Assured Autonomy: Path Toward Living With Autonomous .
Assured Autonomy:Path Toward Living WithAutonomous Systems We Can Trust
The material is based upon work supported by theNational Science Foundation under Grant No. 1734706. Anyopinions, findings, and conclusions or recommendationsexpressed in this material are those of the authors and donot necessarily reflect the views of the National ScienceFoundation.
Assured Autonomy:Path Toward Living WithAutonomous Systems We Can TrustWorkshop OrganizersUfuk Topcu, Chair (The University of Texas at Austin)Nadya Bliss (Arizona State University and CCC); Nancy Cooke (Arizona State University); Missy Cummings (DukeUniversity); Ashley Llorens (Johns Hopkins University, Applied Physics Laboratory); Howard Shrobe (MassachusettsInstitute of Technology); Lenore Zuck (University of Illinois at Chicago)
ASSURED AUTONOMY: PATH TOWARD LIVING WITH AUTONOMOUS SYSTEMS WE CAN TRUST1 Introduction.12 Major findings. 33 Recommendations.54 The challenges, trends, and needs. 64.1 Safety and verification. 64.2 Security. 74.3 Certification and regulation. 84.4 Human-system integration and trust.94.5 Ethics.104.6 Privacy.114.7 Societal impacts. 124.8 Education. 134.9 Governance and policy. 135 Use cases.155.1 Urban air mobility.155.2 Space.165.3 Smart cities and response to systemic disruptions. 175.4 Mobility.186 Workshop participants and contributors. 20
1. IntroductionThe future has arrived – what do we want the future to be?Science and technology create new options and opportunities for society. The printing press democratized knowledge, theindustrial revolution replaced manual labor with machines, and the internet connected everyone. Each advanced society, eachalso disrupted society in the advance. This new age of autonomous systems portends replacing human intelligence with machineintelligence everywhere human intelligence and judgment are used. It, too, can be expected to profoundly advance and disruptsociety. However, the path there is not free and not straightforward. It will be a whole-of-society endeavor to tame and shape ourautonomous systems and assure that technology is safe, secure, fair, equitable, dependable, and accountable. How autonomoussystems will be managed is at the crossing of science, technology, society, policy, and governance.Society will respond, as societies have, through their institutions of government, professions, industries, education, and leaders.Throughout history it has never been a smooth process to balance the public interests, acceptable risks, norms, national security,and industry’s profit motive. Public-private partnerships that nurture innovation can try the patience of governments, theirstakeholders, and the public when the supply chain is too slow to respond and when risks turn into liabilities.Autonomous systems will continue to face the same societal pressures as technology revolutions of the past have faced.However, these systems are different from earlier technologies. Replacing humans and human intelligence with machines andmachine intelligence is replacing within existing frameworks of laws, ethics, morality, norms, as well as an existing technologybase. A software flaw in an autonomous system, for example, is no longer just a bug; rather, it is a potential violation of law andmorality.Assured autonomy, the subject of this workshop series, is central to the discussion of the future of autonomous systems andthe revolution they will surely bring. Lack of assurance manifests in longer development time, at more cost, more patching postdeliverable, the acceptance of new and often unjustified risks from new sources of deployed autonomous systems, unavailabilityand unaffordability of insurance, and the lost opportunity costs from an important technology not existing or not availablewhen needed most. The current pandemic offers examples of lost opportunities. With irony, the perceived benefits of autonomy,if attempted without the necessary assurance, may be negated by the lack of the assurance. Assured autonomy is aboutunderstanding and mitigating risks of operating autonomous systems in our society.The science underpinning assurance is itself still an active interdisciplinary research area, as this workshop has shown, withcritical implications on the future of autonomous systems and the society in general. Too strong a requirement for assurancewith the enforcement mechanisms of government may stifle innovation while keeping the public safe today but exposing societyto future shocks. Too weak a requirement for assurance places the public at risk today with little protection from future shocks.The Assured Autonomy Workshop SeriesThe challenge of establishing assurance in autonomy is rapidly attracting increasing interest in the industry, government, andacademia. Autonomy is a broad and expansive capability that enables systems to behave without direct control by a humanoperator. To that end, it is expected to be present in a wide variety of systems and applications. A vast range