INTERNATIONAL JOURNAL OF COGNITIVE ERGONOMICS, 2001, 5(1), 37–57
`Copyright © 2001, Lawrence Erlbaum Associates, Inc.
`
`On the Design of Adaptive Automation for Complex
`Systems
`
`David B. Kaber
`Department of Industrial Engineering
`North Carolina State University
`
`Jennifer M. Riley and Kheng-Wooi Tan
`Department of Industrial Engineering
`Mississippi State University
`
`Mica R. Endsley
`SA Technologies
`Marietta, Georgia
`
`ABSTRACT
`
`This article presents a constrained review of human factors issues relevant to adaptive automation
`(AA), including designing complex system interfaces to support AA, facilitating human–computer in-
`teraction and crew interactions in adaptive system operations, and considering workload associated
`with AA management in the design of human roles in adaptive systems. Unfortunately, these issues
`have received limited attention in earlier reviews of AA. This work is aimed at supporting a general the-
`ory of human-centered automation advocating humans as active information processors in complex
`system control loops to support situation awareness and effective performance. The review demon-
`strates the need for research into user-centered design of dynamic displays in adaptive systems. It also
`points to the need for discretion in designing transparent interfaces to facilitate human awareness of
`modes of automated systems. Finally, the review identifies the need to consider critical human–human
`interactions in designing adaptive systems. This work describes important branches of a developing
`framework of AA research and contributes to the general theory of human-centered automation.
`
`1.
`
`INTRODUCTION
`
`Adaptive automation (AA) has been described as a form of automation that allows for dy-
`namic changes in control function allocations between a machine and human operator based
`on states of the collective human–machine system (Hilburn, Byrne, & Parasuraman, 1997;
`Kaber & Riley, 1999). Interest in dynamic function allocation (DFA, or flexible automation)
`
`Requests for reprints should be sent to David B. Kaber, Department of Industrial Engineering, North Carolina
`State University, Raleigh, NC 27695–7906. E-mail: dbkaber@eos.ncsu.edu
`
`Yuneec Exhibit 1026 Page 1
`
`

`
`38
`
`KABER, RILEY, TAN, ENDSLEY
`
`has increased within the recent past as a result of hypothesized benefits associated with the
`implementation of AA over traditional technology-centered automation. Purported benefits
`include alleviating operator out-of-the-loop performance problems and associated issues,
`including loss of situation awareness (SA) and high mental workload. Though the expected
`benefits of AA are encouraging, there are many unresolved issues regarding its use. For ex-
`ample, there is currently a lack of common understanding of how human–machine system
`interfaces should be designed to effectively support implementation of AA.
`In this article, current AA literature is reviewed in the context of a theoretical framework
`of human-centered automation research with the objective of identifying critical factors for
`achieving human–automation integration to support the effective application of AA to com-
`plex systems. We describe branches of a research framework supporting human-centered
`automation that seems to have been neglected by previous literature reviews, including the
`implications of the design of AA on operator workload and the effects of AA on hu-
`man–computer interaction (HCI) and crew interaction. This work is important because an
`optimal approach to AA remains elusive. Developing a unified perspective of the aforemen-
`tioned issues may serve as a basis for additional design guidance to structure AA applica-
`tions beyond that previously provided.
`
`1.1. Human-Centered Automation Theory and AA
`
`A theory of human-centered automation closely related to AA states that complex systems
`should be designed to support operator achievement of SA through meaningful involvement
`of operators in control operations (Endsley, 1995b, 1996; Kaber & Endsley, 1997). Involve-
`ment may occur through intermediate levels of automation (LOAs) or through AA. Both
`techniques may be effective for increasing operator involvement in control operations as
`compared to full automation. Human-centered automation is concerned with SA because it
`has been found to be critical in terms of successful human operator performance in complex
`and dynamic system operations (cf. Endsley, 1995a). AA has been proposed as a vehicle for
`moderating operator workload or maintaining it within predetermined acceptable limits,
`based on task or work environment characteristics, to facilitate and preserve good SA
`(Hilburn et al., 1997; Kaber & Riley, 1999). Therefore AA might be considered a form of hu-
`man-centered automation. Unfortunately, the relation between SA and workload presents a
`conundrum to those designing automation. Optimization of both SA and workload in the
`face of automation can prove difficult. Under low workload conditions associated with high
`levels of system automation, operators may experience boredom and fatigue due to lack of
`cognitive involvement, or interest in, control tasks. Operators of autonomous systems are of-
`ten forced into the task of passive monitoring of computer actions rather than active task pro-
`cessing. Even when attending to the monitoring task, decreased task involvement can com-
`promise operator SA (Endsley & Kaber, 1999; Endsley & Kiris, 1995; Pope, Comstock,
`Bartolome, Bogart, & Burdette, 1994). This is an important issue because operators with
`poor SA may find it difficult to reorient themselves to system functioning in times of system
`failure or unpredicted events. Therefore, automated system performance under failure
`modes may be compromised.
