Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
4 Design Challenge: Robotic Capability for PETMAN This chapter describes the specific robotic capability needed for the PETMAN system, including motion and articulation, power, and heating and cooling, according to the following PETMAN design challenges: 3.2.1 The study will determine the feasibility of designing a PETMAN system to be tethered (T) or free standing/self-contained (O). A tethered PETMAN system design must not compromise the integrity of the individ- ual protection ensemble equipment being tested on the PETMAN system. If a tethered design is selected the design must also minimize the impact to the whole ensemble operation. 3.2.2 The study will determine the feasibility of designing a PETMAN sys- tem to be compatible with all individual protection and ancillary equipment as well as weapon systems defined in 3.3.9-18.104.22.168. Areas to be addressed are donning/doffing and proper size/fit of the individual protection equip- ment. The PETMAN system design shall meet the appropriate 50th percen- tile male anthropometric measurements, as defined in DOD-HDBK-743A, Military Handbook Anthropometry of U.S. Military Personnel, to allow for the necessary fit/seal that each piece of protective equipment requires. 3.2.5 The study will determine the feasibility of designing a PETMAN sys- tem that can simulate fixed skin temperature (by body region), perspiration rate (by body region), and respiration rate (T) and a realistic variability in skin temperature, perspiration rates, and respiration rates based on the amount of physical activity/exertion (O) defined in 22.214.171.124-126.96.36.199. 77
78 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT 3.2.7 The study will determine the feasibility of designing a PETMAN sys- tem that can perform the Man-in-Simulant Test (MIST) exercises defined in 3.3.8 versus all motions. The study will determine the feasibility of the PETMAN system performing the human-like movements utilizing the de- grees of freedom (DoF) and considerations defined in 3.3.7. 3.2.8 The study will determine the feasibility of designing a PETMAN sys- tem with fully articulated hands and feet that simulate human motion, the minimum amount of hand and foot articulation required for the PETMAN system operation and a partial level of hand and foot articulation. Relevant PETMAN Requirements Additional relevant PETMAN requirements are presented in Box 4.1. The next section will describe current technology and the feasibility of meeting the PETMAN system requirements. CURRENT Technology Humanoid robots are those whose body structure resembles that of a human. The first humanoid robot was perhaps the WABOT-1 (Figure 4.1), developed by Katoâs group at Waseda University in 1970-1973. Several humanoid robots have since been developed, usually with a specific purpose in mind, such as to study bipedal locomotion, for enter- tainment, and to assist humans. In 2000, research and development in this field reached a higher level of sophistication when Honda introduced ASIMO. The current version of ASIMO, released in December 2005, is a 34-degrees-of-freedom (34-DoF) robot. It can run, walk smoothly, climb stairs, communicate, and recognize peopleâs voices. ASIMO is 130 cm tall, weighs 54 kg, and can operate for 1 hour on an internal battery. A more recent robot is HRP-2, which was also developed, by Kawada Industries, to assist humans. Videos of this model walking, wearing a suit, lying down and getting up, and performing such complex behaviors as dancing are readily available on the Web (for example, see http://www. plyojump.com/hrp.html). This robot is 154 cm tall, weighs 58 kg, has 30 DoFs, and can operate for about an hour on an internal NiMH battery. The current model, HRP-3P, has the same capabilities, six more DoFs, and â ASIMO: The Honda Humanoid Robot. http://world.honda.com/ASIMO/. Accessed June 14, 2007. â Humanoid Robot HRP-2 âPromet.â Kawada Industries, Inc. http://www.kawada.co.jp/ global/ams/hrp_2.html. Accessed June 14, 2007.