of industrial sectors,including (but by no means limited to) defense, mobility, health care, manufacturing, and civilian infrastructure, are embracingthe opportunities in autonomy yet face the similar barriers toward establishing the necessary level of assurance sooner or later.Numerous government agencies are poised to tackle the challenges in assured autonomy.It is important to note that how we analyze, address, measure, and deploy autonomous systems is embedded in an existingset of laws, norms, policies, and societal values, as well as in disciplines. Controls, computer science, machine learning, artificialintelligence, human factors, communication, perceptual and cognitive sciences, business, law, ethics, and public policy are only afew among the disciplines that address the underlying scientific problems in assured autonomy.1
ASSURED AUTONOMY: PATH TOWARD LIVING WITH AUTONOMOUS SYSTEMS WE CAN TRUSTGiven the already immense interest and investment in autonomy, a series of workshops on Assured Autonomy was convenedto facilitate dialogs and increase awareness among the stakeholders in the academia, industry, and government. This series ofthree workshops aimed to help create a unified understanding of the goals for assured autonomy, the research trends and needs,and a strategy that will facilitate sustained progress in autonomy.The first workshop, held in October 2019, focused on current and anticipated challenges and problems in assuring autonomoussystems within and across applications and sectors. The second workshop held in February 2020, focused on existing capabilities,current research, and research trends that could address the challenges and problems identified in workshop 1. The third eventwas dedicated to a discussion of a draft of the major findings from the previous two workshops and the recommendations.The next two sections of the report outline the findings and recommendations. We intend these findings and recommendationsto be usable by the government agencies for building and refining research and development programs and the science andby technology policymakers for setting priorities. The following two sections provide more details on the opportunities andchallenges in several disciplines and applications.Overview of the major findingsThe major advances in the technology for autonomy over the past decade or so promise a transformation driven by autonomyover the coming decades. We are at the beginning of the transformation. The findings of this workshop series outline thechallenges and shortcomings that need to be addressed so that autonomy can sustainably fuel a long-lasting transformation.The findings embrace the (already vast and likely underestimated) extent and heterogeneity of the challenges to be overcome.They also advocate for coordination among all disciplines — technical and “non-technical” — and all stakeholders. No good can come from autonomy without proper assurance. The goal of assured autonomy is human-centered — to amplify, augment, enhance, and empower humans. Autonomous systems have a diverse set of vulnerabilities. Open operation environments amplify technical challenges. Assurance is context-dependent and not once-and-for-all. Assurance for autonomous systems requires a major re-think. Means for measuring the progress and characterizing the gaps are necessary. Autonomy is a socio-economic opportunity and challenge. The public will perceive and be affected by autonomy unevenly. The challenges in autonomy require interdisciplinary approaches. Strengthening connections between the stakeholders will accelerate progress. A re-envisioning of education and workforce development must be part of the path forward. A national research strategy for assurance is needed.Assurance, in this context of autonomy, cannot be an afterthought, and “assured” is not just one of the qualifications forautonomy. Autonomy can survive as a useful technology and fuel the envisioned transformation only with proper assurances.The disciplined principles and practices—yet to be developed—for assurance will create the medium in which the interdisciplinaryapproach this report advocates can be realized. Finally, while the emphasis should be on generalizable principles and practices,assurance will look different for different stakeholders, as each application will carry different risks and trade-offs.2