`Conversely, cognitive overload may occur when operators must perform complex, or a
`large number of, tasks under low levels of system automation (e.g., complete manual con-
`trol). High workload can lead directly to low levels of SA and task performance, as opera-
`
`Yuneec Exhibit 1026 Page 2
`
`

`
`ADAPTIVE AUTOMATION
`
`39
`
`tors struggle to keep up with the dynamically changing system. Increasing task
`requirements beyond that which the human is cognitively capable of managing can also lead
`to feelings of frustration and defeat, as well as a loss of confidence in ability to complete the
`task. The operator may then become detached from the task, resulting in loss of SA. Again,
`the loss of SA can lead directly to poor human–machine system performance.
`The first situation described above may be due to system and task design. The second sit-
`uation may result from operator reactions to a difficult task. It should be noted that, between
`these two extremes, it has been found that SA and workload can vary independently
`(Endsley, 1993). The challenge for AA research is to identify the optimal workload, or func-
`tional range, under which good levels of operator SA and total system performance will be
`possible.
`The key issues that must be addressed to meet this need include determining how the de-
`sign of automation or AA methods affect operator workload and how system information
`should be communicated to operators to facilitate SA under AA. Several studies have dem-
`onstrated positive results in terms of operator SA when applying AA as an approach to hu-
`man-centered automation of complex systems. For example, Kaber (1997) observed
`improvements in SA in a simulated automatic dynamic, control task (“radar” monitoring
`and target elimination) when using a preprogrammed schedule of periodic shifts of task
`control between intermediate- and high-level automation and manual control, as compared
`to fully autonomous or completely manual control. Although important for establishing
`preliminary system design guidelines and providing insights into methods of AA, this work
`and other recent studies (e.g., Kaber & Riley, 1999) have been conducted using specific task
`and operational scenarios and, therefore, results may have limited generalizability to a
`broad range of systems.
`Unfortunately, at this point there exists no theory of AA that can optimally address SA
`and workload tradeoffs across all types of complex systems (e.g., air traffic control, produc-
`tion control, and telerobotic systems). This article seeks to address this issue by supporting
`the concept of human-centered automation and presenting an understanding of aspects of
`the relation of AA to SA and workload not previously explored in detail.
`
`1.2. Previous Research
`
`Preliminary or casual reviews of AA research have been published (cf. Parasuraman,
`Mouloua, Molloy, & Hilburn, 1996; Scerbo, 1996), summarizing empirical studies of the
`concept, which make inferences toward a general theory of AA. For example, Scerbo’s work
`includes a brief review of traditional automation, proposed AA mechanisms and strategies,
`and potential benefits and concerns with the implementation of AA. Our work complements
`this effort by discussing some new issues, such as
`
`1. Failures in AA design to consider operator workload requirements associated with
`managing dynamic control allocations between themselves and automated systems
`in addition to maintaining system task responsibilities.
`2. The need to determine how human–computer interfaces should be designed to sup-
`port effective human–automation communication under AA.
`3. The need to evaluate the impact of implementation of AA on human crew interac-
`tions in systems control.
`
`Yuneec Exhibit 1026 Page 3
`
`

`
`40
`
`KABER, RILEY, TAN, ENDSLEY
`
`These issues are considered in the context of the human-centered automation theory with the
`intent of developing a more complete knowledge of AA.
`
`2. WORKLOAD AND AA
`
`Unfortunately, it has been observed through empirical study of AA that operators of many
`complex, dynamic systems may experience workloads above desired levels as a result of
`concentrating on control function allocations and maintaining task responsibilities simulta-
`neously (Kaber & Riley, 1999; Scerbo, 1996). An increase in human operator workload as-
`sociated with introduction of automation in complex systems is not a new issue. Selcon
`(1990) observed that fighter aircraft pilot perceptions of flight workload increased signifi-
`cantly with the introduction of automated decision aids into aircraft cockpits.
