to the scope of the artificial intelligence *mhnology required by Telerobot, one of the conclusions to dmw from this document is that our goals for Telembot artificial intelligence w i l l be diflkult to achieve unless we make masimal use of dl available research resources. NASA probabIy does not have h e desire or capability to fund aLl of the artificial intelligence research which will be required by Telerobot, nor should NASA duplicate research conducted uuder other auspices. W e will need to draw extensiidy u p n research results generated by the Systems Autonomy Program at the NASA Amea Research Center laboratory, and externally funded laboratories at major universities and within the AI industry.

Download

The Use of Causal Explanations in Learning

ABSTRACT

Presentation at HAI-2014: Methodology for Study of Human-Robot Social Interaction in Dangerous Situations

Presentation AAAI-SS-15 APEX DJA Final

(Is there) A Future for Lying Machines?

Computers will (increasingly) deceive us because there are just too many strategically beneficial... more

Robust Human-Machine Problem Solving

This study of the state of robust human-machine problem solving research in the information scien... more This study of the state of robust human-machine problem solving research in the information sciences is motivated by several fundamental needs presented by the USAF mission: 1. Help humans interact and cooperate 2. Facilitate the consensus creation of knowledge 3. Formulate and facilitate prediction, prescription, decisions and planning 4. Perform complex tasks with ever-increasing volumes and sources of data 5. Achieve a network effect from adding resources Although in the conduct of this study, the stated and implied needs for information sciences by AF Advanced Research Laboratories were analyzed from a number of perspectives, this report will not present this information in detail. Nor will this report suggest solutions to those particular needs. The purpose of the study is to identify the basic research that will enable the formulation of those solutions. The remainder of this report will focus on research in specific areas of information sciences where advances could have both...

Assessing Current Research on Autonomous Systems and Related Technologies in Australasia, Final Report

This is the final report for the grant “Assessing Current Research on Autonomous Systems and Rela... more This is the final report for the grant “Assessing Current Research on Autonomous Systems and Related Technologies in Australasia.” The purpose of this study was to gather and synthesize current information, assess its importance and make recommendations to the AFOSR Asian Office of Aerospace Research and Development regarding scientific program development. The report focuses on research in Australia, New Zealand, South Korea, Japan, Taiwan, and the People’s Republic of China. ... The purpose of this study was to gather and synthesize current information regarding information sciences research in Australasia, assess its importance, and make recommendations on possible information exchanges, investments, and transition opportunities. The primary focus was on assessment of information sciences S&T in Australia and New Zealand, our strongest allies in the region. This report also provides assessments of Korea and Taiwan, and some information on activities in China (PRC). The organizati...

Full Paper: Autonomy Technology Challenges of Europa and Titan Exploration Missions

Emerging Cyber-Security Issues of Autonomy and the Psychopathology of Intelligent Machines

The central thesis of this paper is that the technology of intelligent, autonomous machines gives... more The central thesis of this paper is that the technology of intelligent, autonomous machines gives rise to novel fault modes that are not seen in other types of automation. As a consequence, autonomous systems provide new vectors for cyber-attack with the potential consequence of subversion, degraded behavior or outright failure of the autonomous system. While we can only pursue the analogy so far, maladaptive behavior and the other symptoms of these fault modes in some cases are similar to those found in humans. The term "psychopathology" is applied to fault modes of the human mind, but as yet we have no equivalent area of study for intelligent, autonomous machines. This area of study must be created in order to document and explain the symptoms of unique faults in intelligent systems, whether they occur in nominal conditions or as a result of an outside, purposeful attack. By analyzing algorithms, architectures and what can go wrong with autonomous machines, we may a) gai...

Download

Knowledge systems support for mission operations automation

ABSTRACT

Artificial intelligence for monitoring and diagnosis of robotic spacecraft

In this thesis the application of artificial intelligence to monitoring and diagnosis of robotic ... more In this thesis the application of artificial intelligence to monitoring and diagnosis of robotic spacecraft is discussed. Several software prototype systems were developed to serve as testbeds for the research and to evaluate the effectiveness of the approach against real problems and current techniques used in NASA's planetary exploration program. Software prototypes were used to investigate the verification of robot plan execution. New artificial intelligence algorithms for monitoring and diagnosis of robot systems were designed, programmed, and tested. These included plan analysis for monitoring, sensor planning, generation of expected sensor values, and diagnosis of execution failures caused by hardware, environmental or plan anomalies. Testing was performed on a laboratory telerobotic hardware testbed for satellite servicing and on a mobile planetary rover robot operating in natural terrain. Artificial intelligence algorithms, software prototypes, and more advanced, operationally capable systems for monitoring ground support systems and actual spacecraft in flight were designed, programmed, and tested. A ground support system that served as one test domain was the mirror cooling circuit of the 25-foot Space Simulator at the Jet Propulsion Laboratory (JPL) in Pasadena, California. A prototype monitoring system for this device based on a theory of "predictive monitoring" was developed and tested. Mission operations for the Voyager II spacecraft served as another test domain for an intelligent spacecraft health-monitoring and diagnosis system. This system was successfully tested in support of telecommunications operations during Voyager II's encounter with the planet Neptune in 1989. This was the one of the first artificial intelligence systems to be used in planetary spacecraft operations at NASA/JPL. Subsequently, this system was adapted and tested in support of operations of the Magellan spacecraft telecommunications subsystem and the Galileo spacecraft power and pyro subsystem. Some of the specific artificial intelligence algorithms that were developed for monitoring and diagnosis included the use of heuristic and causal model-based reasoning techniques for predictive generation of sensor values, sensor selection planning, dynamic alarm limit checking, hierarchical procedure specialists for fault diagnosis, and integration of Al with conventional systems in full-scale monitoring and diagnosis applications. In support of this overall program of research, novel software engineering tools for artificial intelligence research and application development were also developed and will be discussed in the thesis. The application of artificial intelligence techniques to the monitoring and diagnosis of robotic space systems was shown to be very effective with specific benefits in the areas of systems autonomy, spacecraft safety, ground operations productivity and automation. As a result of this work in part, artificial intelligence is now considered by senior mission designers to be an enabling technology for on-board automation of planetary rovers and for automation in mission operations at the Jet Propulsion Laboratory.

