6
Design Challenge: An Integrated PETMAN System

This chapter addresses overall integration of the PETMAN system as described in section 3.2 of the requirements document (Appendix B):


The feasibility study will be based on the PETMAN system performance requirements defined in 3.3.1-3.3.14. The focus of the feasibility study will be on the significant design challenges defined in 3.2.1-3.2.8 of the PETMAN system design in the feasibility study, however all requirements must be considered while conducting the study. The feasibility study will determine the requirements trade-offs and cost differentials associated with each design approach considered in the feasibility study.

PETMAN SYSTEM DESIGN OVERVIEW

PETMAN will be a complex system consisting of many functional subsystems required to perform a variety of behavioral and sensing tasks. A discussion of full system integration must address physical, communication, and control integration. A typical engineering systems approach to this problem will create an overall systems architecture design that identifies the specific submodules and their functions and interrelations. The engineering systems approach is important to modularize the design and evaluation process and to provide a systematic framework for the development process. Principal engineering functions—such as sensing, actuation, modeling, and calibration—are identified, and the analytic design required is focused on those elements. In addition, the engineering systems view provides a basis for the information system architecture (both hardware



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



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 111
6 Design Challenge: An Integrated PETMAN System This chapter addresses overall integration of the PETMAN system as described in section 3.2 of the requirements document (Appendix B): The feasibility study will be based on the PETMAN system performance requirements defined in ..-... The focus of the feasibility study will be on the significant design challenges defined in ..-.. of the PETMAN system design in the feasibility study, however all requirements must be considered while conducting the study. The feasibility study will determine the requirements trade-offs and cost differentials associated with each design approach considered in the feasibility study. PETMAN SySTEM DESIgN OvERvIEW PETMAN will be a complex system consisting of many functional subsystems required to perform a variety of behavioral and sensing tasks. A discussion of full system integration must address physical, communica- tion, and control integration. A typical engineering systems approach to this problem will create an overall systems architecture design that identi- fies the specific submodules and their functions and interrelations. The engineering systems approach is important to modularize the design and evaluation process and to provide a systematic framework for the develop- ment process. Principal engineering functions—such as sensing, actuation, modeling, and calibration—are identified, and the analytic design required is focused on those elements. In addition, the engineering systems view provides a basis for the information system architecture (both hardware 

OCR for page 111
 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT and software). For PETMAN, the three main kinds of integration that must be considered are • Physical integration, which deals with packaging of all the com- ponents deemed appropriate, available, and supportive of desired performance levels. • Controls integration, which addresses the complexity of the multiple subsystems that make up PETMAN and their interdependence. • Communication, which addresses the smooth and seamless ex- changes between the subsystems and external monitoring stations to attain desired performance. The feasibility of full system integration depends on the degree of complexity of PETMAN’s physical subsystems, that is, the number of actua- tors, the number of sensors, the mode of communication between system components, and performance demands with respect to power, endurance, and complexity of physical tasks (exercises). Integration also depends on the features of the preferred design characteristics of the subsystems, such as skin construction and physical characteristics, wired or wireless com- munication, and modular or discrete assemblies. Figure 6-1 is a conceptual diagram of a PETMAN systems design. The system is organized as a hierarchy of subsystems associated with the major functions. Task Planner The task planner provides the overall coordination of the system and designates the main system behaviors that are required and their key char- acteristics, including timing, rate of motion, sensitivity of sensors, and type of protective ensemble. The task planner also supports the principal user interface and access to software systems. Sensor Subsystem The sensor subsystem supports the array of sensors (as described in Chapter 3) that detect chemical signatures in the space between the protec- tive ensemble and the skin. A sensor planning module is responsible for planning and monitoring the execution of the sensor functions. A given application may require a particular subset (by location or type) of sensors, and the sensor planning module selects, calibrates, and confirms the sensor requirements. The sensor planning module also selects the preprocessing and postprocessing required for management of the sensor data acquired. The sensor data themselves are received by a sensor control and process-