`There are two general cases in which perceived workload increases may occur in appli-
`cations of AA. First, operators may perceive increased cognitive load in monitoring com-
`puter management of function allocations between themselves and automated subsystems
`(Endsley, 1996). This may be due in part to operator anxiety about the timing of allocations
`and the need to complete a particular task during system operations. It may also be attrib-
`uted to an additional load on the visual channel in perceiving task-relevant information on
`“who is doing what.”
`The second involves implementation strategies of AA where the human has the task of
`managing function allocations in addition to performing routine operations. Under these
`circumstances, workload increases may be even greater than that associated with monitor-
`ing computer-based dynamic control allocations (Selcon, 1990). Additional problems indi-
`cate operators may have trouble in identifying when they need to switch from manual to
`automated modes or vice versa (Air Transport Association, 1999). Failures to invoke auto-
`mation or manual control have been identified as occurring due to operator overload,
`incapacitence, being unaware of the need for a different LOA, or poor decision making
`(Endsley, 1996).
`Kaber and Riley (1999) studied the effect of AA on operator workload during
`dual-task performance involving a primary dynamic control task and an embedded sec-
`ondary monitoring task. Participants in this study were provided with a computer deci-
`sion aid that either suggested or mandated DFAs between manual and automated control
`of the primary task based on participant performance in the secondary task. The authors’
`objective was to maintain secondary task performance within 20% of optimal perfor-
`mance observed during testing in the absence of primary task control. Average secondary
`task performance levels during dual-task functioning were within approximately 30% of
`optimal secondary task performance. It is important to note that when the primary task
`was fully automated, secondary task performance was within 5% of optimal. However,
`automated primary task performance may not have been superior to AA of the task.
`Kaber and Riley attributed the observed decrease in performance (indicative of increased
`workload) to the need for individuals to monitor automated dynamic control allocations
`or to manage them, which was not considered in establishing optimum secondary task
`performance baselines or the design of the dual-task paradigm. This is an important issue
`that needs to be considered by future research to ensure that AA achieves the objectives
`of human-centered automation (i.e., moderating workload and maintaining SA). Methods
`
`Yuneec Exhibit 1026 Page 4
`
`

`
`ADAPTIVE AUTOMATION
`
`41
`
`for dealing with AA-induced workload must be devised. A critical step to developing
`such techniques would be to evaluate operator workload associated with the implementa-
`tion of general AA strategies separate from system task workload. These workload com-
`ponents could then be used to drive AA design.
`
`3.
`
`INTERFACE DESIGN FOR AA
`
`In addition to considering the effects of AA on workload, the effects on operator SA must
`also be considered. Implementation of AA may introduce added complexity into system
`functioning and control. Consequently, operators require advanced interfaces that are useful
`for dealing with this complexity to enhance, rather than hinder, system performance. AA
`will require extra attention to developing interfaces that support operator SA needs at vary-
`ing LOAs and in ways that support their ability to transition between manual and automated
`control and back again.
`Scerbo (1996) suggested that the success of AA will in large part be determined by sys-
`tem interface designs that include all methods of information exchange (e.g., visual, audi-
`tory, haptic, etc.). With this in mind, one goal of the interface design for AA systems is akin
`to that of HCI research, that is, to facilitate the transmission of information to and from the
`human and system without imposing undue cognitive effort on the operator in translating
`the information. There are many other general human factors interface design principles for
`complex systems that may have applicability to interfaces for AA, including, for example,
`the list provided by Noah and Halpin (see Rouse, 1988). However, what is needed at this
`point are high-level and specific interface design recommendations that are presented in the
`context of systems to which AA is most common, such as aircraft.
`
`3.1. AA and Cockpit Interfaces
`
`Although aircraft systems currently support a crude level of AA (pilots may shift between
`manual and automated control at will), a number of problems with this process have been
`noted. For instance, today’s automated flight management systems do not adequately sup-
`port pilots in coordinating between information meant to support manual flight and that
`meant to support automated flight (Abbott, Slotte, & Stimson, 1996). For example, the
`American Airlines Flight 965 aircrew that crashed in Cali, Columbia, in 1995 was forced to
`struggle with paper maps and displays that used different nomenclatures and provided dif-
`ferent reference points, making it very difficult to coordinate between manual and automated
`operations (Endsley & Strauch, 1997). They furthermore had only partial information pro-
`vided through any one source and, therefore, were required to integrate cryptic flight plan in-
`formation in working memory. These discrepancies leave pilots faltering in trying to work
`with systems that do not support their operational needs. The systems interfaces are poorly
`designed in terms of providing the SA needed for understanding the behavior of the aircraft
`in automated modes, and predicting what a system may do in any given situation has proven
`erratic. Attempts by pilots to make dynamic shifts in LOAs in situationally appropriate ways
`have been shown to be fraught with problems (Air Transport Association, 1999; Endsley &
`
`Yuneec Exhibit 1026 Page 5
`
`

`
`42
`
`KABER, RILEY, TAN, ENDSLEY
`
`Strauch, 1997), and aircraft interfaces do not allow pilots to track shifts and to effectively and
`efficiently adapt to them.