DESIGN CHALLENGE: ROBOTIC CAPABILITY FOR PETMAN 79 Box 4.1 Relevant PETMAN Requirements 3.3.2 The PETMAN system shall be capable of operation for twelve (12) hours prior to requiring operational maintenance, three (3) months prior to preventive maintenance and six (6) months prior to calibration. (T) The PETMAN system shall be capable of operation for twenty-four (24) hours prior to requiring operational maintenance, six (6) months prior to preventive maintenance and twelve (12) months prior to calibration. (O) Operational maintenance is defined as the required maintenance procedures to prepare the PETMAN system for each test trial, for example, filling a perspiration reservoir, changing agent samplers or decontami- nation before the next trial. Preventive maintenance is defined as a maintenance event performed prior to a failure in order to prevent its occurrence. 3.3.10 The PETMAN system shall be compatible with the following weapons sys- tems. The PETMAN system must be able to hold/grip and aim the weapon IAW the field manual (FM) FM 3-22.9, Rifle Marksmanship Field Manual. 188.8.131.52 M4 Modular Weapon 184.108.40.206 M24 Sniper Rifle 220.127.116.11 M16A2 Rifle 5.56 MM 18.104.22.168 XM8 Lightweight Assault Rifle 3.3.13 The PETMAN system shall record the following system parameters over time: skin temperature, respiration rate, perspiration rate, and total mass (in nanograms) of chemical vapor that penetrates/permeates through the protective ensemble. The PETMAN system shall record the start and stop time of each mo- tion in 1-second increments. the ability to adjust to slippery surfaces when walking. HRP-3 is 160 cm tall and weighs 65 kg. Hubo, a robot developed at the Korea Advanced Institute of Science and Technology (KAIST), is similar in size (56 kg and 125 cm), shape, and function to ASIMO. It is a 41-DoF robot and can operate for about 1 hour on a Li-polymer battery. All the above examples use electric motors for actuation. Humanoid robots developed by Sarcos use hydraulic actuation. Some of the differences between the two methods are described later in this chapter. Robots such as â This site is in Japanese: HRP-3P. Kawada Industries, Inc. http://www.kawada.co.jp/mechs/ hrp-3p/index.html, http://pc.watch.impress.co.jp/docs/2005/0909/hrp3.htm. â Humanoid Robot Research Center. Hubo Lab. http://www.hubolab.com/. Accessed June 14, 2007.
80 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT Figure 4.1 WABOT-1, developed at Waseda University in the early 1970s, was the first humanoid robot. SOURCE: WABOT-1, Humanoid Robotics Institute, Waseda University.
DESIGN CHALLENGE: ROBOTIC CAPABILITY FOR PETMAN 81 the Sarcoman and the Sarcos Animation Figure are mounted so that they do not have to balance. Numerous videos on the Internet demonstrate the ability of Sarcoman to perform a variety of tasks that require dexterity. These types of robots have been used to test spacesuit designs, in the enter- tainment industry, and in robotics development and research laboratories. Dexter, a system designed by Anybots, and Lucy, a system still under development at Vrije Universiteit Brussel, both operate with pneumatic actuation. This type of actuation mechanism is still being developed. Dexter is 175 cm tall, weighs 61 kg, and does not have arms. Lucy is only a pair of 150-cm-tall, 30-kg legs. The goal of these robots is to study the use of pneumatic actuation in bipedal walking robots. Robonaut, a teleoperated robot developed by NASA, was designed to work with astronauts in zero-gravity environments. Inasmuch as astro- nautsâ legs generally work as one unit in space, developing a system that could walk was not necessary. Instead, dexterity in the upper body was con- sidered of paramount importance. The most recent version of Robonaut, Robonaut R1b, is essentially an upper body that can be mounted either on one âlegâ, on a two-wheel base, or on an all-terrain vehicle appropriate for extreme environments. Robonaut is intended to work cooperatively with humans, be controlled by humans, and use tools designed for humans, so its human-size hands are complex, with five fingers, 12 DoFs in the hand itself, and two DoFs in the wrist. The development of the hand was closely tied with the development of the arm and would not be easily transferable to a different system. The state-of-the-art humanoid technology is impressive. It can read- ily perform many simulated exercises. However, none of the humanoid robots reviewed above is suited for PETMAN applications, although they constitute a technologic basis for what is or is not feasible. The major limitations of the current humanoid robots for PETMAN applications are of four types: â¢ Proportion. Most humanoid robots look like a man in a space suit. They have been adjusted to human proportions in the verti- cal direction (such as height, foot length, and arm length), but their legs and arms are too fat to accommodate joint actuation mechanisms. âTele-robotics: High Performance Humanoid Robot. Sarcos. http://www.sarcos.com/telespec. atr.html. Accessed June 14, 2007. ââAbout the Robots.â Anybots. http://www.anybots.com. Accessed June 14, 2007. âBipedal Robot Walking Lucy. Department of Mechanical Engineering, Vrije Universiteit Brussel. http://lucy.vub.ac.be/index.htm. Accessed June 14, 2007. âDecember Tests at JSC. Robonaut. NASA Johnson Space Center. http://robonaut.jsc.nasa. gov/. Accessed June 14, 2007.