OCR for page 111
 DESIGN CHALLENGE: AN INTEGRATED PETMAN SYSTEM FIguRE 6.1 Conceptual design of the integrated PETMAN system. ing module that has direct interfaces with the sensor elements. The sensor control and processing module could be either a centralized or a distrib- uted information system. Data received and preprocessed are stored in the sensor data archive (either centralized or distributed and either onboard or remote). Retrieval of stored data is managed by the sensor planning module. Environmental (under-Ensemble) Subsytem The environmental subsystem determines and controls the conditions in the protective ensemble—understood to be principally air temperature and humidity. The environmental planning module determines these required conditions, including the mode of control that is intended for the subsystem in various portions of the ensemble-skin air space. For example, one mode might require a set skin temperature and a set perspiration (water-trans- port) rate at the surface of the skin, and another mode might estimate the activity of the body and use a metabolic model to control air temperature and humidity in accord with simulated body metabolism; the second mode would require more extensive control systems and more elaborate models to govern heat and water-vapor release. The actuators for the environmen- tal system would probably be embedded at or beneath the skin surface.

OCR for page 111
 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT Motion Subsystem The motion subsystem would provide for the control of all motions of limbs, head, and chest associated with the performance requirements. A critical assumption regarding the performance requirements for PETMAN is that the motion requirements can all be predetermined for each phase of the study. That is, each arm motion, walking motion, or other maneuver is preprogrammed and noninteractive with unknown constraints or events in the local environment. That assumption is critical to the systems design and greatly simplifies the requirements. If all motions under known con- ditions are preprogrammed, many of the kinematic and dynamic control requirements can be precomputed, and execution is greatly simplified. The one exception to the assumption may be related to balance. If true dynamic balance is required (that is, no external stabilizing elements or other mecha- nisms are used), true dynamic control is required to support the system, and both software and hardware requirements (for example, joint sensing and control sample rates) are increased. The motion planning module would interpret motion behaviors (such as walking) set by the task planner and translate them into elemental limb and joint motions (for example, moving the knee) to achieve complex co- ordinate motion. The overall degrees of freedom (DoFs) of the mannequin are not known, but one might expect 16-48 DoFs that must be specified (by the motion planner) and controlled (by the motion controller). Such coordinated control typically requires a comprehensive kinematic model of the system that would constrain the motions in an integrated fashion. The introduction of dynamic balance would require still further modeling of system dynamics. The motion control module would integrate joint and limb sensors with command signals to joint and limb actuators to execute the motion plan. Specific control algorithms for motion and balance could be derived with a variety of analytic, learning, or heuristic approaches. Such approaches have been extensively studied. For dynamic motions (such as jumping), it is less clear that performance could be achieved without more extensive development. Several specific issues arise in relation to the motion subsystem. The use of anthropomorphic hands for general tasks is not well understood, and assumptions about manipulating objects to mimic normal human mo- tions may not be realistic. The choice of hand design and the sensor suite available (for example, tactile and force sensors) will be important in deter- mining the range of manipulation motions and the tasks that are feasible. A second open question may be related to chest-wall motion. This mo- tion might be coupled to transport of air to the respirator, but that is not necessary for the functional design. Chest-wall motion might be achieved

OCR for page 111
 DESIGN CHALLENGE: AN INTEGRATED PETMAN SYSTEM through separate mechanisms and coordinated through control functions with air transport. Taken independently, it is technically feasible to construct almost all the PETMAN subsystems with some degree of compatibility with individual threshold or objective performance requirements. However, as requirements are grouped, power requirements increase, weight is increased, more pack- aging volume is needed, higher control processing power is required, and more complex controls and communication schemes are needed. On the basis of current state of the art of mannequins and humanoid robots, an autonomous nontethered mannequin that performs at the objective level is not feasible; it would entail the sacrifice of some critical performance requirements. That assessment is based on a comparison of the autonomous operation attainable with state-of-the-art technology of portable energy plants and their charging frequency, power rating, and weight requirements. In contrast, the feasibility of approaching more of the performance demands, especially of performing exercises and carrying weapons at de- sired rates and endurance, increases appreciably when power is supplied externally, that is, with a tether. Once a tether is accommodated for power, control and communication can be appreciably simplified because higher power controls can be communicated with externally. Accordingly, we give special attention here to how performance re- quirements may be met to help in the decision of whether to tether or not to tether. TETHERINg COMPATIbILITy WITH FuNCTIONAL REquIREMENTS Table 6.1 provides some insight into the compatibility of tethering with critical functional requirements at the objective and threshold levels. Com- patibility is given a qualitative assessment that indicates whether tethering or nontethering low (L), medium (M), or high (H) compatibility with the desired or required performance. Requirements are listed in the table by paragraph number of the PETMAN proposal (see Appendix B) with key identifying statements. Tethering would be required to supply locomotive power for the man- nequin’s joint motors, to supply fluids necessary for simulating biologic functions, and to communicate control signals and data to and from differ- ent sensors and actuators. The concept of tethering may also be expanded to include external manipulation of the mannequin’s limbs if self-actuation proves too demanding within the constraints of objective or threshold performance levels. The individual performance requirements of PETMAN compete for

