Automation Aids Research Articles

Autonomy and Automation Reliability in Human-Robot Interaction: A Qualitative Review

The effectiveness and reliability of automation aids are critical topics in the area of human-robot interaction (HRI). As more tasks are subsumed by robots and autonomous systems, it is important to examine the relationships between these entities and their human operators. Research to date has covered various manipulations of autonomy, but this broad body of research is in need of focus and consistency. The current study presents a qualitative overview of research regarding levels and reliability of autonomy/control and the effects they have on important HRI-relevant outcome variables. Results indicate that autonomy and automation aids operate uniquely for different tasks, and that there are many complex factors that can affect not only performance but also usability, confidence, and safety. Unresolved issues in the field and challenges and opportunities for future research are also presented.

Automation Dependency and Performance Gains under Time Pressure

Objective: An experiment tested a technique for encouraging appropriate human-automation interaction. Background: Operators often fail to make optimal use of automated aids, particularly when the aids are highly reliable. One way to discourage automation disuse might be to encourage automation dependence through time pressure. Methods: Fifty-two participants performed a simulated security screening task, searching for knives hidden in cluttered baggage x-rays. Participants were assisted by a diagnostic aid that was either 95%, 80% or 65% reliable, and were given instructions that asked them to make speeded or unspeeded decisions. Results: Participants showed higher levels of automation dependence under time pressure. This benefited overall performance in the 95% reliable condition. Conclusion: Time pressure encouraged heuristic dependence on automation aids, and benefited overall human-automation performance when the automation was highly reliable. Application: Data suggest a method for mitigating automation disuse.

Effects of in-vehicle navigation assistance and performance on driver trust and vehicle control

Driver trust in in-vehicle navigation systems may be an important factor influencing use of such automation. This research investigated the trust effect of in-vehicle navigation aiding performance that degraded over time and assessed differences in trust among automated versus human aids in a simulated driving task. Twenty participants drove through a suburb area following traffic signs and directions from one of the navigation aids, the human aid via a cell phone or the in-vehicle automation aid presented on a separate simulator display, with different levels of aid information accuracy in each trial (100%, 80% and 60%). A control condition was also investigated to assess any benefit of aiding in which participants were presented with a telemarketing survey instead of navigation information. A 2×4 mixed experimental design was used with the two levels of aid type (manipulated between-subjects) and the four levels of information accuracy (manipulated within-subjects). Driver trust was measured using a subjective survey of initial trust expectations, as well as subjective ratings of trust at the close of each trial. Driving speed control deviations and the number of errors in driver conformance with posted signs and navigation instructions were recorded. In agreement with prior work in other domains, results demonstrated drivers had a higher initial trust expectation for the automated aid than the human aid. However, once the automated aid was used, trust declined dramatically across degrading aid performance conditions. In general, there was a trend for trust to decrease more sharply with the automated aid than the human aid, but there was no statistical difference among the aids. Trust declined as both aids degraded in performance. Driving errors also increased as the navigation aid performance decreased, and the control condition produced the highest number of errors. This study demonstrated the role of driver trust in in-vehicle navigation aid use and has implications for designing navigation systems that support trust and performance.

Managing Multiple Uavs: Are we Asking the Right Questions?

The limiting factor in how many unmanned air vehicles (UAVs) a single operator may be able to simultaneously control successfully may be constrained by the workload experienced by that operator. Recently, the focus on the cause of the workload and performance variation primarily stemmed from the number of concurrent UAVs controlled by a single operator. The present experiment systematically investigated the number of UAVs controlled and the number of targets each UAV had to provide information on during its flight. Additionally, an automation aid was evaluated to determine if cueing facilitated performance. The results suggest that the tasks of each UAV are just as much or perhaps even more important to consider when evaluating the source of increased operator workload and performance differences. Further, the automation, as employed in this experiment was limited in providing a practical benefit to the operator.

