Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 26
2007-2008 Assessment of the Army Research Laboratory 3 Human Research and Engineering Directorate INTRODUCTION The Soldier Systems Panel of the Army Research Laboratory Technical Assessment Board (ARLTAB) reviewed programs of the Human Research and Engineering Directorate (HRED) within the Army Research Laboratory (ARL) during visits to HRED’s primary site at Aberdeen Proving Ground, Maryland, on July 17-19, 2007, and on June 23-25, 2008. A subset of the panel met with several of the HRED scientists to review the quickly developing program in neuroergonomics on both December 17, 2007, and June 9, 2008. In the briefings during 2008, the HRED presenters acknowledged benefiting from earlier comments of the panel, and this was evident in their subsequent presentations. As general background, HRED is organized as two divisions to conduct research and development efforts to enhance soldier performance. The Soldier Performance Division conducts a broad-based program of soldier-centered basic and applied research and technology testing and evaluation directed toward maximizing battlefield effectiveness. In contrast, the Human Factors Integration Division conducts laboratory and field data analyses, develops modeling and human simulation programs, and performs applied research to ensure that soldier performance requirements are adequately considered in technology development and system design. Tables A.1 and A.2 in Appendix A respectively show the funding and staffing profiles for HRED and the other directorates and indicate the relative levels of effort now devoted to basic and applied research and service activities. The framework for the assessment, as presented by the HRED Acting Director, emphasized HRED’s dual objectives of providing science and technology to enable transitional capabilities for the smaller, smarter, lighter, and faster future force, while also seeking opportunities to accelerate technologies directly into the current force. The HRED Acting Director has been applying a new analytical planning framework, referred to as the Mission and Means Framework (MMF), to plan and coordinate various projects. This approach, although new and unproven, may have merit in that it provides a strong mission-oriented context for various projects; there is concern, however, that it may stifle creative think-
OCR for page 27
2007-2008 Assessment of the Army Research Laboratory ing and breakthrough innovations by HRED scientists who wish to study more fundamental problems that do not easily fit into a particular defined mission. In other words, HRED must more seriously consider ways in which it can best achieve the goal of performing cutting-edge applied research in addition to maintaining its cutting-edge basic research, given that the large majority (approximately 90 percent) of its funding is from 6.2-level or above sources, which target applied research. Because of the existing funding allocation for mission-oriented work, it is important to begin this discussion by revisiting the definition of “applied research,” which in the Army is defined from the DoD Financial Management Regulation which states: Applied research is a systematic study to understand the means to meet a recognized and specific need. It is a systematic expansion and application of knowledge to develop useful materials, devices, and systems or methods. It may be oriented, ultimately, toward the design, development, and improvement of prototypes and new processes to meet general mission area requirements. Applied research may translate promising basic research into solutions for broadly defined military needs, short of system development. It includes studies, investigations, and non-system specific technology efforts. The dominant characteristic is that applied research is directed toward general military needs with a view toward developing and evaluating the feasibility and practicality of proposed solutions and determining their parameters. Applied research precedes system specific technology investigations or development.1 A significant proportion of HRED staff resources is allocated to the support of Army programs under development and related field work; that work is not assessed by the Board, whose focus is on the research supportive of the program and field support. HRED is in a unique position to improve the Army’s development and use of advanced technologies with targeted, cutting-edge, applied research that will assist in determining the feasibility of various weapons systems and design concepts early in the development process. More specifically, HRED is developing new methods, models, and human performance databases to aid in designing specific Army technologies and in evaluating and improving existing mounted and dismounted soldier systems. To excel in such endeavors, however, will require HRED to continue to perform complex, human-centered, scientifically sound studies that are motivated and defined by a staff and management that have a deep understanding of the complexity of various soldier-system interactions associated with tasks performed within emerging hardware and software technologies. CHANGES SINCE THE PREVIOUS REVIEW Five major changes are evident in the Human Research and Engineering Directorate since the previous published review:2 Planning for a major new program in neuroergonomics. This initiative is expected to result in fundamental studies of how to use newer neurological monitoring methods and data along with cognitive and human performance models to improve the decision-making and performance capabilities of soldiers in the field. Because this is such a new area of investigation for the Army, a collaborative program that includes multiple universities, using a Collaborative Technology Alliance (CTA) arrangement, will be the best way to meet the high expectations for this program over the next 5 to 10 years. 1 Department of Defense, DoD Financial Management Regulation, Vol. 2, Ch. 5, Washington, D.C., June 2006. 2 National Research Council, 2005-2006 Assessment of the Army Research Laboratory, Washington, D.C.: The National Academies Press, 2007.