OCR for page 111
 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT TAbLE 6.1 Comparative Assessment of Capability of a Tethered and a Nontethered System in Meeting the PETMAN Functional Requirements Design Challenges and Non- Functional Requirements Tethered Tethered 3.2 Feasibility Study Design Challenges 3.2.1 ntegrity of individual protection ensemble I M H equipment 3.2.2 ompatible with individual protection, C M H ancillary equipment, and weapon systems • Donning and doffing M M • Meets 50th percentile males H L 3.2.3 aterials of construction not degraded by M M M chemicals M M • Can be decontaminated 3.2.4 Integrated into ensemble sampling system H M 3.2.5 ontrol of skin temperature and perspiration C H M and respiration rates 3.2.6 ot affected by specified environmental N H M chamber conditions 3.2.7 a- Man-in-Simulant Test (MIST) exercises See 3.3.8 See 3.3.8 - System performing human-like movements b below below 8.8.8 Fully articulated hands and feet that simulate L L human motion 3.3 Performance Requirements 3.3.2 Operation for (T/O): - (12/24) h before operational maintenance a M L - (3/6) months before preventive maintenance b M L c- (6/12) months before calibration M M 3.3.3 Anthropometric requirements; 50th percentile M L male 3.3.4.1 Skin temperature fixed/variable M M 3.3.4.2 Perspiration rate fixed/variable M M 3.3.4.3 Respiration rate fixed/variable M M 3.3.5 ompatible with chemical-breakthrough C H M sampling technologies 3.3.6 Not affected by chamber environmental M L conditions: 3.3.6.1, 3.3.6.2, 3.3.6.3 (for 0-10 mph), 3.3.6.4, 3.3.6.5 L L 3.3.6.3 Not affected by wind speed up to 161 mph 3.3.7 Articulated and robotic, like a human in H L aesthetics and proportions: 3.3.7.1-9, -11, -12, -13, -14, -15, -16: arms, legs, H M torso L L 3.3.7.10 hands

OCR for page 111
 DESIGN CHALLENGE: AN INTEGRATED PETMAN SYSTEM TAbLE 6.1 Continued Design Challenges and Non- Functional Requirements Tethered Tethered 3.3.8 Simulation of exercises: 3.3.8.1 Standing M M 3.3.8.2 Walking at 4.8 km/h (3 mph) H M 3.3.8.3 Marching (12-in.-high step) at 4.8 km/h H M 3.3.8.4 Jumping jacks—jumping, landing M L 3.3.8.5 Sitting H M 3.3.8.6 Standing to squatting to standing position M L 3.3.8.7 Reaching arms in all directions M M 3.3.8.8 From standing to lying prone and return M L 3.3.8.9 Kneeling on one knee M M 3.3.8.10 Kneeling on both knees M M 3.3.8.11 Low crawl M L 3.3.8.12 High crawl M L 3.3.8.13 Aiming weapon in various positions M L 3.3.9 Compatible with protection and ancillary equipment: M M 3.3.9.1 Suits M L 3.3.9.2 Boots M M 3.3.9.3 Gloves M M 3.3.9.4 Masks M M 3.3.9.5 Helmets M L 3.3.9.6 Combat boots M M 3.3.9.7 Ballistic-protection vests M M 3.3.9.8 Pistol holsters M M 3.3.9.9 Battle dress and combat uniform M M 3.3.9.10 Physical training gear, T-shirt, shorts, and M M so on 3.3.9.11 Skin-exposure reduction paste 3.3.10 Compatible with weapons systems: 3.3.10.1 M4 modular weapon H M 3.3.10.2 M24 sniper rifle H M 3.3.10.3 M16A2 rifle, 5.56 mm H M 3.3.10.4 XM8 lightweight assault rifle H M 3.3.11 Capable of being decontaminated M M 3.3.12 Uses common commercially available parts H M 3.3.13 Records system measures at 1-s intervals: Skin temperature M M Respiration rate M M Perspiration rate M M Penetrating mass (in nanograms) of chemical M M vapor 3.3.14 Programmable to perform exercises or H M motion NOTE: H = high capability, M = adequate or medium capability, L = little or no capability.