Linear temporal characteristics of heart interbeat interval as an index of the pilot's perceived risk

The concept of adaptive automation is the first commendable step towards machine monitoring and support of the human operator behaviour and performance. Despite a large number of studies, the concept is still not sufficiently mature and more experimental studies are required to enable its application in the real world. As a small contribution in this direction, this experimental study made an effort to develop some techniques for online assessment of the pilot's perceived risk of the consequences resulting from their inability to execute a task adequately. Linear dynamical characteristics of 60 pilot's heart interbeat intervals were used for the assessment. Past research has shown that such ‘perceived risk’ influences the pilot's need for an automation aid. The preliminary results of this study suggested that the frequency and power of the heart interbeat interval dominant cycle, together with the characteristics of the energy contained in the power spectrum, can be sensitive indices of the level of perceived risk. The study techniques and preliminary results appear promising for incorporation into the algorithm of an intelligent online monitoring and alerting system on future aerospace vehicles to define the moments when the human supervisory controllers require an automation aid.

Unmanned Aerial Vehicle Flight Control: False Alarms versus Misses

Thirty-two undergraduate pilots from the University of Illinois School of Aviation performed simulated military reconnaissance missions with an unmanned aerial vehicle (UAV). Pilots were required to: a) navigate the UAV through a series of mission legs, b) search for possible targets of opportunity, and c) monitor system health. They performed the missions under three types of auditory auto-alert aids (a 100% reliable system, a 67% reliable system with automation false alarms, and a 67% reliable system with automation misses), as well as a non-automated baseline condition. Results indicate that while reliable automation can benefit performance relative to baseline in the automated task, the unreliable automation aids reduced performance to that of baseline or worse. The automation false alarms and misses harmed performance in qualitatively different ways, with false alarm prone automation appearing to cause more damage than miss prone automation to both the automated task and the concurrent target search task.

Case-Based Reasoning for Knowledge Discovery

The Georgia Tech Research Institute's goal in the Novel Intelligence in Massive Data (NIMD) Program, is to investigate certain aspects of an intelligence analyst's preferences and analytic strategies used in the process of discovering new knowledge. We are analyzing search strategies used by analysts in an attempt to understand their current task model. Based on this understanding, we will design and prototype a software tool that applies case-based reasoning in combination with other advanced reasoning techniques to help analysts perform knowledge discovery. The main objective of our work is the development, validation and incremental improvement of a set of knowledge discovery automation aids that significantly reduces the manual searches done by intelligence analysts and increases the quality and quantity of derived intelligence. We believe that these technical improvements will depend on our explicit understanding of the cognitive issues implied by the use of advanced reasoning techniques, integrated into a next generation NIMD prototype.

Impact of Contextual Information on Automation Brittleness

Brittleness, the inability to provide accurate assistance in all situations, is frequently an issue with complex automation and automatic decision aids. This paper examines a method of mitigating the impact of brittleness on overall system function. Using the task of planning an emergency descent for a commercial aircraft, this study found that the presence of contextual information in the presentation of an automatically generated emergency descent procedure might aid in mitigating the effects of automation brittleness. By providing pilots with rationale as to the design of the descent procedure, the pilots were better able to correctly determine why a provided procedure was or was not feasible.

Distributed Air/Ground Traffic Management for En Route Flight Operations

Distributed Air/Ground Traffic Management (DAG-TM) is a detailed instantiation of the Free Flight concept, formulated to serve as a foundation for NASA’s long-term research on advanced air traffic management concepts. Potentially revolutionary methods of air traffic management will be necessary to achieve the mature state of Free Flight. DAG-TM is a human-centered operational concept that seeks to minimize the impact of air traffic management constraints by leveraging new procedures and technological innovations in automation aids, information sharing, and supporting technologies. Under the DAG-TM paradigm of operations, flight crews, air traffic service providers (controllers), and aeronautical operational control personnel (flight planners/dispatchers) utilize distributed decision-making to maximize user flexibility and system throughput. This paper introduces the DAG-TM concept and provides a detailed description of the en route concept elements; the main purpose is to disseminate the DAG-TM concepts to the air traffic management community, and to motivate further exploration.

