Thermophysiological aspects of wearable robotics: Challenges and opportunities

ABSTRACT

Technological advancements in the last two decades have enabled development of a variety of mechanically supporting wearable robots (i.e. exoskeletons) that are transitioning to practice in medical and industrial settings. The feedback from industry and recent controlled studies is highlighting thermal discomfort as a major reason for the disuse of the devices and a substantial barrier to their long-term adoption. Furthermore, a brief overview of the devices and their intended applications reveals that many of the potential users are likely to face thermal comfort issues because of either high exertion or medically related high heat sensitivity. The aim of this review is to discuss these emerging thermal challenges and opportunities surrounding wearable robots. This review discusses mechanisms, potential solutions, and a platform for systematically measuring heat transfer inhibition caused by wearing of an exoskeleton. Lastly, the potential for substantial metabolic rate reduction provided by exoskeletons to reduce worker thermal strain in warm-to-hot conditions is also considered.

KEYWORDS:

• Thermal comfort
• wearable robotics
• exoskeleton
• personal cooling

Introduction

Exoskeletons, as wearable robotics are commonly known, are increasingly sighted not only on the silver screen but also on factory floors and rehabilitation clinics. As indicated by the name, these devices inspire to have either the protective or the supportive function of the outer shell of insects [1,2] (see an example of an ant exoskeleton [3] in (Figure 1(a)). The former function dominated for most of human history: armor was designed to protect users from harm (see an example of medieval armor in Figure 1b). It is only in the last 130 years that inventors started thinking of the supportive function, as illustrated by the 1890 U.S. Patent [4] for “apparatus for facilitating walking, running, and jumping” shown in Figure 1c. In the 1960s, General Electrics constructed the first full-body and hydraulically powered exoskeleton prototype that weighed 680 kg and was never made operational [5,6] (see Figure 1d). Within the last 20 years, sufficient technological progress has been made to allow engineers to develop practical exoskeletons that can mechanically support the users in executing physical tasks such as lifting a heavy box in a factory [6].

Figure 1. From natural to early human-made protective and supportive exoskeletons: (a) Saharan silver ant exoskeleton is covered by triangular micro-hairs that decrease the radiative heat flux on their bodies [3] (images reused with permission from American Association for the Advancement of Science), (b) medieval armor protected knights from mechanical harm but could contribute to heat illness and exhaustion (part of Map of Jerusalem, c.1200 this work is in the public domain in their country of origin and other countries and areas where the copyright is the author’s life plus 100 years or less), (c) illustration of “apparatus for facilitating walking, running, and jumping” from 1890 U.S. Patent [4] and (d) photograph of “Human Augmentation Research and Development Investigation MANinpulator” or Hardiman worked on but never made operational by General Electrics in the 1960s [5] (image credit Museum of Innovation and Science, Schenectady).

What the knights, the modern factory workers, and even actors wearing a fake super-suit [7] have in common when wearing their mechanically supporting exoskeletons is a thermal problem: they often were or are way too hot! This unfortunate thermal feature of the medieval armor, along with dehydration, was famously exploited by Sultan Saladin to exhaust the crusader army of Guy de Lusignan prior to their devastating defeat at the battle of Hattin in 1187 [8]. In the modern context, early evidence is emerging that thermal discomfort associated with wearing exoskeletons is a leading cause of their disuse by factory workers and is a suggested topic for improvement [9–14]. In contrast, many insects thrive in their shells in scorching environments. One of their secrets? The exoskeletons are often multifunctional, providing both mechanical and thermal functions. For example, Saharan silver ants have exoskeletons that decrease the radiative heat flux on their bodies [3]. In particular, the microscale hairs with triangular cross-sections on the ants’ exoskeletons enhance short-wave radiation reflectivity and increase long-wave radiation emissivity (see Figure 1a) [3]. So, in this review I will argue that we should learn from bugs and include thermal consideration along with the primary biomechanical aspects in wearable robotics design. Specifically, I will provide a brief overview of exoskeletons and early evidence of thermal challenges posed by wearing the devices. I will also discuss opportunities to resolve these issues, a platform for systematic thermal testing of the devices, and potential for their use in decreasing thermal strain of workers in warm-to-hot conditions.