82 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT â¢ Operating time. Electric humanoids look attractive for PETMAN applications because they are battery-operated and obviate a power cable, but the continuous operating time is limited to about an hour. â¢ Robustness. Current humanoid robots are not designed for robust- ness. Data are not available to estimate the reliability of current technology for PETMAN applications. â¢ Primitive hands. Most humanoid robots have too few degrees of freedom for hands. Attempts to introduce additional degrees of freedom will make the arms heavier and possibly fatter. Movement and Degrees of Freedom Current humanoid robots typically have lower maximal velocities than humans. For example, most of the joints of H7, a bipedal robot developed at the University of Tokyo, are limited to about 1 revolution/s. In addition, the range of motion of most joints is less than a humanâs. Tables 4.1 and 4.2 give ranges of motion of HRP-2 and Hubo. Note that the human ankle has a yaw degree of freedom not listed in these tables. In light of the available technology, the PETMAN requirements de- scribing degrees of freedom are too vague and do not highlight the specific needs of the project. We recommend that the sponsor add specific range-of- motion and maximal-velocity requirements for the desired behaviors to any broad agency announcement (BAA). Motion-capture studies of movements and donning and doffing of soldiers performing MIST exercises should be used to refine BAA requirements for range of motion and maximal joint velocities and to estimate forces and torques necessary for the desired man- nequin behaviors. Motion-capture studies are often carried out in this way to understand complex movement such as in medical or sports science.10 Actuation and Transmission Technology The technology used for actuation of the PETMAN system will be a key feature of any development plan. The type of actuation in a robot affects range of motion, sizes of limbs and joints, power use, heat generation and â See JSK website: http://www.jsk.t.u-tokyo.ac.jp/research/h6/H6_H7.html. Accessed Octo- ber 9, 2007. 10â Brady, R. 2000. Foot Displacement but Not Velocity Predicts the Outcome of a Slip Induced in Young Subjects While Walking. Journal of Biomechanics 33(7):803-808; Brown, W. M., L. Cronk, K. Grochow, A. Jacobson, C. K. Liu, Z. PopoviÄ, andÂ R. Trivers. 2005. Dance Reveals Symmetry Especially in Young Men. Nature 438:1148-1150; Davis, I. 2006. Biomechanical Factors Associated with Tibial Stress Fracture in Female Runners. Medicine and Science in Sports and Exercise 38(2):323-328.
DESIGN CHALLENGE: ROBOTIC CAPABILITY FOR PETMAN 83 Table 4.1 Ranges of Motion of the HRP-2 Robot (a) Standard (b) HRP-2 Joint Human Head R -50 deg. to 50 no existence. deg. P -50 deg. to 60 -30 deg. to 45 deg. deg. Y -70 deg. to 70 -45 deg. to 45 deg. deg. Right Arm Shoulder R -90 deg. to 0 -95 deg. to 10 deg. deg. P -180 deg. to 50 -180 deg. to 60 deg. deg. Y -90 deg. to 90 -90 deg. to 90 deg. deg. Elbow P -145 deg. to 0 -135 deg. to 0 deg. deg. Y -90 deg. to 90 -90 deg. to 90 deg. deg. Wrist R -55 deg. to 25 no existence. deg. P -70 deg. to 90 -90 deg. to 90 deg. deg. Right Hand P 0 deg. to 90 -16 deg. to 60 deg. deg. Waist R -50 deg. to 50 no existence. deg. P -30 deg. to 45 -5 deg. to 60 deg. deg. Y -40 deg. to 40 -45 deg. to 45 deg. deg. Right Leg Hip R -45 deg. to 20 -35 deg. to 20 deg. deg. P -125 deg. to 15 -125 deg. to 42 deg. deg. Y -45 deg. to 45 -45 deg. to 30 deg. deg. Knee P 0 deg. to 130 0 deg. to 150 deg. deg. Ankle R -20 deg. to 30 -20 deg. to 35 deg. deg. P -20 deg. to 45 -75 deg. to 42 deg. deg. NOTE: R: Roll axis, P: Pitch axis, Y: Yaw axis SOURCE: Kaneko, K., F. Kanehiro, S. Kajita, H. Hirukawa,T. Kawasaki, M. Hirata, K. Akachi, and T. Isozumi. Humanoid Robot HRP-2. Proceedings of the 2004 IEEE International Conference on Robotics & Automation, New Orleans, LA. April 2004. Â© 2004 IEEE.