OCR for page 28
2007-2008 Assessment of the Army Research Laboratory A continued increase in the number of HRED-relevant projects that have been funded by the Army Research Office (ARO). In 2004, the ARO funding for such projects was only about $0.5 million. In 2006 it was almost nine times higher, at about $4.5 million, and in 2008 it will have increased again, to about $7.5 million. Such a trend appears to support a strong interest by the scientific community at large in the types of research areas being investigated within HRED. At a minimum, the ARO projects provide access to a broad array of experts for future collaboration. The Board looks forward to learning more about such collaborative efforts and hopes that these result in additional joint (ARL and outside collaborators) publications, visitations, workshops, and seminars. The near completion of the very sophisticated Environment for Auditory Research (EAR) facility. This facility is composed of five different sound exposure laboratories and a control room. The facility can present to subjects a large array of auditory experiences, some of which may be unique to this facility, raising the possibility of conducting original basic research that could not be performed anywhere else. A shift in the activities within the area of network science research. The 2005-2006 assessment report recognized the potential value from the work being done in social networks and the cognition area, and it encouraged continued activity that leveraged interactions with others within and outside ARL. Since that report, there has been considerable evidence of the latter in the form of grant writing, workshops, and other activities discussed more completely later in this report. A very positive development also has been HRED’s creating collaborative research efforts, such as the Multidisciplinary University Research Initiative (MURI) with Central Florida University, establishing the Davis Fellow position, and initiating other joint projects with university faculty and independent researchers in this area. HRED staff have focused their attention on developing a network science portfolio that leverages multiple sources of funding to support basic research on the cognitive and social impacts of networked operations. This includes a basic research (6.1) Network Science Army Science Objective (ASO) that is being funded through ARO and a significant new Network Science Collaborative Technology Alliance funded through the Computational and Information Sciences Directorate (CISD). Notable in these thrusts is the emphasis on multidisciplinary collaboration, as well as outreach to leverage the expertise in academia and industry to advance the state of the art in network science as it applies to emerging needs of the Army. A change that is not positive: the shift away from an emphasis on the development of human modeling capabilities that were highlighted in the 2005-2006 assessment report. There have been some enhancements in the usability of the Improved Performance Research Integration Tool (IMPRINT), which is the primary human-performance-modeling algorithm and software developed by HRED to predict soldier task time requirements and mental workloads. These enhancements include improved IMPRINT usability (e.g., improved formats of outputs and better accessibility) and functionality (e.g., enhancements to workload scales) over the past 2 years, during which congressional funding supporting the IMPRINT program shrank (from approximately $2 million in FY 2006 to $1.14 million in FY 2007) and then evaporated (replaced in FY 2008 by $270,000 in ARL mission funding). The FY 2008 funds have been devoted largely to the maintenance and upkeep of IMPRINT, and there appears to be a risk that HRED will continue its recent practice of largely restricting its use of the IMPRINT human task analysis model in its current state to perform first-cut human factors analyses of anticipated hardware design problems. Although the IMPRINT model is useful in its current state, many different empirical studies are underway or have been completed within HRED and elsewhere that could continue to enrich and improve the IMPRINT model, and this was encouraged in the previous ARLTAB assessment report. Unfortunately, little has been done to translate new empirical findings into the model to make it more robust and valid for future design analyses. Furthermore, it was not clear that the existing model is being used very often to structure and plan future empirical studies conducted within HRED. An exception
OCR for page 29
2007-2008 Assessment of the Army Research Laboratory to this was a study conducted in 2007 of the task demands on a gunner in a mounted combat system. IMPRINT was used to determine whether this soldier could also control an unmanned ground vehicle (UGV) while moving. The IMPRINT model showed that it might be feasible, and in so doing provided several well-structured hypotheses for subsequent study. Follow-up empirical studies indicated that the gunner would miss too many targets while attempting to control the UGV (see the later discussion of human-robot interactions). This example is highlighted here because it illustrates two common research issues: the model was helpful in the early planning of a complex empirical study, and the IMPRINT model’s ability to make accurate predictions about human behaviors needs to be improved in order to better simulate certain types of complex soldier tasks. ACCOMPLISHMENTS AND ADVANCEMENTS Environment for Auditory Research Facility The completion of the Environment for Auditory Research facility is a very significant event. The listening laboratories in this facility will allow unique studies of how different sounds are identified, localized, and used for communication in a variety of environments. Given the uniqueness of the five different physical listening laboratories, however, it is very important to specify the physical acoustics of each of the rooms and publish the results soon so that other experts from a variety of disciplines can appreciate (and possibly use) the facilities in the future. In terms of collaboration, the EAR facility group is in conversation with one or two research groups, including the Air Force acoustics research group. Work within the Sensors and Electron Devices Directorate (SEDD) on acoustic devices