Overview of the recommendationsThe recommendations aim to outline the high-level characteristics of an approach toward addressing the technical, economical,and societal challenges for the adoption of autonomy at scale and the creation of a long-term transformation fueled by autonomy.They also identify the potential roles for different stakeholders in the implementation of this approach.We emphasize that this report intends to be an invitation — not a final statement — for a comprehensive effort that isinterdisciplinary beyond all past efforts on autonomy.2. Major findingsOpen operation environments amplifytechnical challengesThe following will distill several major findings from theworkshop series. These findings point to challenges andopportunities (at times simultaneously) in autonomy.They include overarching properties that appear acrossdisciplines and the need for change in how we do research,educate students, and/or operate autonomous systems.Autonomous systems are to operate in environments thatcannot be confined within clear boundaries. By the verynature of autonomy, these environments will evolve overtime, may change abruptly, drastically, and unexpectedly,and may include adversaries. These characteristicsamplify the challenges in every element of assurance,from actionable specifications of behavior to verificationto certification.No good can come from autonomy withoutproper assuranceOur ability to establish appropriate assurance forautonomous systems will determine how effectivelywe can manage their risks and benefits. Lack ofappropriate assurance may drive the cost of developmentuneconomically high, delay adoption of autonomoussystems at scale, or result in catastrophic consequences.Lack of assurance can potentially negate the perceivedbenefits of autonomy.Assurance is context-dependent and notonce and for allAn assurance argument is useful only in its intendedcontext, i.e., the environment characteristics, theassumptions in reasoning, the fault and threat models,etc. Therefore, it is critical to explicitly and unambiguouslyexpress what the outcome of assurance does and doesnot guarantee and how robust or sensitive it is to shiftsin the context. Additionally, it is desirable for an assuranceargument to continually adapt to the changes in thecontext in a manner transparent to all stakeholders.Human-centered autonomyThe goal of assured autonomy is to develop systemsthat are not only trustworthy, but that will also amplify,augment, enhance, and empower humans. Such humancentered autonomy can be achieved only by accountingfor human-autonomy integration and human valuesthroughout the design and lifetime of autonomoussystems and developing (human) participatory approachesfor the specification, assessment, and governance ofthese systems.Assurance for autonomous systemsrequires a major re-thinkAutonomous systems feature a level of complexity that isunprecedented in systems for which there exist principlesand practices for systematically establishing assurance.This demands a major re-think of assurance — rather thantrying to merely re-purpose the existing practices.Autonomous systems have a diverse set ofvulnerabilitiesMeasuring the progress and characterizingthe gapsThe development of autonomous systems relies onan uncontrollable, deep supply chain and open-sourcesoftware. Consequently, these systems inherit thevulnerabilities of their building blocks and potentiallyintroduce new vulnerabilities due to the integration ofheterogeneous functionality.The breadth and interdependencies of the challengescomplicate measuring the progress and raise a need fornew benchmarks and testbeds. Furthermore, a catalogof the existing capabilities and acceptable reference3
ASSURED AUTONOMY: PATH TOWARD LIVING WITH AUTONOMOUS SYSTEMS WE CAN TRUSTsystems may help prioritize for more pressing needs andadjust the performance expectations.measures for safety, security, and privacy — three of theproperties autonomous systems are expected to respect.Deferring the progress for any of them may underminethe progress for the others. Similarly, coordination towardaddressing the technical, regulatory, legal, and ethicalchallenges may create not only an impact greater thanotherwise possible but also cross-fertilization.Autonomy is a socio-economic opportunityand challengeAutonomy will potentially affect virtually all aspects ofsociety. Some effects are immediate, such as removingthe driver from vehicles; some effects are indirect andrealized over time. For example, the implications ofdriverless vehicles can be traced all the way to the newbusiness opportunities, the need for new employmentopportunities, the changes in the future of work, urbanplanning, law, governance, and policy, and the implicationsfor the environment. A holistic perspective that embracesall socio-economic aspects — coupled with the technicalaspects — of autonomy is necessary for proper incentivesfor assurance and adoption of autonomy at scale.Strengthening connections between thestakeholders will accelerate progressBuilding stronger connections between the stakeholders —academia, the industry, the regulators, the policymakers,and the public — would improve efficient use of theresources and better align objectives and approaches.Re-envisioning education and workforcedevelopmentAn important — if not the most important — need foraffordably developing and deploying autonomous systemsis an appropriately trained workforce. Arguably noacademic program offers the necessary interdisciplinaryapproach, and recent studies point to a growing mismatchbetween how academia perceives the gaps in educationand training and how industry and government do.The public