Overview of wearable robots or exoskeletons

Wearable robotics can be defined as devices worn by human operators to supplement or replace a limb’s function completely [15]. The technology “extends, complements, substitutes or enhances the human function and capability or empowers or replaces the human limb where it is worn” [15]. Highlighting synonymity to wearable robots, exoskeletons are defined by the ASTM F3323-19 standard as a “wearable device that augments, enables, assists, and/or enhances physical activity through mechanical interaction with the body” [16,17]. These definitions are different from portable (e.g. smart phone or laptop) or wearable electronics (e.g. smart watch), whose thermal comfort and safety have been reviewed elsewhere [18–20].

Exoskeletons can be classified according to their functions as empowering robotic exoskeletons that extend human strength above natural level, orthotic robots that restore lost or weak function to natural level, and prosthetic robots that substitute a lost limb [15]. Since thermal issues surrounding prostheses and orthoses have been reviewed elsewhere [21–25], I will focus the discussion on the former two classes of devices. The supported region can further distinguish exoskeletons as full-, lower-, upper-body or even more-specific body part (e.g. arm or back) devices [17].

Exoskeletons are also classified by actuation technology and energy source as passive, active, and semi-active [2,17]. The passive devices use load-bearing elements such as springs and dampers to store and release energy harvested from human movement to support postures and motions [17,26]. The active devices use electrical, hydraulic, or pneumatic actuators to support human movement by providing additional strength [17,26]. Semi-active devices use a combination of active and passive elements. Finally, wearable robots can also be classified based on their construction materials as the classical rigid or newer soft devices referred to as exosuits [27–29]. Figure 2 shows examples of a variety of exoskeletons used in diverse applications, including a passive and rigid arm-support exoskeleton used by a worker in a car assembly line [13], a passive and soft lower back supporting exoskeleton used during lifting of a box [30], an active and rigid lower body exoskeleton used for walking by an individual recovering from stroke [31], and an active full-body exoskeleton enabling a worker to hold a 200-pound weight.

Figure 2. Examples of current exoskeletons: (a) a passive and rigid arm-support exoskeleton used by a worker in a car assembly line [13] (reprinted with permission of Taylor and Francis), (b) a passive and soft lower back supporting exoskeleton used by male and female workers in a warehouse (image courtesy of Herowear), (c) an active and rigid lower body exoskeleton used by an individual recovering from stroke [31], and (d) an active full-body exoskeleton enabling a worker to hold a 200 pound weight (source: Sarcos Technology and Robotics Corporation).

Many commercial exoskeletons are marketed for specific uses in medical, elderly support, military, and industrial applications [2,15]. In clinics, a variety of the devices have been proven advantageous to rehabilitation and autonomy recovery after stroke [32–34] and spinal cord injury [34–36] as well as for individuals with neurological conditions such as multiple sclerosis [37–41]. Remarkably, exoskeletons enable some individuals with paraplegia to carry out routine ambulatory functions [35,36]. Since physical strength of many individuals decreases with age, numerous exoskeletons have also been developed to support the elderly in moving around safely (e.g. preventing falls), remaining active, and even delaying retirement from physically demanding jobs [42–44]. In turn, devices targeted for military applications increase the physical abilities of soldiers so that, for example, they can carry heavier loads [1,45,46]. Exoskeletons intended for industrial settings are marketed as an approach to reduce fatigue and in some cases strengthen and enhance worker productivity, and as devices that increase the user’s comfort and long-term occupational health outcomes [2,17,26].

Many occupational exoskeletons are designed to address biomechanical burdens workers face due to excessive and repeated physical tasks in, among others, automotive [47], construction [48], agriculture [14], retail [49], and steel [50] industries. For example, lifting is a common source of low back disorders, which despite the implementation of government-mandated interventions, still account for about 40% of work-related musculoskeletal disorders reported in the United States [30]. In general, musculoskeletal issues correspond to over half of the work-related health problems (e.g. 60% in the European Union [17,26]) and substantially reduce impacted individuals’ quality of life and work activities. The use of exoskeletons during occupational tasks reduces the user’s physical stress and strain [17,26,51]. These benefits are projected to reduce cumulative tissue damage and risk of musculoskeletal disease in the long term [30]. In the last 5 years, many commercial exoskeletons [30,51] have been implemented in industrial practice [30]. This transition has motivated research into potential unintended side-effects and user experiences [51]. Naturally, a positive user experience, including thermal comfort, is fundamental to the acceptance and widespread adaptation of the devices in workplaces and beyond [17].