and algorithms may also be applicable. This type of cooperation is certainly needed in such an important area. One of the advances that this facility provides is the opportunity to examine auditory capabilities and communication limitations in a variety of very complex acoustic environments, both real and virtual. However, research in auditory communication performance may not exploit the full capabilities of this facility. From a medical standpoint, loss of hearing is a major issue in the Army. In this context, it is good to know that this facility has a medical audiologist on the staff, particularly given the strong interaction between the ability to segregate sound sources and hearing loss. However, collaboration with other groups concerned about noise-induced hearing loss should be given some priority. Groups at the U.S. Army Aeromedical Research Laboratory at Fort Rucker, Alabama, are performing studies to document the full extent of health hazards due to noise exposure in the Army and are developing programs in accordance with Army Regulations 40-10 to provide systems and materials to preserve hearing. Integrated hearing protection and hearing communication systems also are being developed at Fort Belvoir, Virginia, in the Program Executive Office-Soldier Programs. All of these would appear to be relevant in the planning of future research in this new facility. The EAR facility staff is quite strong, but it is not clear how the staff will balance applied development work and publications in peer-reviewed journals. In the past they have published a good number of technical reports and conference proceedings, with some on the Internet, but more publications in peer-reviewed journals are necessary in order to gain the credibility and attract and retain the best people to this new facility. As for the physical facility, of the three anechoic rooms the Distance Hall is the most novel anechoic chamber (for human psychophysical testing) by virtue of its size and configuration. The multiple speakers allow the simulation of variously placed sound sources and reflective walls. As a result, this space provides an opportunity to examine the psychoacoustics of distance perception and motion perception, as well as to answer more fundamental questions concerning the psychoacoustics properties of
OCR for page 30
2007-2008 Assessment of the Army Research Laboratory environments with multiple (virtual) reflections and sources. This space provides a unique opportunity to evaluate the sensitivity of human observers to differences in sound source depth, as well as changes in depth (motion). The speaker (distant) arrays also provide virtual walls to study complex sound reflection effects on depth perception. The other two anechoic spaces are impressive in size and development. The Sphere Room appears to be slated for basic and applied sound source localization research, including the rapid measurement of head transfer functions both with and without helmets. The Dome Room has characteristics appropriate for experiments concerning the perception of sounds in space and the motion of sounds in azimuth with sound reflections. The new physical facility is impressive, and several well-qualified staff members are in place. It is not as clear, however, that they have identified the research questions that this HRED facility can uniquely answer. For example, What are the classic questions in psychoacoustics that can be revisited in this new facility? How will the staff see the work through from conception to publication? For each component of the laboratory, the general tenor of the research to be conducted was provided to the Board, but details were lacking. This was particularly true of the Listening Laboratory, where the description was limited to the auditory capabilities of the space and not the science that it will subserve. The overwhelming weakness was the absence of a coherent research plan as to how the EAR facility would attain its lofty goals. Given that construction has been in the works for several years, one would expect a complete description of the research that is to be initiated over the next couple of years, including the hypotheses to be tested, the measurements to be made, the source of experimental subjects, and other experimental factors. The absence of details at this time is a concern and in a few instances was alarming. In summary, EAR is a wonderful facility, and some excellent people are in place. Now is the time to reach out to other experts, to set up workshops and visitations, and to plan carefully an equally strong program of both basic and applied research that will result in this being seen as a first-class national resource in the future. HRED should consider collaborating with acoustics experts among the staff of the ARL Sensors and Electron Devices Directorate. Night-Vision Research Night-vision enhancement technologies are rapidly being deployed in both the civilian and the defense sectors. HRED has been involved in human factors studies of night-vision devices for many years, but perhaps because of the limited nature of the briefings provided on this topic, it is not clear how HRED’s work fits in with the broader work in this area. A central concern has been sensor fusion—the need to combine multiple sources of information into a display that a user can interpret. Groups in many fields are working on these issues. Military laboratories in several countries publish in this area. In addition, major automobile-manufacturing groups are involved in fundamental studies of night-vision enhancement technologies. It is not clear whether the HRED researchers have collaborated or consulted with the many groups around the world that are studying this problem. Certainly collaboration would be appropriate with the visual scientists and engineers at the Army’s Night Vision Laboratory at Fort Belvoir. An HRED-organized workshop on this topic might provide a means to focus some of the activities better. The research on sensory fusion that was reported, while of importance, does not seem likely to make significant contributions to an understanding of the perceptual issues in sensor fusion. The specific projects do not appear likely to produce results that could be published in the peer-reviewed literature—an indication that the approach being taken is not current. For example, the sensor fusion algorithms under study were very basic. A quick survey of the literature suggests that other laboratories