84 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT Table 4.2 Ranges of Motion of the Hubo Robot Angle Range Hip Yaw ââ 0 ~ +45Â° Roll â31Â° ~ +28Â° Pitch â90Â° ~ +90Â° Knee Pitch â10Â° ~ +150Â° Ankle Pitch â90Â° ~ +90Â° Roll â23Â° ~ +23Â° SOURCE: Presentation by Jun-Ho Oh, April 2, 2007. distribution, deliverable power and control, ease of joint movement when the system has no power (backdriveability), and more. In practical terms for the PETMAN system, the method of actuation will affect the ability of the robot to perform exercises with the desired forces and range of motion, the ease of maintaining a stable temperature, the likelihood of achieving an- thropomorphic dimensions, the ease of donning and doffing the individual protection ensemble (IPE), and the need for a tether. Several actuation technologies, including electric, hydraulic, and pneu- matic systems were considered, but there is no clear leading technology. Most commercial humanoid robots use electric torque motors and har- monic drives, but there are many other possibilities, including: â¢ Electric systems with battery power â¢ Electric systems with a tether to deliver power â¢ Hydraulic systems with an onboard battery-powered hydraulic pump â¢ Hydraulic systems with an onboard hydraulic pump and an exter- nal power source (electric tether) â¢ Hydraulic systems with an external hydraulic pump (pressurized hydraulic fluid in the tether) â¢ Pneumatic systems with an onboard battery-powered air compressor â¢ Pneumatic systems with an onboard air compressor and an external power source (electric tether) â¢ Pneumatic systems with an external air compressor (pressurized air in the tether) Advantages and disadvantages of various systems are described in Table 4.3. All-electric systems exist, but the overall perspective that emerged in our discussion is that hydraulic and pneumatic systems may also be feasible.
DESIGN CHALLENGE: ROBOTIC CAPABILITY FOR PETMAN 85 Table 4.3 Advantages and Disadvantages of Electric, Hydraulic, and Pneumatic Actuation Type of Advantages Disadvantages Actuation Electric - Precisely controlled, high- control - Inappropriate form factor bandwidth for some joints (ankle, wrist, - Can be sealed and cleaned. fingers). - Rotary actuation (compact design - Complicated design for linear is possible for some joints) actuation - Motor-reduction gear set-harmonic - Requires heat dissipation drive can produce large torques - Extensive gearing is typical - Joint torque feedback can be used with poor backdriveability and to achieve backdriveability and little torque control good torque control Hydraulic - Can produce large forces - Requires hydraulic fluid lines, - Linear actuation can be with space requirements appropriate for some joints - Needs to handle hydraulic- - Precisely controlled high- control fluid leakage bandwidth - Requirse hydraulic pump -Joint torque feedback can be used - Inappropriate form factor for to achieve backdriveability and some joints; complicated design good torque control for rotary actuation - Pump may need heat dissipation - Difficult to achieve human range of motion with rigid linear actuators Pneumatic - Compact - Pneumatic system must be - Lightweight closed; air cannot be exhausted - Inherently backdriveable into ensemble - Difficult to control because of air compressibility and transmission delay; low-control bandwidth - Requires air lines and air compressor with associated space requirements - Range of motion may be low - Bulky size or high air pressure needed for large force - Inappropriate form factor for some joints; complicated design for rotary actuation
86 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT To avoid the use of a tether, an onboard hydraulic pump or air compressor and associated power source may need to be developed. Ability to Don and Doff without Modification of PPE Although the robot does not need to don and doff the protective gear autonomously, it does need to be possible for a human to put the protec- tive gear on the robot and take it off. That may require that the robot be backdriveable or that it be able to move in such a way as to assist in don- ning and doffing. Power and Energy The threshold power requirements for the PETMAN system call for a tethered system. With