`At a very basic level, system displays for supporting manual and automated control need
`to be consistent and coordinated to allow smooth transition from one mode of operation to
`another. In the context of aviation systems, Palmer, Rogers, Press, Latorella, and Abbott
`(1995) stated that interface design should
`
`1. Foster effective communication of activities, task status, and mission goals, as well
`as the development of useful and realistic conceptual models of system behavior.
`2. Enhance operator awareness of his or her own responsibilities, capabilities, and limi-
`tations, as well as those of other team members.
`3. Support DFA that is quick, easy, and unambiguous.
`
`The latter recommendation is directed at AA and supporting pilot performance when shifts
`in LOAs occur. These are important recommendations because the way in which an interface
`presents information to the user will impact what is perceived, how accurately information is
`interpreted, and to what degree it is compatible with user needs or models of task perfor-
`mance (all of which may critically influence operator development of good SA on modes of
`operation of a complex system).
`Unfortunately, the application of AA to complex systems like aircraft often increases
`rather than decreases the amount of information an operator must perceive and use for task
`performance, including data on system automation configuration and schedules of control
`function allocations. On the basis of Palmer et al.’s (1995) recommendations, interfaces for
`AA must support integration of such data regarding “who is doing what” with task-relevant
`data. And they should ensure that all information is presented in a cohesive manner; there-
`fore, function allocation information should have meaning to current task performance. For
`example, aircraft automated vertical flight control modes should provide guidance on the
`operation of different types of speed control (e.g., speed controlled via elevators with maxi-
`mum thrust or idle thrust) and altitude control (e.g., vertical speed or altitude controlled via
`the elevators and speed controlled via throttles) on the basis of current phase of flight and
`current flight segment, as well as the current LOA for flight control (Feary et al., 1998).
`In addition to these, interfaces are needed to facilitate the development of strong mental
`models regarding how such a complex system will function across many classes of situa-
`tions. Lehner (1987) stated that accurate mental models are important because HCI can re-
`main effective even when there is significant inconsistency between the problem-solving
`processes of the human and the decision support system, although system error conditions
`may occur in which recovery is only possible by one method of operation. Cockpit inter-
`faces for supporting mental models of automated systems in aircraft operations have been
`found to be very poor, leading to significant difficulties in understanding system behavior
`(McClumpha & James, 1994; Wiener, 1989).
`In particular, mental model development can be affected by system response feedback on a
`user’s actions through an interface in addition to consistently displayed system state informa-
`tion. Feedback allows the operator to evaluate the system state in relation to his or her control
`actions, goals, and expectations of system functioning. Both individual and team feedback of
`knowledge of system states and responses have been shown to optimize human–machine per-
`formance (Krahl, LoVerde, & Scerbo, 1999). Lack of feedback forces the human into an
`open-loop processing situation in which performance is generally poor (Wickens, 1992).
`
`Yuneec Exhibit 1026 Page 6
`
`

`
`ADAPTIVE AUTOMATION
`
`43
`
`Although the need for good SA and good mental models are fundamental to the opera-
`tion of automated systems in general, achieving them can be even more challenging with the
`added complexity of AA. System interfaces need to support the understanding of not just
`one system, but multiple systems, in that at different levels of AA, the system may operate
`in very different ways.
`
`3.2. Dynamic (Cockpit) Displays For AA
`
`Morrison, Gluckman, and Deaton (1991) also raised general interface design issues that
`should be considered when implementing AA in the airplane cockpit. They stated that auto-
`mated tasks may require new interfaces and cues so that (a) the status of the automation is
`clearly indicated to the human, (b) effective coordination of task performance is facilitated,
`(c) monitoring of the automated task by the human is encouraged, and (d) manual perfor-
`mance of the task after automation is not negatively affected. These interface characteristics
`are similar to the design recommendations made by Palmer et al. (1995). Unfortunately, they
`do not offer specific interface design guidelines for AA. However, like many other AA re-
`searchers, Morrison et al. are proponents of using adaptive interfaces, or displays that change
`dynamically, according to changes in AA control allocations to ensure the effective coordi-
`nation of task performance.