Current thermal comfort challenges associated with exoskeletons

In feedback from early exoskeleton adopters in the industry, thermal discomfort associated with wearing the devices is emerging as a major reason for the disuse of the devices and a substantial barrier to their long-term adoption [11,12]. In the industrial context, the thermal discomfort of workers wearing such devices is not surprising, considering the large attachment areas of the exoskeletons that inhibit heat dissipation and high user metabolic rates (e.g. ~200 Wm⁻² even after reduction associated with wearing an exoskeleton [9]). However, thermal issues are also likely to arise even when metabolic rates are much lower.

Many potential elderly and medical exoskeleton users have various heat perception and sensitivity issues that might impact their experience with the devices. For most humans the whole-body sweat rate decreases with age, which leads to reduced heat dissipation capacity and increased susceptibility to heat illness [52]. Since wearing an exoskeleton will further inhibit their heat dissipation capacity, elderly users might experience thermal discomfort even when wearing the device to aid in safely moving around or conducting moderate physical activities. In turn, for 60% to 80% of individuals with multiple sclerosis, neurological symptoms including cognitive, sensory, and motor impairments are temporarily exacerbated by even small environment or exercise-induced body temperature increase [53–55]. When wearing of an exoskeleton to aid gait rehabilitation, an individual might cool less effectively, experience faster body temperature increase, and potential symptom exacerbation.

Many individuals recovering from stroke or with spinal cord injury also experience thermal symptoms that might alter, although in harder to foresee ways, their experience with exoskeletons. Specifically, spinal cord injury can affect an individual’s thermoregulation by altering heat production by reducing muscle mass and heat dissipation by blood flow redistribution and reduced sweating capacity [56–58]. Similarly, over 60% of patients recovering from stroke report alteration in thermal sensation and may have plegic limb skin 1 to 5°C colder than unaffected regions [59,60]. Thus, a significant fraction of exoskeleton users can be expected to be impacted by thermal interaction with the device. Considering the plethora of research and development activity around exoskeletons, it is surprising that wearable robotics’ thermal aspects have not received any attention prior to 2019 [17]. In this Section, I will provide an overview of the most comprehensive trial published on the thermophysiological impacts of exoskeletons [9] and use it to discuss external and internal thermal and physiological alterations induced by wearing an exoskeleton.

Liu et al. [9] measured skin temperature at multiple sites and metabolic rate and recorded thermal comfort, thermal sensation, and sweat feeling metrics of 10 male package handlers repeatedly lifting a box with (EXO) and without (WEXO) an exoskeleton in cold (10°C and 50% relative humidity) and temperate (20°C and 50% relative humidity) indoor conditions. The skin temperature measurements on forehead, chest, upper arm, forearm, hand, anterior thigh, anterior calf, and foot of the right side of the body were reported as mean skin temperature calculated using the Gagge/Nishi’s equation [9]. The subjects wore a 3 kg back-supporting passive exoskeleton that included spacers on the chest and thighs and a supporting structure containing a damping mechanism at the hips (see Figure 3a). During lifting, the lower back muscle activity was reduced by the spacers providing support to the upper body. The damping mechanism around the hips provided the support force (see Figure 3b). The subjects were young and healthy male package handlers with at least 1 year of work experience and wore long underpants, straight trousers, long-sleeved sweatshirt, socks, boots, and a jacket totaling about 1.3 clo (0.2 m²°CW⁻¹) insulation. The experimental procedure consisted of sequential 40 minute preparation, 20 minute adaptation, 20 minute rest, 5 minute questionnaire, 20 minute lifting, and 5 minute questionnaire stages. The lifting activity was designed according to the National Institute of Occupational Safety and Health (NIOSH) equation [61] and consisted of lifting and lowering a 10 kg box to and from a 75 cm platform five times per minute. The physiological parameters were continually measured, while the subjective metrics were quantified during the before and after lifting questionnaire stages [9].

Figure 3. Thermophysiological and thermal perception impacts of wearing an exoskeleton quantified by Liu et al. [9]: (a) photograph of the passive, back-supporting device and (b) schematic of its use; (c) the measured metabolic rate, and recorded (d) thermal sensation vote, (e) thermal comfort vote, and (f) sweat feeling index metrics. The figures are licensed under open access Creative Commons CC By 4.0 license.