OCR for page 111
 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT power resources and controllability and are often in conflict. For example, extensive articulation is required for human-like aesthetics, but when cou- pled with required exercises with real weapons the weight and size would probably become too great to meet the 50th percentile of male subjects. Similarly, increased weight requires higher power, which conflicts with long performance times in nontethered systems. Accordingly, the assessment of compatibility depends on the priorities associated with each of the PETMAN performance requirements. Deter- mining such priorities would move the PETMAN system into the realm of meeting more objective level requirements. The assessment is highly subjective, but it is indicative of the poten- tial opportunity for meeting the performance requirements when either approach is adopted. Variations in judgments are possible, although the overall assessment is likely to be the same. The overall systems design of a PETMAN is strongly influenced by the choice of a nontethered or tethered operation. With the nontethered ap- proach, all resources—including those required for power, gases, liquids, and heating and cooling functions—would need to be housed and managed inside the mannequin itself. The principal advantage of nontethered opera- tion is that the ensemble remains entirely intact in accord with the goal of using standard ensemble equipment without modification as it would be worn by a human. Any tether access to the mannequin would need to penetrate the ensemble at some point or require customization of some part of it to enable access. As described below, a tether connected to the head, torso, or boot heel of the mannequin may constitute an acceptable compromise. A review of available technologies suggests that fully self-contained nontethered operation may be difficult to achieve in the short term because of the space and weight requirements associated with onboard power and materials. Current battery technologies will realistically support electric motor drives only for self locomotion of the mannequin with simple move- ments. Similarly, hydraulic systems pose major challenges to the nontethered design in that volume and power for an onboard hydraulic pump, storage tank, accumulator, and necessary battery power source are unrealistic with current technologies. On the basis of those assessments, it is important to consider the principal options for tethered operations and to identify their effects and risks. The key favorable and unfavorable aspects of the use of a tether include the following: Favorable: • Provides access for ample power, gases, liquids, and communication. • Dramatically reduces the space and weight requirements for inter-

OCR for page 111
 DESIGN CHALLENGE: AN INTEGRATED PETMAN SYSTEM nal components of the mannequin, making the 50th percentile male envelope more realizable. • Might make it possible to operate for the targeted operational periods and maintenance frequencies. • Allows most of the functional requirements to be met. Unfavorable: • Requires some modification of the ensemble to enable penetration by a tether that contains cables, tubes, and fibers. • May disturb the natural mobility, dynamics, and range of motion for the mannequin because of the forces and weight of the tether itself. • May require specially actuated support capability to manipulate the tether as the mannequin moves. On the other hand, the key favorable and unfavorable aspects of the nontethered approach include the following: Favorable: • Maintains the integrity of the ensemble for testing without penetration. • Provides good aesthetics. Unfavorable: • Has inadequate power to meet many of the motion and actuation requirements for full mannequin functionality. • If required power is provided internally, weight and size are likely to far exceed those of the 50th percentile male. • Is not expected to meet targeted operational periods and frequency of maintenance if the 50th percentile male envelope is observed. • Posture stability is not yet controllable for most of the exercise movements. Tether Attachment Options Three broad options for tether attachment to the mannequin are dis- cussed below.

OCR for page 111
0 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT Head a. Back of head. Tether connection to the back of the head would require modification of the helmet and all hood components of the ensemble. b. Face. Integration of the tether connection into the face mask may be easier to accommodate as an equipment modification, but there may be disruption of normal sealing of the mask because of forces applied by the tether. c. Both. Back and face tethers could interfere seriously with head and neck motion. Without active suspension of the cable, a balancing man- nequin would be affected by such forces applied to the head. Abnormal motions might affect penetration in the head and neck regions. Torso For the back or front of the torso, the principal areas of every test en- semble would need to be modified to introduce a tether interface. Teth- ers attached to the torso could also interfere substantially with motion of the mannequin because of applied forces, and active manipulation of the tether might be needed. Foot or Boot Tether attachment to the sole or heel of the boot may have some ad- vantages, especially that any penetration may be limited to one shoe or boot of a pair, allowing full testing opportunity for the other shoe or boot of the same pair. • The modification of a boot may be more acceptable than modifi- cation of other components of the ensemble, particularly because the sole and heel of the boot may be less susceptible to the leakage under study. In addition, it may be more acceptable to use a small number of custom boot models that are customized with tether connectors, and these choices may have minimal effects on other aspects of the tests. • Some forms of attachment of the tether to the sole of a boot could directly interfere with foot placement for all standing and walking motions. Such attachments may lead to an alternative approach to generate motion externally; for example, each foot would be moved by an external manipulation mechanism without actual ground contact in achieving the motion. Such an approach would lead to a quite different overall systems design (external-internal