`Introducing dynamic displays into adaptive system interface design is currently a critical
`research issue. Dynamic displays can allow for consideration of operator information re-
`quirements as well as styles of interaction through their configuration. For example, dy-
`namic displays implemented in the aircraft cockpit can present specific interface features
`based on different modes of automated flight and functional roles of pilots under different
`modes. They can also allow pilots to select or deselect features according to their individual
`information needs and styles of flying the aircraft (Wiener, 1988). By allowing for flexible
`configuration of displays and meeting pilot information requirements, SA may be enhanced
`and performance made effective across modes of aircraft operation.
`Dynamic displays have, however, been noted to cause human–machine system perfor-
`mance problems depending on how they are implemented. If dynamic displays are opti-
`mized to include just the information that supports a particular mode of operation, the global
`SA that is needed to support operators’ knowledge of when to switch modes may be lacking.
`That is, they also need information that will alert them to the need to switch from automated
`to manual control and the information that will support such a transition smoothly. Display
`interfaces that are optimized for automated control may lack sufficient information to allow
`operators to build up this level of understanding. The same can be said of the transition from
`manual to automated control, although this may not be as difficult. Norman (1990) noted
`that designers often leave critical information out of automated displays in the belief that
`operators no longer need that information.
`From the opposite perspective, Wiener (1988) pointed out that there is a potential toward
`display clutter and ill-considered symbols, text, and color in many dynamic display designs
`for complex systems. This is brought about by the designer attitude that if it can be included
`in the interface, then it should. This approach to interface design strays from the theory of
`human-centered automation (Billings, 1997). A number of AA research studies have been
`conducted to establish interface design approaches to address this tendency. For example,
`direct manipulation interface design was proposed by Jacob (1989) as an interface style for
`use in AA systems to offset some performance disadvantages associated with dynamic dis-
`
`Yuneec Exhibit 1026 Page 7
`
`

`
`44
`
`KABER, RILEY, TAN, ENDSLEY
`
`plays that are linked to different modes of automation and to address the lack of transpar-
`ency of system functions through interfaces under high LOAs. The lack of function
`transparency has been associated with mode awareness problems (Sarter, 1995).
`Ballas, Heitmeyer, and Perez (1991) studied the direct manipulation interface style to de-
`termine whether it would be particularly effective in an intelligent cockpit implemented with
`AA. Two important components of direct manipulation that were anticipated to improve per-
`formance included (a) reduced information processing distance between the users’ intentions
`and the machine states and (b) direct engagement without undue delay in system response and
`with a relatively transparent interface. Ballas et al. found that using direct manipulation and
`maintaining a consistent interface style could offset the negative effects of changing controls
`and displays (dynamic displays). They also speculated that direct manipulation would en-
`hance SA in assessment tasks in particular, and, consequently, have the potential to reduce au-
`tomation and dynamic display-induced performance disadvantages.
`In support of these findings, other research has shown that adaptive systems providing
`indirect manipulation and opaque interfaces have negative effects on human–computer
`communication and overall system performance, as they may restrict human interaction
`with the system (Scerbo, 1996). Sarter and Woods (1994) observed an automation opacity
`problem with adaptive systems and claimed that user data interpretation becomes a
`cognitively demanding task rather than a mentally economical one.
`On the basis of this research in the context of adaptive systems, Scerbo (1996) encouraged
`designers of interfaces for AA to include as many information formats as possible to allow
`data to flow more freely between the human and system. In this way, operators may be able to
`communicate more naturally because information translation would not be limited to one or
`two formats. However, it is important to ensure that multimodal interface capabilities of con-
`temporary, complex systems are not exploited to the extent of causing information overload,
`as previously observed by Wiener (1988) in historical dynamic displays.
`
`3.3. Summary of Interface Design Research For AA
`
`Some general guidelines for AA have been presented in context, but they do not offer the
`specificity needed to fully support design. Further applied work is needed in this area to eval-
`uate the degree to which the designs of dynamic displays support human performance with-
`out increasing cognitive and perceptual loading. In addition, work should be done to explore
`the effects of using multiple display formats, as some researchers have suggested, for meet-
`ing specific operator information requirements and simultaneously ensuring global aware-
`ness of system states and changes among modes of operation. In particular, careful attention
`needs to be paid to the extra demands associated with detecting the need for, and effecting,
`smooth transitions between AA modes.