Attitudes and Personality in Trust of Intelligent, Autonomous Agents

Autonomous Agents and Human Interpersonal Trust: Can We Engineer a Human-Machine Social Interface for Trust?

There is a recognized need to employ autonomous agents in domains that are not amenable to conven... more There is a recognized need to employ autonomous agents in domains that are not amenable to conventional automation and/or which humans find difficult, dangerous, or undesirable to perform. These include time-critical and mission-critical applications in health, defense, transportation, and industry, where the consequences of failure can be catastrophic. A prerequisite for such applications is the establishment of wellcalibrated trust in autonomous agents. Our focus is specifically on human-machine trust in deployment and operations of autonomous agents, whether they are embodied in cyber-physical systems, robots, or exist only in the cyber-realm. The overall aim of our research is to investigate methods for autonomous agents to foster, manage, and maintain an appropriate trust relationship with human partners when engaged in joint, mutually interdependent activities. Our approach is grounded in a systems-level view of humans and autonomous agents as components in (one or more) encompassing metacognitive systems. Given human predisposition for social interaction, we look to the multi-disciplinary body of research on human interpersonal trust as a basis from which we specify engineering requirements for the interface between human and autonomous agents. If we make good progress in reverse engineering this "human social interface," it will be a significant step towards devising the algorithms and tests necessary for trustworthy and trustable autonomous agents. This paper introduces our program of research and reports on recent progress.

Download

Knowledge-Based Diagnosis for Aerospace Systems

This document discusses autonomous diagnosis of aerospace systems. The need for automated diagnos... more This document discusses autonomous diagnosis of aerospace systems. The need for automated diagnosis in aerospace and the approach of using knowledge-based systems are examined. Research issues in knowledgebased diagnosis which are important for aerospace applications are treated along with a review of recent relevant research developments in Artificial 1ntelligence.The design and operation of some existing knowledge-based diagnosis systems are described. The systems described and compared include the LES expert system for liquid oxygen loading at NASA Kennedy Space Center, the FAITH diagnosis system developed at the Jet Propulsion Laboratory, the PES procedural expert system developed at SRI International, the CSRL approach developed at Ohio State University, the StarPlan system developed by Ford Aerospace, the IDM integrated diagnostic model, and the DRAPhys diagnostic system developed at NASA Langley Research Center. iii This technical report is a preprint of a chapter which will be published as part of a book titled, "Machine Intelligence and Autonomy for Aerospace Systems." The book will be a volume of the AlAA Series Proa ress in Astron-and Aeronautics. However, the earliest estimate of the date of publication of this book is some time in the spring of 1988. I have decided to issue this technical report in the belief that the technology of knowledge-based systems for automated diagnostic applications is of substantial immediate interest to engineers, technical staff, and aerospace managers who are currently developing, plan to develop, or are simply curious about the area. The reader will note that I have to a large extent avoided discussion of much of the background, rationale, or basic technology behind knowledge-based systems. Many other high quality references for this already exist. I have also chosen not to describe simple rule-based systems for diagnosis in much detail. This particular implementation approach was one of the first to be applied to the problem of diagnosis (as part of the MYCIN system developed by Shortliffe at Stanford in the late 1970's) and its success instigated the development of many commercial products for developing rule-based systems. Instead, the chapter primarily discusses the overall architecture of knowledge-based approaches to diagnosis which may or may not be suitable for implementation using rulebased systems.

Download

Autonomous task level control of a robot

Proceedings of ROBEXS, 1986

ABSTRACT

Shared Awareness, Autonomy and Trust in Human-Robot Teamwork

Teamwork requires mutual trust among team members. Establishing and maintaining trust depends upo... more Teamwork requires mutual trust among team members. Establishing and maintaining trust depends upon alignment of mental models, an aspect of shared awareness. We present a theory of how maintenance of model alignment is integral to fluid changes in relative control authority (i.e., adaptive autonomy) in human-robot teamwork.

Download

David Atkinson

Uploads

Papers by David Atkinson

Log In