OCR for page 111
 DESIGN CHALLENGE: AN INTEGRATED PETMAN SYSTEM drive), considered below, but might accommodate the boot-based tether connection. • The attachment of the tether to the back of the heel of the boot may be acceptable if a small number of modified boots could be used in a variety of tests. Attachment to the heel would still permit natural motion of the foot and boot and would not interfere with standing, walking, jumping, or crawling motions. External ma- nipulation of the tether to avoid applied forces to the foot and leg may be necessary. External-Internal Drive of the Mannequin The self-contained (nontethered) mannequin design would require all actuators to be inside the body, and this poses a challenging design prob- lem to fit actuators into a body of the required size and the problems of managing power and cooling to support the actuators. An alternative strategy could configure external manipulation of the mannequin limbs with external mechanisms. A manipulator with four to six DOFs might be attached to each foot and to each hand. The external drives would provide the primary forces used to move the feet and hands and might, in principle, provide much of the balance and dynamic support needed for the set of prescribed motions. The mannequin would still require some internal drive mechanisms, but the power, DoFs, and range of motion for the powered actuators would be reduced. In addition, passive joints could be introduced into the mannequin to support the needed DoFs. In principle, the resulting system could be used to generate the required motions. There would be a loss of realism in the lack of actual boot contact with the ground. The execution of walking, marching, jumping, and crawling motions could still be accomplished to mimic the required motions. This approach may have merit for nontethered mannequins for the performance of tasks and exercises that require higher loads and speeds than may be possible with internally packaged actuators, such as crawling with weapons. SOME INTEgRATION ObSERvATIONS The design of the PETMAN system—mannequin, controls, software, subsystems, and the rest—is challenging. However, all the individual sub- systems appear manageable independently, and the challenge is in the inte- gration within the constraints of size, weight, and functional requirements. Accordingly, a systematic approach to the design that focuses on attaining the most functionality at the early stages and prepares the design to accom- modate remaining functionality at later stages is desirable.

OCR for page 111
 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT The nontethered mannequin could be considered as the ultimate ob- jective, but it should not now be the overriding one. It appears advanta- geous to begin with a tethered approach and a target of meeting threshold performance levels. The design process would determine the configuration and characteristics of the subsystems and system components, and it should be unencumbered by power requirements for locomotion or control. Once the threshold levels are achieved, higher-priority objectives could be in- corporated progressively to the extent that volume and weight restrictions allow. A follow-up phase could deal with progressive enhancements as technology is developed and experience leads to more ingenious designs and inventions. There is a high level of uncertainty in the feasibility of attaining many of the PETMAN requirements. That uncertainty makes it impossible to predict the cost and time requirements of a PETMAN-like mannequin solicitation. A phased approach may be useful, with the initial phase being a development phase that addresses the high-priority requirements. That phase should result in a realistic set of specifications that address those re- quirements, and focus on objectives that can be attained within the desired timeframe. A minimum of one year may be necessary for the development phase. The follow-up phase may then focus on the realization of a mannequin with acceptable capabilities. The following are areas to be addressed in a development phase: • Respiration volume and space requirements. • Actuation of hands and feet within the volume available in the arms and legs of the 50th percentile torso. • Stability and control of the high-activity exercises. • Decontamination. • Level of functionality without a tether. • Heating and cooling. • Maintainablity and reliability. PRACTICAL APPROACH TO MANNEquIN DEvELOPMENT A practical approach to developing the PETMAN system may follow these steps: 1. Design the mannequin so that all actuators and controls are ex- ternally powered and a tether penetrates at the least disturbing location, possibly one of the boots. 2. Operate the electric motors at powers higher than their continuous- operation ratings and improve their duty cycle by cooling supplied