OCR for page 31
2007-2008 Assessment of the Army Research Laboratory are experimenting with much more sophisticated algorithms for cue combination. Many of these seem to be based on Bayesian approaches in which an effort is made to determine which of several signals is providing the most information at this location and at this time. Similar approaches are used in other areas of vision research (e.g., models of eye movements and models of depth cue combination). Once again, contact with researchers who work on cue combination, Bayesian theories, and ideal observer approaches would be useful. In this context, some of this collaboration could be established by greater participation at national vision and/or machine vision conferences. The Board, in referring to past HRED research, especially that discussed in the 2005-2006 assessment report, noted that when presenting an image to one eye that is different from the image visualized by the other eye, a phenomenon known as sensory rivalry occurs. This can result in perceptual errors and cognitive distraction, particularly when dynamic displays are involved. The research reported during the current assessment did not deal with this important topic specifically, but rather relied on eye-movement studies to indicate how and when sensory fusion would occur. It was not clear what hypotheses were being tested with this approach. The work being done is not being published well. Although HRED has produced several conference proceedings and technical reports, there are only a few papers in refereed journals over the past several years. The dearth of significant publications should be of concern both to the responsible researchers and to those responsible for directing the overall effort. Cognition and Neuroergonomics Collaborative Technology Alliance The panel’s select group of experts on neuroergonomics was very favorably impressed with the current accomplishments of the neuroscience group in HRED. The use of the CTA mechanism, joining industry and academic groups, is appropriate. The HRED group now includes outstanding new, young scientists with excellent backgrounds. The motivation of the group appears high, and the members have demonstrated major advances over the past year in their understanding of the field and what they can contribute. The neuroscience group’s organization of the 1-day workshop held on May 8, 2008, was an excellent method of informing neuroscientists and cognitive scientists of the Army’s needs and of quickly evaluating various research groups that could respond to a CTA announcement. This workshop allowed 18 different research groups to present projects (self-selected) on a variety of topics relating to applications in the broad field of neuroscience to address Army needs. Some were excellent—for example, the project on independent component analysis of electroencephalography. Others were less so. The HRED scientists reviewed these presentations at a special meeting on June 9, 2008, with a subset of the panel members. The HRED staff indicated that the pending CTA announcement would not cover all areas addressed by the workshop attendees but would focus on a few areas that seemed most promising for basic research and that would have clear applications to the Army’s needs. In particular, the staff indicated their intention to emphasize the development of a portable monitoring system of neurological functions that could predict when a soldier’s cognitive abilities were being overly stressed by particular tasks and the environment to the extent that performance would be compromised. The details of the CTA announcement were not available at the time that this report was drafted, but HRED should develop advanced cognitive performance models to form the basis for hypotheses to be tested, and to relate various neurological measurements to the prediction of human performance capabilities and mental workload.
OCR for page 32
2007-2008 Assessment of the Army Research Laboratory Human-Robotics Interaction Current military operations are employing robotic systems in unprecedented numbers and roles. Many types of unmanned aerial vehicles (UAVs) are used for intelligence, surveillance, and reconnaissance (ISR); for targeting and for tactical intelligence; and even for tactical engagement. Unmanned ground vehicles are being employed to deal with threats. Most of the systems in use require the attention of at least one or more operators for each vehicle. The potential ground combat leverage available through unmanned and robotic systems will never be realized until the ratio of operators to unmanned systems is reduced. Although the use of multiple robotic systems that incorporate all types of unmanned vehicles in military operations has been advocated, multiple robotic systems are unrealistic under current operating protocols. The Army’s Future Combat Systems include, conceptually at least, multiple types of robotic components; their leverage could be increased significantly by one-to-many (human-to-robot) control. Achieving technological superiority through the one-to-many control paradigm is one key approach to countering the higher leverage obtained by potential enemies employing an asymmetric approach to engaging U.S. forces. During the 2007 meeting at HRED, a number of HRED human-robotic interaction projects were described. The HRED human-robotics work was of relatively high quality. Much of the presented research focused on single platform systems, however. The research focus was on fundamental human-robot interaction questions, unlike the previous review during which much of the research focused on robotic perception, mission packages, and supervisory control of single platforms. The HRED group has a unique opportunity to begin a program of research that focuses on understanding how multiple UGVs per operator can be controlled in real-world scenarios (see the discussion in the “Opportunities and Challenges” section below). Network Science: Social Networks and Cognition As the 2005 report of the National Research Council entitled Network Science points out, interacting networks in the physical, information, cognitive, and social domains are ubiquitous in U.S. military operations, and they are increasing