Automation Distrust Impacts on Command and Control Decision Time

Research into how an air defense command and control soldier will interact with automation and decision aid recommendations is evolving to show that, when trust in the automation is in question, a degradation in decision time of at least 30% will exist when compared to an operator that is not distrustful. Current work involves methods to decrease the temporal effects of distrust on operator authorization, including the accuracy of decisions when reliability is in question, using configural displays and other holistic display methods.

Cutting Stock Problem: A Heuristic Solution Based on Human Intuitive Strategies

The goal of the cutting stock operation is to maximize the material utilization by laying out maximum number of parts on a sheet. Historically, laying of parts on a sheet has been performed by humans. As the manual process is very time consuming, a number of heuristic and mathematical algorithms have been developed in recent years. However, these algorithms do not guarantee higher utilizations than those achieved by human operators. This study investigated the feasibility of extracting and generalizing the involved human strategies in laying out parts. The results indicated that it is possible to extract and form generalized strategies leading to higher utilization. A group, trained with these strategies, achieved higher utilization than the untrained group. The trained group also achieved higher utilization than the most promising mathematical algorithm. This implies that these strategies can be used to train new operators. Using these strategies and automation aids such as a graphics software, the layout process will be much faster and produce high material utilization.

Future air traffic management: technology, operations, and human factors implications

The NASA, the FAA and Eurocontrol have initiated programs of research and development to provide flight crew, air line operators and air traffic managers with automation aids to increase capacity in enroute and terminal areas, to support the goals of safe, flexible, predictable and efficient operations, and to ensure, and to enhance the safety of operations in the response to demands for capacity enhancement. Achievement of these goals is enabled by increased precision of knowledge of aircraft and asset position in space and time, by increased density and bandwidth of transmitted information among the airspace assets and users, and by increased service and information to the flight deck, ATC and airline operations centers. These aiding technologies are intended to support exploitation of timely and dynamic information on atmospheric hazards, traffic fluctuations, and airspace utilization. These systems are intended to support the human operators who maintain the authority and responsibility for safe, efficient operations through cognitive aiding technologies, to serve as cognitive orthoses (Reason, 1988) [12].

Cognitive Performance for Multiple Operators in Complex Dynamic Airspace Systems: Computational Representation and Empirical Analyses

This paper presents a set of studies in full mission simulation and the development of a predictive computational model of human performance in control of complex airspace operations. NASA and the FAA have initiated programs of research and development to provide flight crew, airline operations and air traffic managers with automation aids to increase capacity in enroute and terminal area to support the goals of safe, flexible, predictable and efficient operations. In support of these developments, we present a computational model to aid design that includes representation of multiple cognitive agents (both human operators and intelligent aiding systems). The demands of air traffic management require representation of many intelligent agents sharing world-models, coordinating action/intention, and scheduling goals and actions in a potentially unpredictable world of operations. The operator-model structure includes attention functions, action priority, and situation assessment. The cognitive model has been expanded to include working memory operations including retrieval from long-term store, and interference. The operator's activity structures have been developed to provide for anticipation through knowledge of the intention and action of remote operators. System stability and operator actions can be predicted by using the model. The model's predictive accuracy was verified using the full-mission simulation data of commercial flight deck operations with advanced air traffic management techniques.

Applications of Artificial Intelligence Techniques to Component‐Based Modular Building Design

Research in construction automation at the University of Reading led to the formulation of a computer‐integrated, component‐based construction system. The Reading Building System was rationalized for automation following a systematic study of the construction processes involved in the design and erection of a variety of building types, especially high‐tech offices. Computer‐aided design (CAD) packages were written that used Parts Set components as primitives and that offered flexibility in design that was so often lacking in earlier approaches to system building. At the same time, a family of automation aids was developed to manipulate the parts that were modeled in the CAD programs.