The exoskeleton had a negligible impact on the mean skin temperature but did substantially reduce the steady-state metabolic rate of the package handlers. It is important to point out that temperature, thermal sensitivity, and contribution to thermal comfort sensation can vary significantly between skin regions [62,63]. Consequently, a negligible change in mean skin temperature that is an average of multiple sites might not fully reflect potential significant temperature rise at the device attachment location and associated thermal discomfort. In turn, wearing of the exoskeleton did substantially reduce the average metabolic rate during lifting from 313 $\pm$ 32 Wm⁻² to 225 $\pm$ 27 Wm⁻² and from 273 $\pm$ 45 Wm⁻² to 198 $\pm$ 33 Wm⁻² in the temperate and the cold conditions, respectively (see Figure 3c). Other studies have reported a similar metabolic rate reduction associated with wearing an exoskeleton [64]. The devices assist muscle groups specific to the target motion and decrease the muscles’ activity [64].

Despite the reduction in metabolic rate, however, package handlers wearing the device during lifting in the temperate conditions reported increased sweating and feeling warmer, and thus overall feeling more uncomfortable than those without the device (see Figure 3d-f). In contrast, workers wearing the exoskeleton when lifting in the cold conditions sweated less and felt warmer, and overall were more comfortable than those without the device. Liu et al. [9] concluded that by increasing the local thermal insulation of the clothing and reducing the exerted effort, the exoskeletons improve workers’ comfort and thereby are well suited for use in cold conditions. In temperate conditions, the increased local thermal insulation associated with the device negates the thermal benefits of reduced metabolic rate and makes the device very uncomfortable. Next, I discuss the local and whole-body impacts of wearing an exoskeleton from a heat transfer perspective.

Wearing an exoskeleton reduces the user’s local ability to cool by covering skin with thermal insulation and preventing sweat evaporation [9,10,13] (see Figure 4a). These processes are difficult to avoid and minimize with body interface size since the devices must attach directly to the body to apply the assistive force and have a specific area to avoid uncomfortable pressure on the skin [10]. An example body interface consists of a thin and soft memory foam layer (~0.3 cm) that is supported by a thicker rigid plastic layer (~2.5 cm) [10]. With thermal conductivities of about 0.04 Wm⁻¹°C⁻¹ and 0.2 Wm⁻¹ C⁻¹ [65], the two layers have a total thermal resistance of around 0.2 m²°CW⁻¹, which is an order of magnitude higher than typical short or long sleeve shirt (0.1 to 0.15 clo [66] or 0.015 to 0.023 m²°C W⁻¹). In other words, the body interface is a thermal insulation layer that locally negates most of the conductive heat transfer from the body to the environment. As in the case of casts, thermal aspects of the body interfaces could be improved by three-dimensional printing of partially hollow attachments [67]. Conductive heat loss could also be enhanced by the use of more thermally conductive materials in the body interface, such as carbon fiber or metals on the outer layer [9] and potentially novel soft composites (e.g. elastomers with liquid metals [68]) in the inner layer. However, heat conduction alone is unlikely to be sufficient to replace the heat sink provided by sweat evaporation.

Figure 4. (a) Schematic of human-surrounding heat transfer processes impacted by wearing an exoskeleton, (b) dual-mode exoskeleton body interface that temporarily loosens when the device is not used [10], (c) example phase-change cooling garment (collar in this case [74] image reused with permission) that is used to directly cool skin and could be integrated into rapidly swappable packets into the mode-switching body interfaces, (d) schematic of the 35-independently controlled surface zones and (e) photograph of ANDI thermal manikin and walking motion stand during protective garment testing (image source: Thermetrics), and (f) schematic of the total clothing ($I_t$), the intrinsic clothing ($I_{cl}$), and the environment () thermal resistances.

By its solid nature and direct attachment, the body interface also prevents the evaporation of sweat. The inhibited cooling ability can increase local skin temperature by several degrees Celsius [10] and, along with increased wetness, contributes to the user’s increased thermal discomfort. The increased temperature and wetness could also alter local friction [69,70], leading to more skin chafing, which is a major issue with exoskeleton use [71]. In particular, temperature mechanically softens the skin tissue, substantially increasing the surface friction coefficient [69]. Beyond these negative local impacts, attaching the body interfaces also has whole-body consequences.