in importance with the growing efforts to transform the U.S. military into a force capable of network-centric operations.3 In this context, the domain of network science is defined broadly within ARL. It encompasses the range of phenomena emerging from the introduction of network-centric operations and needing to be understood and addressed to support the warfighter. Such phenomena range from understanding the physical characteristics of networked operations (e.g., characteristics of sensors and radios), through the communication level (routing, self-configuring networks), through the information level (e.g., secure information flows), and on up to the impact of network characteristics on the performance of individuals and teams operating in a networked environment. HRED research activities and goals in network science cover social network research as well as research on the impact of network operations on human cognition. The early in-house studies in network science presented by HRED in 2007 were not very impressive, but they did allow the HRED staff to become more familiar with the methods and models needed to perform high-quality research in network sciences. It would appear that, appropriately, the staff has now planned a much more aggressive program of research, as noted below. HRED is to be particularly commended for its success in leveraging ARO funding to advance the basic research foundation relating to human decision making as it is affected by Army command-and-control structures. An ASO program under ARO funding currently planned for the 2008-2012 period 3 National Research Council, Network Science, Washington, D.C.: The National Academies Press, 2005.
OCR for page 33
2007-2008 Assessment of the Army Research Laboratory will involve conducting experiments and observations to explore interacting network effects on human decision making. HRED is successfully leveraging ARO funding to foster greater interdisciplinary dialogue within and outside ARL. For example, a Davies Fellowship under ARO sponsorship has been obtained to fund a mathematics professor from West Point Military Academy to work with HRED behavioral scientists to explore the impact of network structures on human decision making. HRED also organized a Network Strategic Technology Initiative workshop that brought together leading researchers in network science to contribute to the Army’s understanding of the state-of-the-art research in network science and its relevance to warfighter issues. HRED, in conjunction with CISD and SEDD, is initiating a major new Network Science Collaborative Technology Alliance that is intended to develop the basic research foundation to enable modeling, design, analysis, prediction, and control of secure tactical communications, sensing, and command-and-control (decision making) networks. This includes research on the impact of networked processes on individual and distributed team decision making, from both individual cognitive and social network perspectives. The research thrust to be taken is in the planning stages. Nevertheless, this is clearly an important research initiative that has the potential to advance the state of the art in network science as it relates to critical Army needs, and the Board looks forward to following the progress of the research. HRED is to be commended for initiating 6.2-level network science research that is directly focused on an urgent Army research need—analyzing the impact of complex dynamic network-centric environments on individual and team cognitive and collaborative performance. As explained in HRED’s description of its 6.2 Tactical Human Integration of Networked Knowledge program, network-centric operation involves an abundance of information received from many different sources (human and sensor) and presented across multiple modes (text, audio, visual). Human decision makers are currently unable to make effective use of this information for reasons including information overload as well as network bandwidth and other hardware constraints that introduce lags, information loss, and degradation, leading to performance problems that include attention misdirection and poorly calibrated trust in the information received. HRED is initiating a multiyear program—the Tactical Human Integration with Networked Knowledge (THINK) Army Technology Objective—scheduled to start in FY 2009 and to extend through FY 2012, to analyze the contributors to performance problems and identify methods for overcoming these problems. This work will bring together experts in cognitive science, social network science, and computer science to work collaboratively to develop and evaluate methods to train and improve information sharing, decision making, and collaboration in networked operations. It will also aim to develop improved methods and guidelines for information aggregation and alerting so as to enable more effective attentional focus on high-priority issues, produce better trust calibration, and improve the quality of individual and distributed collaborative decision making. This program includes laboratory and field experiments intended to validate the impact of proposed enhancements in training, social network organization, and new information-aggregation and alerting concepts. Biomechanical Modeling Research In the previous ARLTAB assessment report, it was noted that biomechanical analysis tools were used to understand the basis for injuries incurred by persons lifting heavy components during bridge building. The ability to merge new cognitive models and biomechanical models of soldiers into the IMPRINT framework has been recommended in the past, although apparently funding has not been provided recently to pursue this development. Nonetheless, two recent biomechanical studies both deal with very real and important problems for the Army and use biomechanical modeling methods that
OCR for page 34