OCR for page 111
 DESIGN CHALLENGE: AN INTEGRATED PETMAN SYSTEM through the tether. This approach reduces the actuators to their smallest practical size. 3. Consider using the highest-speed motors with higher-ratio speed reducers, harmonic drives, and other cycloidal reducers and allow- ing compact packaging. 4. Assess whether the actuators and other subsystem components can be packaged within the target volume and size constraints of the 50th percentile male population. If not, apply acceptable compro- mises and trade-offs to reduce the number of actuators or their power or size requirements to acceptable and packageable levels; some low-priority functionality could be sacrificed, beginning with the ones that are most demanding of power and space. 5. Reevaluate the possibilities of improving functionality and regain- ing compromise losses by replacing some electric actuators with air actuators, especially those with small stroke-angle movements; consider trade-offs to use oversized motors rather than motor- reducer combinations. 6. Reassess the tethering of controls and communication versus self- contained approaches and wireless communication to reduce the size, rigidity, and weight of the tether and help with its maneuver- ability and management. 7. Allow for efficient use and distribution of cooling air to reduce its supply-line size; for example, if not all actuators are cooled simultaneously, cooling air may be cascaded from one function to another for cooling particular areas or heating others as needed to maintain body-surface areas at desired temperatures; consider circulating heated air with blowers and fans. 8. A whole-system energy distribution study may be performed to help to distribute the cooling and heating media most efficiently. Ideally, if the net balance is an energy surplus, only cooling is re- quired, although in reality some heating will be necessary in some locations and more cooling will be added; such inefficiency should be minimized. SySTEMS ARCHITECTuRE AND SOFTWARE CONSIDERATIONS The architecture for all the systems reviewed takes the form of a hier- archic structure that integrates task definition and planning, model-based execution and representation, and behavior-based sensor reflex control. This type of architecture is generally well understood, the implementation is feasible, and such a modular approach is important for the reliable and systematic development of PETMAN. The dancing humanoid by Ikeuchi’s

OCR for page 111
 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT group is an example of behavior control similar to what is needed for the PETMAN project (see www.cvl.iis.u-tokyo.ac.jp). Software systems for PETMAN must support several classes of users to meet the stated needs fully. These systems include • Operations testing and monitoring. This would be a user-oriented system interface designed for nontechnical users and prepared with clear interactions for routine setup and running of tests, collection and preview of data, and periodic general monitoring as required. • Test development and routing maintenance. This system would support periodic revision and development of new test procedures within the basic framework of the PETMAN system. It would in- tegrate simulation and modeling capabilities so that a user could create new exercises and behaviors and visualize them in the system design mode. No formal programming language interaction should be required. • Expert programming. This system would provide expert access to the modular computer software codes that underlie the hierarchic system and would be used by experts, perhaps vendor staff, trained to work with it at this level. Existing software tools and systems are available to create all those components. In addition, it is recommended that state-of-the-art software engineering methods and procedures be instituted to support the creation of reliable and maintainable software. The software structure of PETMAN will be critical for both the func- tionality and the usability of the system. As suggested in the system archi- tecture overview discussed earlier, the system will most likely be hierarchic in structure, and the software should be well organized, modular, and clearly structured to implement key capabilities for planning, modeling, control, sensing, and actuation. In practice, it may be appropriate to use different programming environments and languages for the different lay- ers of software. Although the planning and task-development environment might use an abstract representation to define sequences of mannequin behaviors, the low-level embedded software that drives motors and sensors will involve a quite different style of programming and implementation. From a broad systems perspective, software development itself is feasible with existing software methods and programming environments. However, demonstrated experience in development at all three levels of capability should be examined carefully. Clear benchmarks for development will be critical and require adequate effort and resources. A key approach to programming takes into account the variety of users of the system. At least three groups of users must have effective and efficient