a power line attached to the robot providing un- limited, uninterrupted power, it would be feasible to meet either the 12-h (T) or 24-h (O) operational maintenance requirement. Without a tether (O), however, feasibility is more difficult to assess. Examples of other ro- botic systems that perform legged locomotion suggest power requirements of around 300 W at the battery terminals during moderate walking.11 ASIMO uses a NiMH 38.4-V 10-Ah battery; HRP-2 uses a NiMH 48-V 14.8-Ah battery; and Hubo uses a 24-V, 20-AhLi-polymer battery.12 Use of Li-polymer batteries at 150 W-h per kilogram would require batteries weighing 2 kg for an hour of operation. New models of ASIMO run for 1 h between charges.13 Hubo runs for about 60 min between charges.14 Considering only the power requirements for motion and actuation (not including simulation of human physiology, sensors, and so on) and a test schedule that has the robot just standing at least 50 percent of the time, a two hour untethered test seems feasible with current technology before charging is required. Although plugging the mannequin into the wall (or plugging a cable into the mannequin) is the most obvious method of recharging the system, other possibilities could be considered: 11â Based on Dr. Saoshi Kagamiâs humanoid, which consumes 300-400 watts (100 watts to power the CPU), weighs 58 kg, is 1.58 m tall, and walks at 2.1 km/hr. See Bekey, G., Ambrose, R., V. Kumar, A. Sanderson, B. Wilcox, and Y. Zheng. 2006. WTEC Panel Report on Inter- national Assessment of Research and Develooment in Robotics, World Technology Evaluation Center, Baltimore, Md., page 87. 12â Jun Ohâs talk to committee. 13â Honda ASIMO. http://asimo.honda.com/InsideASIMO.aspx. Accessed October 9, 2007. 14â Humanoid Robot Research Center. Hubo Labs. http://www.hubolab.com. Accessed June 14, 2007.
DESIGN CHALLENGE: ROBOTIC CAPABILITY FOR PETMAN 87 â¢ It is feasible to recharge a battery through electric contacts on the boot or on the face plate to allow longer tests. â¢ It is feasible to charge the robot inductively to allow longer tests. Any method of recharging should be evaluated in light of any neces- sary IPE modifications and the length of time needed to recharge the system fully. Heating and Cooling Current humanoid robots may produce 100-200 W of heat more than a human being. In the IPE, removing that heat will be difficult. For the mannequin to maintain âskinâ temperatures similar to those of a human, the system will most likely require heating and cooling at different points in the operation. In addition, motors, actuators, and other components may create hotspots in the robot, and the extremities will probably be cooler. An electric system seems to be susceptible to severe overheating in an enclosed ensemble, and some active means for circulating coolant through internal instrumentation and to actuators in the limbs may be needed. For example, Hubo uses fans for cooling the main computer, control boards, and DC motors, and HRP-2 has a cooling system for the leg actuators. A hydraulic system may not require cooling but would still need temperature regulation to simulate skin temperature. An onboard hydraulic pump and its power supply will probably need to be cooled. Thus, to maintain a human-like skin-temperature gradient, it will probably be necessary to develop both heat-distribution and heat-dissipation or cooling systems. Having considered various cooling systems, it is feasible to develop distribution and cooling systems for integration into a test mannequin.15 For example, it may be possible to preload the system with a disposable coolant, such as liquid nitrogen or solid CO2, with a circulating liquid and vent exhaust gases as part of the breathing simulation. Spacesuit technolo- gies may be adaptable to the mannequinâs needs, but they usually require the presence of a convenient radiator. A tethered cooling system may be useful for all actuation types. Tethered or Untethered As discussed above, the combination of power and actuation demands and the volume constraints of the human body shape suggest that a tethered system may be a useful option for using current technologies. Once a tether is introduced, it might be used for electric power, hydraulics, communica- 15â See Chapter 2.