`
`4. AA AND HCI
`
`Communication is a critical factor in achieving effective human–automation integration.
`Most researchers agree that effective communication among complex system components is
`critical for overall system success (see Scerbo, 1996). This is due to each individual compo-
`nent or member of the system (the human or computer) possessing knowledge and informa-
`tion that other members may not. Thus, each member must share information to make deci-
`sions and carry out actions.
`
`Yuneec Exhibit 1026 Page 8
`
`

`
`ADAPTIVE AUTOMATION
`
`45
`
`Within the context of human–human teams, this need has been termed shared SA. Shared
`SA is defined as the degree to which team members have the same awareness of information
`requirements for team performance. It is related to team SA, which is “the degree to which
`each team member has the information needed for his/her job” (Endsley & Jones, 1997, p.
`47). Shared SA incorporates not only information on system states, but also the effect of
`task status and actions of other team members on one’s own goals and tasks (and vice versa)
`and projections of the future actions of other team members. For a human–machine team,
`the same need exists. The machine will have certain expectations of human behavior built
`into it and needs to ascertain what actions have or have not been taken in relation to its pro-
`gramming. The human operator needs to have an understanding of not only what the ma-
`chine has done, but also what it is doing and will do next. Failures in this shared SA among
`humans and machines are well documented (Wiener, 1989).
`The tendency for sharing of information between parties may change with changes in
`system function allocation and LOAs associated with AA. This must be considered in AA
`design and interface design. The question can be raised as to whether the human operator
`will be able to continue communicating with automation effectively without performance
`implications when the mode of system automation changes dynamically, regardless of the
`quality of the interface design. The mode of system automation, the structure of the opera-
`tor’s role, and operator workload may inhibit critical information flow and, in the worst
`case, only allow the human to observe the system. Because of the manner in which automa-
`tion is structured in a supervisory control system, human operators are not permitted in-
`volvement in active decision making on a routine basis during system operations. Process
`control interventions can be used for error prevention, but they do not provide for regular
`communication between the operator and system automation. This is unlike other forms of
`automation, such as batch processing systems, where operators are involved in active con-
`trol of the system and communicate with the automation in planning and decision-making
`tasks on a regular basis, although the communication may be related to future processes.
`Active versus passive decision making has been identified as a critical factor in operator
`out-of-the-(control) loop performance problems, including a loss of SA (Endsley & Kiris,
`1995). Under supervisory control, operators are normally provided with high-level summa-
`ries of system functions handled by automation (Usher & Kaber, 2000). This form of feed-
`back may be sufficient for monitoring the safety of system states, but it is usually inadequate
`for decision making toward planning operations and so forth. This problem extends beyond
`interface design as it is rooted in the adaptive structuring of the system and the natural be-
`havior of human operators, although it may be affected by interface design changes. Re-
`search needs to identify how effective human–automation interaction can be maintained
`across LOAs regardless of changes in the role of the operator in order to promote SA and
`performance when DFAs occur.
`
`4.1. Establishing a Human–Automation Relationship and Potential
`Problems
`
`To ensure effective human–automation communication under AA, Bubb-Lewis and Scerbo
`(1997) said a working relationship between the human and the system must be developed.
`Muir (1987) offered some suggestions for developing this relationship in adaptively auto-
`mated systems, including providing operators with information such as the machine’s areas
`of competence, training operators in how the system works, providing them with actual per-
`
`Yuneec Exhibit 1026 Page 9
`
`

`
`46
`
`KABER, RILEY, TAN, ENDSLEY
`
`formance data, defining criterion levels of acceptable performance, and supplying operators
`with predictability data on how reliable the system is. Also, appropriate feedback mecha-
`nisms and reinforcement during training are important ingredients in developing a relation-
`ship and creating an effective human–computer team.
`Key problems in training and performance that can serve to undermine the human–auto-
`mation relationship and interaction under AA include human information misinterpretation.
`This problem may stem from the inability of the human to assess the intention of the com-
`puter system (Bubb-Lewis & Scerbo, 1997; Mosier & Skitka, 1996). As a result of these
`misinterpretations, Suchman (see Bubb-Lewis & Scerbo, 1997, p. 96) claimed that systems
`can lead humans “down the garden path,” sometimes never reaching a solution. To prevent
`this type of problem, Woods, Roth, and Bennett (1990; also see Bubb-Lewis & Scerbo,
`1997) suggested presenting the machine’s current state, goals, knowledge, hypotheses, and
`intentions to the human in a clear and