2007-2008 Assessment of the Army Research Laboratory could in the future enhance the physical-task-analysis models within the IMPRINT program. The first of these studies explored the effects of the mass of various handheld weapons during dynamic targeting motions. The study had not been completed, but the biomechanical analysis methods were appropriate, and preliminary results showed that adding mass to today’s weapons has a deleterious effect on quick aiming motions. The second biomechanics study was to determine the injury risk factors during long marches. It explored how certain foot and leg motions could produce stress fractures in the lower leg bone. Once again, biomechanical models were used to understand subtle motions and provided a means to predict how certain types of gait motions raised the risk of injury. These continuing studies illustrate the use of human modeling methods to plan the research, analyze the results, and eventually provide simulations of soldier performance characteristics in a variety of task scenarios. As in the cognitive modeling area, this type of modeling and empirical research should be continued so as to produce more useful and accurate simulations of soldiers’ physical endeavors. OPPORTUNITIES AND CHALLENGES It should be clear from the preceding sections of this chapter that many worthwhile applied research and development actions have taken place since the previous review. There are many opportunities and challenges that exist within each of these areas. The most significant of these are discussed below. Neuroergonomics The opportunity to work in the rapidly developing area of neuroergonomics is noteworthy. To lead this effort, HRED has assembled an excellent team of researchers who have engaged many others in the field to define a coherent program of basic and applied research. This approach is appropriate. The challenge will be to find the effective mix of neurological, cognitive, and human performance scientists and to bring them together in a collaborative team to further the understanding and modeling of human decision making and actions that affect a soldier’s effectiveness. It also will be necessary to ensure that the HRED-supported research is complementary to a number of other similar studies supported by the Department of Defense in this arena and to ensure that this complementary research is supported by ARO. Advanced cognitive performance models should be developed to form the basis for hypotheses to be tested and to relate various neurological measurements to the prediction of human performance capabilities and mental workload. Environment for Auditory Research Facility The near completion of the EAR facility provides another opportunity for HRED. This facility could allow the researchers using it to become leaders in the area of auditory performance. The team of researchers appears to be well prepared to perform cutting-edge auditory studies. Their challenge will be to develop a formalized and coherent set of studies that take full advantage of this outstanding facility, while meeting the unique needs of the Army both to protect soldiers against noise-induced hearing loss and to improve auditory communication and performance. There is a risk that both investigator and laboratory time could be absorbed by short-term practical questions, such as how different helmet designs influence sound localization. If time is not set aside for more exploratory basic studies, the scientists may not remain at the cutting edge of the research, and it will be difficult to attract the best scientists to the laboratory, thus losing the advantage now provided by such a well-conceived physical facility.
OCR for page 35
2007-2008 Assessment of the Army Research Laboratory Vision During the 2007 briefings, HRED discussed the problem of binocular rivalry. This phenomenon arises when different images are presented to a person’s two eyes. This is a specific problem within the broader problem of sensor fusion and cue combination. It arises in HRED settings when vision enhancement and communication devices present one set of imagery to one eye and another set to the other eye. This is a very important problem, and HRED has many opportunities to enhance the knowledge base in this area. Given that this was a research topic in the past, it is disappointing that no results seem to have appeared in the scientific literature. More generally, the rate of publication in this area is very modest. The list of publications during the 2005-2008 period appears to show 10 publications from the vision group. Of these, only one has been submitted to a peer-reviewed journal. All of the other papers appear to consist of conference proceedings or ARL technical reports. In general, the vision group is working on interesting problems that have very significant basic and applied potential. The information presented by HRED suggests that this potential has not been realized. If the group is to continue working in this area, it should develop much more extensive contact with the broader community working in this and related areas, and HRED should consider the sort of commitment of resources and personnel that have positioned the auditory group to make significant contributions in this arena. Network Science In the network science area, major opportunities exist at ARL generally and in HRED in particular to build on their existing unique network capabilities. More specifically, the opportunity for ARL is to create a domain that captures the unique characteristics of the research laboratory by assembling staff that can address the cross-disciplinary problems inherent in this area and ensuring that they address issues affecting the mission of supporting the soldier. Network science methodology in general and network metrics in particular are still in the early stages of development. ARL is well positioned to advance the state of the art in these domains if it can pull together an effective team of people from the physical, mathematical, software, and social and behavioral sciences. Unlike many other research institutions, ARL can gain access to network data gathered from varied simulations and field exercises as well as from real interactions among network members—these members including not only soldiers, but also robots and other network-based information agents. Essentially, such contexts present unparalleled opportunities to develop new network research paradigms as well as to assess the reliability and validity of existing network metrics. The primary challenge for the network science group at HRED is to execute the current research projects and become more involved with future streams of research. A great deal of energy is being expended and much activity is occurring, but there is little completed work in this area available for evaluation by the Board. This is understandable, because the network science program is early in its development. A second challenge is to move interesting, ongoing efforts and future planned work to completed outcomes, and to publish this work quickly so as to establish and enhance credibility with peers in this arena. Workload Modeling The HRED human factors design aid referred to as the IMPRINT workload simulation model continues to represent a major success story for the directorate. It appears not to have evolved in functionality a great deal since the previous ARLTAB review. This is unfortunate, because there is a great need to have a robust human factors (HF) analysis tool for planning research studies and for assisting Army contractors in meeting HF requirements within the MANPRINT (Manpower and Personnel