Future ATC automation aids based upon AI technology

The authors discuss two knowledge-based systems that are currently under development by the US Federal Aviation Administration, namely a simulation system for air traffic controller training and an automated problem resolution capability that is the central feature of the Automated En Route Air Traffic Control System (AERA 2). For each of these systems, the manner in which knowledge-based system technology is applied is discussed, along with the knowledge bases. The process for building, refining, and evaluating the knowledge bases is also presented. In the air traffic controller training system, ATCoach, the representation of ATC knowledge has been shown to be highly beneficial for both ATC training and airspace modeling. In AERA 2 automated problem resolution (APR) evaluations, the AI/knowledge-based system technology has proven to be indispensable in the representation of the experts' knowledge and the effective factoring of the knowledge into APR. It is concluded that these applications offer great potential for success and, consequently, represent only the first step in applying AI principles to ATC.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A Silicon Compiler System Based on Asynchronous Architecture

As one of the most promising approaches for the automatic generation of VLSI chips, the concept of "silicon compilers" was presented. A silicon compiler is a software system which accepts the behavioral or functional description of the chip and directly translates it to the layout patterns of the VLSI masks. In order to realize a design automation aid for full-custom VLSI chips for nonprofessional chip designers such as programmers, we have selected asynchronous architecture as the target model of the compiler and have been developing a silicon compiler system. A VLSI chip based on this architecture model consists of asynchronous modules which hold the control information internally. The behavior of the chip is controlled only by local communication among these modules. The silicon compiler system accepts the purely functional specifications of the chip, written by the hardware specification language ISPC, and directly translates it to the layout patterns in the form of CIF codes. By selecting an effective architecture model, it is possible to realize the silicon compiler system through simple translation programs which analyze the functional description of the chip and extract from it the configuration of components and the control information and through a library of the asynchronous modules which are generally defined with suitable parameters.

A microprogramming design laboratory

Sophisticated design aids are available to help automate microprogram development both for large, complex computers and for microprocessor-based systems. These aids include procedures for source code verification, object code generation, microprogram simulation, and interactive debugging. For various reasons, such aids are generally not well-developed for microprogrammable minicomputers. This paper describes a microprogramming laboratory developed for the Interdata 85 minicomputer which uses both the minicomputer itself as well as the time-sharing and batch facilities of a large computer center. Experience with the laboratory has shown that the facility for interacting with a microprogram is as important as the availability of complete diagnostic information. The argument is made that design automation aids for microprogramming minicomputers could be produced efficiently from standardized assembly and simulation programs.

A microprogramming design laboratory

Sophisticated design aids are available to help automate microprogram development both for large, complex computers and for microprocessor-based systems. These aids include procedures for source code verification, object code generation, micro-program simulation, and interactive debugging. For various reasons, such aids are generally not well-developed for micro-programmable minicomputers. This paper describes a microprogramming laboratory developed for the Interdata 85 minicomputer which uses both the minicomputer itself as well as the time-sharing and batch facilities of a large computer center. Experience with the laboratory has shown that the facility for interacting with a microprogram is as important as the availability of complete diagnostic information. The argument is made that design automation aids for microprogramming minicomputers could be produced most efficiently from standardized assembly and simulation programs.

A microprogramming design laboratory

Sophisticated design aids are available to help automate microprogram development both for large, complex computers and for micro-processor-based systems. These aids include procedures for source code verification, object code generation, microprogram simulation, and interactive debugging. For various reasons, such aids are generally not well-developed for microprogrammable minicomputers. This paper describes a microprogramming laboratory developed for the Interdata 85 minicomputer which uses both the minicomputer itself as well as the time-sharing and batch facilities of a large computer center. Experience with the laboratory has shown that the facility for interacting with a microprogram is as important as the availability of complete diagnostic information. The argument is made that design automation aids for microprogramming minicomputers could be produced efficiently from standardized assembly and simulation programs.

Automation aids tile manufacture

Automation aids tile manufacture