88 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT tion and sensing signals, and heating and cooling. However, the integration of the tethered system has two major implications: â¢ Penetration of the ensemble. The tether must pass through the protective ensemble at some point and would therefore typically require custom modification of each ensemble used in each test. In addition, the presence of the tether may introduce some abnormal behavior of the ensemble fabrics and materials that could alter the pattern of penetration. â¢ Forces induced by the tether. The tether must be designed to mini- mize mechanical forces exerted on the mannequin due to the con- nection of the tether. Some existing technologies are able to support the weight of a tether and also may be used to move a tether, reduc- ing unnatural dynamic forces that may result. The presence of a tether raises the issue of where to place the connec- tion. From a broad perspective, the head (helmet) and the back might be logical sites, but both may require modification of important regions of the ensemble and may influence critical performance of the ensemble and air motion. An alternative is to consider the foot (heel or sole). The boot would need to be modified, but perhaps a smaller range of boot designs could be an acceptable trade-off. A tether on the back of the heel might minimize interaction with the mannequin motion and still allow normal foot place- ment in standing, walking, jumping, and crawling tests. In summary, we believe that a tether incorporating electric power, cool- ant, and command signals could be flexible and run through the face plate or the side of the boot to allow longer tests. It may be possible to design a mannequin for which a tether is used only on longer tests. Motion and Control Simulation and monitoring of the Man-in-Simulant Test (MIST) exer- cises by the mannequin will require movement programming and move- ment data recording, both of which are well-established capabilities. For example, programming of humanoid dancing by Ikeuchiâs group is an example of behavior control similar to what is needed for this project,16 and walking, climbing a ladder, sitting, lying down, getting up, and use of small weights have all been demonstrated by robots mentioned earlier in this chapter. As mentioned earlier, many of the required motions could be programmed with the aid of motion-capture technology. 16â Computer Vision Lab: IKEUCHI Laboratory, Institute of Industrial Science, University of Tokyo. www.cvl.iis.u-tokyo.ac.jp. Accessed June 14, 2007.
DESIGN CHALLENGE: ROBOTIC CAPABILITY FOR PETMAN 89 However, for the PETMAN system in particular, specific technical is- sues should be considered. The control systemâs need to handle different weights and weight distributions of different ensembles and carried weap- ons increases the complexity of the required motion programming. It may also be necessary to adapt to the load distribution of water or sweat as it moves from storage areas to the boots and gloves. In addition, because of the human shape and function requirements of the system, there is concern about whether the torques required for jumping jacks, marching, and crawling are achievable with current actuators and transmissions. The recommended motion-capture study should highlight any limitations in this regard. Also, proposing groups should be encouraged to consider a variety of actuation and transmission typesâincluding elec- tric, hydraulic, and pneumatic, perhaps in combinationâto accommodate the full range of required motions. Anthropomorphic Form and Function The PETMAN anthropomorphic shape requirements place severe con- straints on the volume of internal systemsâsuch as power, cooling, com- puting, and communicationâand on the mechanisms themselves, including actuators and associated gearing. Review of current technologies suggests that both constraints will present challenges to the PETMAN designers. Packaging power, cooling, computing, and communication in the torso will be difficult with todayâs technologies for an untethered system. As men- tioned above, the batteries for self-contained operation for the anticipated maximal time of 1-2 h would have substantial weight and volume. Similarly, a computer system, communication electronics, sensor instrumentation, gas exchange, heating-cooling apparatus, and all attendant âplumbingâ would need to be accommodated. The current generation of self-contained humanoid robots with electric motors all exceed the proportional volume and shape requirements of a mannequin because of the bulk of electric actuators and gears. Specifically, it may be difficult to reduce the size of the ankle and wrist to human size and form with electric rotary actuators and harmonic drives (the current technology). These systems also have relatively simple hands and feet and therefore do not meet the functional requirements of PETMAN. The use of hydraulic systems could reduce the bulk of actuators and gearing and meet the form requirements. Similarly, some combination of electric and hydraulic actuators might be engineered to achieve a compromise. Control poses another important challenge to achieving human dimen- sions. Current control approaches, such as zero moment point (ZMP) con- trol, make use of a wide foot. It may be difficult to modify those approaches in the period allocated for development.
90 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT Hands and Feet The PETMAN requirements for human-like hands and feet integrated into a humanoid robot are beyond the state of current technology. Design- ing a robotic hand that fully mimics a human hand is quite difficult, and current robot hands lack human degrees of freedom, range of motion, and speed.17 Available models would be difficult to integrate into a new system, are larger than an average human male hand, and do not meet the threshold (T) requirements of PETMAN. Substantial additional research would be required to develop fully human-like hands as well as feet for a PETMAN system. However, more limited hands and feet are available, and the current requirements go beyond the needs of the project. A useful reference point is the National Aeronautics and Space Administration (NASA) Robonaut sys- tem,18 which uses articulated thumbs and fingers to perform basic grasping and manipulation functions but lacks a number of the specific motions of a human hand. We recommend that a motion-capture study be performed to determine the hand functions required for PETMAN. For feet, a simple study of foot mockups with toe joints should be performed to determine the joints and flexibility necessary to don and doff boots and to accomplish other desired movements. Reliability and Maintenance Evaluating the reliability and maintenance schedule of a PETMAN system is difficult for several reasons. It is not clear what the wear and tear characteristics of the PETMAN tests will be. Reliability has not been emphasized in humanoid-robot research, so the degree to which compo- nents will need to be oversized to make current robot designs reliable is not known; this may have implications for system size, weight, and thermal issues. Finally, given the small number of humanoid robots in existence, we do not have extensive data on maintenance requirements. Cost-Benefit Analysis Developing a PETMAN system that can perform all the desired motions will require substantial modifications of current humanoid-robot designs. Current commercially available humanoid robots cost about $1 million. The estimated cost of developing the motion system of PETMAN would be 17â Shadow Dexterous Hand, Shadow Robot Company, http://www.shadowrobot.com/hand/ Accessed October 9, 2007. 18âRobert Ambrose, National Aeronautics and Space Administration, presentation to com- mittee (see Appendix D).