OCR for page 36
2007-2008 Assessment of the Army Research Laboratory Integration) program. The challenge will be to prioritize the functionality most needed and to develop software projects that would most effectively meet the Army’s needs to improve the design of future human-hardware task systems and to better train future warriors. To do so will require that the IMPRINT model be better able to simulate the effects of various complex perceptual and cognitive tasks, perform dynamic biomechanical motion and vibration simulations, and provide predictions of operator effectiveness and mental loads when the operator is controlling multiple UGVs. There also is the need to develop an IMPRINT model that is capable of analyzing the workload imposed on soldiers while operating as a team, and for long periods under high mental and physical loading. There also is a need to understand what type of formal training is required so that users can fully implement and accurately use the IMPRINT model to simulate various design scenarios, especially since the model can be operated with different levels of complexity and functionality. Robotics As has been stated previously, a real benefit of using UGVs will be realized when several UGVs can be controlled by a single operator. Although the present work on the interactions of an operator and a single UGV are extremely important, ARL should consider developing an enterprise-wide program dealing with the semiautonomous coordination of multiple robotic systems (i.e., the networking of information and autonomous actions taken by the robotic systems with varying degrees of human intervention to further mission objectives). To accomplish such coordination will require research to further define the supervisory control structure for groups of UGVs and UAVs, using smaller groups of human operators than are now required. Building on a strong foundation of research on individual robots, ARL also can bring to bear significant talent from HRED to address human-system integration, along with robotics-related work in the Sensors and Electron Devices Directorate and the Computational and Information Sciences Directorate. This cross-directorate work will provide a means to address actual, rather than simulation-based, platform-to-platform and platform-to-human communications and mutual awareness. A key advantage of this approach is that it will position ARL to serve Army needs in a large variety of contexts and concepts of operations, regardless of the specific properties of the individual unmanned systems that are ultimately developed or acquired by the Army. It is believed that commercial developers are outpacing the Army’s in-house efforts in individual robotic systems, despite the fact that multiplatform coordination and supervisory control of large numbers of robotic systems represent new frontiers that for now are relevant mostly to military rather than commercial users. Strategic application of resources to these problems can position ARL to enable the Army to excel in the operational use of unmanned, remote robotic systems. Biomechanics As to research addressing the physical requirements of manual tasks of soldiers, there is a very large opportunity to enhance the modeling of the musculoskeletal system of soldiers in order to predict their physical performance capabilities in such tasks. Most of the existing modeling has relied on overly simplistic, structural representations of both the anatomy and the physiology that govern human exertions of all kinds. The types of complex perceptual-motor tasks required of soldiers demand that the highest quality of biomechanical modeling and empirical studies be available if the Army is to understand the environmental and task factors that affect a soldier’s performance capabilities. The staff at HRED appears to have the fundamental biomechanical knowledge and some of the physical resources
OCR for page 37
2007-2008 Assessment of the Army Research Laboratory necessary to push forward on such research. The challenge will be to structure the research in such a way that it provides further insights as to how the physical models now used as part of the IMPRINT program can be improved. Because a number of academic and industrial research groups are working in the area of biomechanical modeling, a workshop should be held to explore further which types of existing biomechanical models are most appropriate to guide and enhance the types of research most needed by the Army. OVERALL TECHNICAL QUALITY OF THE WORK There is much to admire about the progress that has been made over the past 2 years at HRED. In particular, there appear to be some new, well-trained researchers on the staff who understand the need to perform applied research that is of high quality in a scientific sense. In this context, the near completion of the EAR facility and the new funding being provided for the neuroergonomics program provide new resources (physical and financial) that are unique. Unfortunately, since formal research plans were lacking in detail, it was made clear to the Board how the staff will best use the new resources in the EAR facility and the new funds for the neuroergonomics program to balance research that could provide both scientific breakthroughs and solve important military problems. If one looks at the publications coming from HRED over the past couple of years in all the areas, not just the two areas mentioned above, the number of peer-reviewed journal papers is not very impressive. Most of the publications