OCR for page 111
 DESIGN CHALLENGE: AN INTEGRATED PETMAN SYSTEM access to appropriate software tools in the final implemented systems, as described below. Real-Time Test Implementation and Monitoring This level of software development should focus on user interaction and capabilities for users to implement and monitor tests effectively. The users are not software programmers or experts in the technology of the system. The software will provide an interface based on pictorial interac- tion and clear display of data and system status. Many interactive tools support these types of interaction, and the implementer might be expected to adopt existing software environments that would support such tools. An important component of the test-operator environment will be the control and organization of data collection. Data should be systematically collected and archived with integrated time references. Preliminary real-time display of data for all sensors in the system should be part of this environment. Test-Protocol Development and Routine Maintenance An onsite technical systems staff will be required to support PETMAN and to provide the capability to modify, design, and implement new proce- dures and tests. This programming environment should provide access to high-level definitions of mannequin behavior and motion and opportunities for users to change procedures, timing, and motions within a bounded set of options. The programming environment should have a strong visualization component. For the modification of behaviors and motions or the design of new tests, the visualization environment should provide animated views of candidate motions and evaluation measures related to forces, energy expenditure, temperature, and other factors. Simulation is an important element of this development process. In the review of background systems (see Appendix D), the simulation systems presented by American Android Corp. and Boston Dynamics were good examples of the types of simulation that might be embodied in the developers’ programming environment. Expert-System update, Revision, and Debugging The detailed internal programming code of the system should normally not be accessed in routine maintenance and operating procedures. However, there must be internal access by the developer or other designated support vendors to modify and update internal software code. Proper software engineering procedures and methods are critical and support any needed modifications at this level of detail. Independent testing and benchmarks

OCR for page 111
 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT of modular components of the hierarchic system should support systematic updating as needed. CONCLuSIONS AND RECOMMENDATIONS Conclusion 6-1: Taken independently, most of the PETMAN objectives are feasible. Collectively, they are likely to be in conflict and will require compromises. Recommendation 6-1: • Set priorities among the requirements to help to meet the objectives that have the highest payoff. • Maintain close contact with the contractors, and jointly decide on design trade-offs. Conclusion 6-2: The nontethered approach to mannequin design has merit. However, a nontethered design that meets even the objective level of the functional requirements of PETMAN would be difficult to achieve with state-of-the art technology. Recommendation 6-2: Make the nontethered approach the ultimate target for mannequin design, but allow tethering for the initial stages. Phase the program to have a functional mannequin from the outset, and allow progressive enhancements that gradually minimize the tether until it is eliminated from the design. Conclusion 6-3: A tethered mannequin design has the best potential to meet critical functional requirements. The tether may require special man- agement to minimize its effects on mannequin movements. Recommendation 6-3: • Begin the design with a tethered approach that minimizes the size and bulk of the tether. • Consider the tether attachment at the heel of a boot to minimize tether interference with the mannequin. • From the onset of the design, observe tether management closely to minimize use of the tether. Conclusion 6-4: A tethered approach burdens the design with the po- tential need for a tether-managing arrangement that follows the manne-

OCR for page 111
 DESIGN CHALLENGE: AN INTEGRATED PETMAN SYSTEM quin and distracts from its aesthetics and agility. Any design that reduces the bulk of the tether would help to minimize its disadvantages. Recommendation 6-4: • Package as much of the motion and control resources inside the mannequin as possible within the target 50th percentile male popu- lation envelope. • Minimize the size of the tether by use of wireless communication and shared or complementary inputs and outputs; for example, cooling water from motors may used to heat the mannequin skin. • Use dynamic braking of motors (which recharges an internal bat- tery) to minimize the external energy supply and the size of the power cord in the tether. Conclusion 6-5: The dynamic stability of marching in high step, jump- ing-jack exercises, and the low and high crawl with weapons is chal- lenging and goes beyond the state of the art of humanoid technology. The PETMAN system could be powered and articulated to perform the exercises, but the controls may take more time to develop Recommendation 6-5: Follow a phased approach to the development of PETMAN to attain more functionality as time goes by. Include all the power requirements in the initial phase, and add the controls as the technology and algorithms develop. Conclusion 6-6: The system architecture takes the form of a hierarchic structure that integrates task definition and planning, and model- and behavior-based execution and representation. Recommendation 6-6: • Require a systematic modular approach to the design of both hard- ware and software systems. • Require explicit linkage of system modules to maintenance, service, and test procedures for system monitoring and repair. Conclusion 6-7: Software systems for PETMAN must support several classes of users to meet stated needs fully, including • Operations testing and monitoring. • Test development and routing maintenance. • Expert programming.

OCR for page 111
 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT Recommendation 6-7: • A formal approach to the introduction of modern software en- gineering principles and practices should be an integral part of software development. • Existing commercial software systems and tools seem to be avail- able to support the functionality anticipated in this system and should be used to support maintenance, reliability, and upgrading of the resulting system.