DESIGN CHALLENGE: ROBOTIC CAPABILITY FOR PETMAN 91 about $10 million. The proposer should be required to address the serious risks associated with the use and development of this system: â¢ The robot falls down and requires expensive repair. â¢ The robot is expensive to decontaminate (see Chapter 5). â¢ The robot cannot be packaged in human form. â¢ Operation for 1-2 h may not be long enough for the sponsorâs needs, but longer tetherless runs cannot be achieved. Feasibility and Potential Alternatives In light of the full list of requested functions, it will be difficult to design an untethered (objective requirement), free-standing, 50th percentile robotic test mannequin with an operational time of two hours. The main design challenge will probably be the integration of all systems (power, control, sensors, respiration, actuation, and so on) within the size constraints; this is discussed in some detail in Chapter 7. Meeting the threshold requirement of a tethered system, which would reduce the number of systems housed in the mannequin, is feasible. It may also be possible to design a system that allows two operational modes: one untethered that meets some of the performance requirements and operational time, and one tethered with greater performance options that is not constrained by battery life and the size of coolant and perspiration reservoirs. Performance might be improved if the sponsor adjusted the design requirements to accommodate only the prescribed motions of the test rather than fully human-like motions (limit DoFs, range of motion, maximal velocities, and so on). With respect to motion and control of the PETMAN system, the con- clusions and recommendations are as follows: Conclusion 4-1: Achievement of the full human range of motion and speed will require the development of new actuators or transmissions. In light of the currently available technology, the PETMAN requirements regarding degrees of freedom are too vague and do not represent the specific needs of the project. Recommendation 4-1a: Specific range-of-motion and maximal-velocity requirements for the desired behaviors should be included in any broad agency announcement (BAA). Recommendation 4-1b: Motion-capture studies of test movements and of donning and doffing should be performed to refine BAA requirements regarding range of motion and maximal joint velocities. This will also be useful for estimating forces and torques necessary for the desired behaviors.
92 SOLDIER PROTECTIVE CLOTHING AND EQUIPMENT An additional motion-capture study should be performed to determine nec- essary hand function required for PETMAN. For the feet, a simple study of foot mockups with toe joints should be performed to determine the foot joints and flexibility necessary to don and doff boots and to accomplish other desired movements. Conclusion 4-2: Achievement of tests longer than two hours will require either limited motion, a recharge-refueling method, a tether, or research on new power sources and actuators With a power line attached to the robot providing unlimited, uninterrupted power, it would be feasible to meet either the 12-h (threshold) or 24-h (objective) operational maintenance requirement. Without a tether, however, the feasibility is more difficult to assess. Conclusion 4-3: It is feasible to develop current distribution and cooling systems for integration into a testing mannequin, but improvement of thermal-regulation technologies (heat distribution, cooling, and so on) will probably be required, especially for an untethered design. The problem is simpler if a tether is available to provide a convenient line for heating and cooling purposes. Conclusion 4-4: Achievement of the human form incorporating all de- sired features may not be possible with current actuation technology. In particular, it may be difficult to reduce the ankle and wrist to human size and form with electric rotary actuators and harmonic drives (the cur- rent technology). Modification of current motion-control approaches may be necessary. The feasibility would increase if a way to run a hydraulic compressor electrically, either with a battery or with electricity delivered through a tether, were developed or if high-performance pneumatic control were demonstrated. Conclusion 4-5: Given the specific, unique requirements of PETMAN and the small number of humanoid robots in existence, there are not enough data available to assess the reliability of the system.