are in technical reports and proceedings. These are certainly helpful: they show that the staff is capable of reporting their methods and findings to a limited extent, and they can represent the only viable outlet for some of the directorate’s research that is not experimental or that involves single large exercises or simulation experiments. However, such reports often lack the clarity needed for others to fully evaluate and hopefully come to respect the cutting-edge work being done in HRED. The six areas of concentrated research reviewed in this chapter are of vital importance not just to the Army, but to society in general. This list includes the following: (1) providing a better understanding and the means to enhance audiometric performance—a major problem for older individuals; (2) understanding neurophysiology at a level that predicts when a person is cognitively incapable of performing certain tasks; (3) presenting networked information to people in a fashion that can be quickly and accurately understood and acted on by one or many people; (4) being able to control multiple unmanned vehicles and tactical resources with minimum human interventions; (5) being able to understand and predict the physical capabilities of soldiers to perform complex and fatiguing manual tasks; and (6) providing the means to be able to perceive objects while in darkened environments. These all are very important and scientifically challenging problems. They all require study by teams with multidisciplinary backgrounds, which appear to be available in HRED for most of the areas. However, closer working collaborations (e.g., co-authored papers with experts outside HRED) are needed in all the areas to complement and enhance the capabilities of the staff. In some cases this is being done well, but not in all the areas. As noted at the beginning of this chapter, the closer alignment of ARO and HRED research will address this recommendation. From a methodological perspective, most of the areas of research conducted by HRED are strictly empirical. In some cases models are used to justify particular types of empirical studies. For instance, the IMPRINT task-analysis model has been used to provide some limited performance and mental-loading predictions associated with the performance of various complex tasks that were being considered for future empirical study. Select biomechanical and workspace-analysis models also have been used to understand the cause of certain types of injuries. HRED should continue this trend; many more future studies should include the use of analytical models during the planning of experiments. Such efforts
OCR for page 38
2007-2008 Assessment of the Army Research Laboratory can often lead to more efficient laboratory studies and, more importantly, can assist in allowing the results obtained from small empirical studies to be compared to other studies to gain general validity and applicability. Since the origins during World War II of organized research that was meant to understand and model human-hardware system interactions and consequences, the military has been the largest benefactor and supporter of such work. One might surmise that after 65 years of such work, there is not much that has not been addressed in this area. Yet the operational complexity of current military systems, not to mention future systems, demands that one know much more about the mental and physical attributes of the soldiers who are expected to operate and maintain these systems under the most arduous conditions imaginable. George Fisher, former chair of the National Academy of Engineering (NAE), noted in his 2000 address to the NAE, that we are in the Dark Ages when it comes to designing systems that are convenient for people to operate. Indeed, it is estimated by some that fewer than 10 percent of currently graduating engineers receiving a bachelor’s degree have had even one ergonomics course, and fewer than about 2 percent of engineers receiving a Ph.D. degree have had such exposure. Given this situation, is it any wonder that the military continues to be plagued by hardware and software systems that are extremely difficult to operate effectively and safely and to maintain? HRED identified six areas of concentrated research and development and requested and supported this review of those areas. The areas selected are, in general, highly appropriate and important not just for improving military operational effectiveness but also for improving the quality of life for all people. Although this report raises questions about the quality of the research in some of these areas, it is clear that most of the staff are capable of performing outstanding applied research in the various areas reviewed. It also has been acknowledged that the facilities are being improved to support the empirical studies that are needed. Continuing to pursue opportunities for more collaborative interdisciplinary research would contribute further to HRED’s studying and modeling of complex real-world conditions. For example, bringing together investigators focused on auditory processing with cognitive neuroscientists might strengthen this area and provide more useful results. A growing awareness of the importance of engaging other similar research groups seems to be taking place in some of the topical areas, but there should be more workshops and visiting senior scientist positions, along with support to publish more papers in peer-reviewed journals and with co-authors from different laboratories. Over the past 2 years, not much tangible in terms of new major findings has resulted. However, there exists a great deal of excellent potential for HRED to become a first-in-class research organization in several areas. To do so will take leadership that understands and respects the complex issues involved in performing cutting-edge, human-centric research and model development. Such leadership must manage the sometimes conflicting research goals resulting from the need for fast evaluations of new technologies and systems that are being rapidly deployed, versus providing scientifically valid, predictive models, methods, and principles to improve the design of future combat systems. With such leadership, HRED can become an outstanding national resource, given the excellent staff and physical facilities that are beginning to be available.