will perceive and be affected byautonomy unevenlyWhat is an acceptable level of assurance for autonomoussystems will evolve over time as people begin tounderstand the benefits and the risks. The impact ofautonomy on different stakeholders and, even moreimportantly, on different demographic groups of thesociety will be different.A national research strategy for assuranceis neededThe challenges in autonomy requireinterdisciplinary approachesAssured autonomy is a national challenge. Yet therecurrently exists no coordinated approach, no organizedcommunity response, and no national research programsearching for answers.The challenges in autonomy are interdependent. Forexample, such interdependencies induce correlationsand trade-offs between the technical requirements and3. Recommendationsfor assured autonomy in the complex fabric of society isinherently interdisciplinary, convergent, crosscutting,and broad. The recommendations presented here offera vision for exploration, innovation, and advancementthat will make our society more prepared for this newtechnological transformation.We stand at the beginning of the next wave oftechnological change that assured autonomy will bring.It will fundamentally alter our relationships to our criticalinfrastructures, our national defense, our lands, ourlives, and livelihoods. In these different contexts assuredautonomy must be viewed. Each context has its own risksand priorities, its own knowns and unknowns, and itsexpectations of consequences. Research and developmentThe recommendations build a vision for progress arounda “network of institutes on autonomy” and brieflydescribe possible roles for the researchers, educators,4
MAJOR FINDINGSNO GOOD CAN COMEFROM AUTONOMYWITHOUT PROPERASSURANCE.GOAL OF ASSUREDAUTONOMY IS TO AMPLIFY,AUGMENT, ENHANCE,AND EMPOWER HUMANS.THE CHALLENGES INAUTONOMY REQUIREINTERDISCIPLINARYAPPROACHES.OPEN OPERATIONENVIRONMENTSAMPILY THE TECHNICALCHALLENGES.MEASURING THEPROGRESS ANDCHARACTERIZINGTHE GAPS.AUTONOMY IS ASOCIO-ECONOMICOPPORTUNITY ANDCHALLENGE.ASSURANCE ISCONTEXT-DEPENDENTAND NOT ONCE ANDFOR ALL.STRENGTHENINGCONNECTIONSBETWEEN THESTAKEHOLDERS WILLACCELERATE PROGRESS.A NATIONAL
Ufuk Topcu, Chair (The University of Texas at Austin) Nadya Bliss (Arizona State University and CCC); Nancy Cooke (Arizona State University); Missy Cummings (Duke University); Ashley Llorens (Johns Hopkins University, Applied Physics Laboratory); Howard Shrobe (Massachusetts Institute of Technology); Lenore Zuck (University of Illinois at Chicago)
transport cost, storage space and environmental footprint. Getinge Bowie-Dick tests conform to EN ISO 11140-4. Getinge Assured AdvaTec Bowie-Dick Pack 134 C* . Getinge Assured Self-Contained Biological Indicator (EO) Getinge Assured Multi-Incubator Getinge Assured Helix Test 3.5, 4, 5.3, 7 & Formaldehyde Biological Indicators. EN ISO 11140 D .
ASSURED FILE TRANSFER What was previously referred to as "trusted downloading" is now "assured file transfer." This short provides guidance on the requirements for assured file transfer, and includes an approved file formats job aid of DCSA authorized file types and formats for assured file transfer procedures. DATA SPILLS
Medicare Assured Diamond and Gateway Health Medicare Assured Ruby will generally cover the drugs listed in our formulary as long as the drug is medically necessary, the prescription is filled at a Gateway Health Medicare Assured Diamond or Gateway Health Medicare Assured Ruby network pharmacy, and other plan rules are followed.
autonomy modes and the capability to shift between them on the fly [7]. II. DESCRIPTION OF THE SLIDING SCALE AUTONOMY SYSTEM Our system was modelled on the INEEL robot control architecture [7], which consists of four discrete autonomy modes: Teleoperation: In this mode, the user controls the robot directly without any interference from .
Creating an Autonomy-Supportive Classroom Environment for Student Success . Objectives Differentiate between an autonomy-supportive vs a . Students in an Autonomy Supportive environment are: - Creative - Intrinsically motivated - Exhibiting greater conceptual understanding
DAC DAC ADC ADC. X19532-062819. Figur e 2: RF-ADC Tile Structure. mX0_axis Data Path ADC VinX0 mX1_axis Data Path ADC VinX1 mX2_axisData Path ADC VinX2 mX3_axis Data Path ADCVinX3 mX3_axis mX1_axis ADC mX0_axis Data Path ADC Data Path ADC VinX_23 VinX_01 Data Path Data Path Dual RF-ADC Tile Quad RF-ADC Tile. X23275-100919. Chapter 2: Overview
Ashalatha Kunnappilly FORMALLY ASSURED INTELLIGENT SYSTEMS FOR ENHANCED AMBIENT ASSISTED LIVING SUPPORT 2019 Mälardalen University Licentiate Thesis 278 Formally Assured Intelligent Systems for Enhanced Ambient Assisted Living Support Ashalatha Kunnappilly ISBN 978-91-7485-425-1 ISSN 16
LIVING AND NONLIVING THINGS. What is a LIVING thing? If something is LIVING - it is ALIVE ! x A plant is a living thing. x A n animal is a living thing. x YOU are a living thing. These are . living things: a tree . a cat a person . What can living th ings