The attachment of the exoskeleton restricts clothing movement, which can substantially reduce the convective loss from surrounding body parts [65]. Specifically, the movement of clothing caused by human motion pumps air within the fabric-skin gap and substantially enhances heat loss from the skin (see Figure 4a). So far, this process has only been measured during walking [66,72] and found to reduce the effective thermal insulation of clothing even by half [66]. Since the motion of air by the skin drives sweat evaporation [65], the reduction of the convection is also associated with proportionally decreased evaporative cooling capability.

Besides reducing the body’s cooling ability, powered exoskeletons can expose the body to additional external heating. Fortunately, full-body exoskeletons had improved substantially from the early 2000s when they required several thousand watts to operate. For example, the early 6000 W power requirement of Sarcos devices, that from a thermal perspective could have been considered an elaborate human cooking machine, was reduced to just 400 W in the latest models [73]. However, when operating in warm-to-hot environments, the waste heat from the device might not dissipate efficiently to the surroundings and provide another substantial heat source for the user. Similarly, in sufficiently hot conditions, the surrounding itself will switch from a heat sink to a heat source for the user. In such conditions, besides application of current approaches to workplace heat safety (e.g. increasing air conditioning where available, providing shading, modifying work-rest intervals, etc.), exoskeletons with modified designs could prove to be beneficial. In the next Section, I discuss potential device design alterations to reduce the whole-body and local heating associated with exoskeleton use.

Opportunities for improvement of thermal aspects of exoskeleton design

While several simple routes can minimize the whole-body thermal exoskeleton impacts, minimizing local heating within the body interfaces is more challenging and may require more advanced solutions. To start with the most straightforward point that even some medieval knights got right, devices intended for outdoor use in, for example, agricultural [14,75] or construction [48] industries should have exterior surfaces that reflect short wave radiation. Instead of putting a white cloak over armor as knights [76], wearable robotics designers have a variety of novel coatings or even fabrics with engineered spectral properties to choose from. These materials reflect nearly all shortwave radiation but are either highly reflective, emissive, or even transparent to longwave radiation [77–79]. The infrared reflective materials are suited for extremely hot conditions like Phoenix in the summer, where the surrounding surfaces are much hotter than skin and thus provide a strong infrared heating source [80]. When the surrounding is not as hot, the highly infrared emissive materials can cool to sub-ambient temperatures by radiating to space through the atmospheric window [77]. Infrared transparent materials such as polyethylene textiles and thin films can help cool the body radiatively in moderate indoor or outdoor conditions. However, polyethylene becomes exponentially more absorbing with thickness, so it is not helpful as a construction material for body interface materials.

The pinching of fabric by the exoskeleton could be minimized by attaching the devices directly to the skin or underneath loose clothing. Alternatively, air movement in the clothing-skin gap can be mechanically induced using ventilated clothing with small integrated fans. These simple garments substantially enhance sweat evaporation [81,82] and are the most cost-effective and implementable personal cooling garment for outdoor workers [83]. Since traditional routes to increase ventilation of the body interfaces, such as partial meshing and adding airflow channels, provide only minor improvements in reported thermal comfort [13], forced ventilation, albeit with smaller fans, could also be used to cool down the body interfaces.

As most people wearing a backpack in hot weather can attest to, even advanced straps with meshes or air flow channels become hot and sweaty. While simulation-driven optimization of the ducts can improve air flow across the back of a moving cyclist [84], a remarkable improvement of comfort with this approach in near stagnant conditions that most exoskeletons will be used in seems unlikely. Since even a face mask can have integrated small fans [85], the addition of active ventilation and intelligently designed air ducts in the foam layer in most problematic body interfaces could be a simple option to improve the thermal comfort of exoskeletons.

Elstub et al. [10] recently demonstrated an even more straightforward and clever way for “on-and-off” convective cooling of the body interface area. A pilot study showed that thermal comfort and skin temperature could be improved by temporarily disengaging and loosening the body attachments when the exoskeleton is not used. The images in Figure 4b show that the loosen interfaces can hang on straps attached to a belt around the user’s waist. The loosening of the form-fitting sleeve after 25 minutes of physical activity caused a rapid drop of skin temperature by 2 to 4°C and improved the thermal comfort reported by two out of the four subjects. In addition, there appear to be quick, practical, and easy technical solutions to implementing the mode-switching approach [86]. Consequently, the dual-mode body interfaces provide a simple and promising way to improve exoskeleton user thermal comfort in scenarios where the device attachments can be relatively frequently disengaged.

The soft interface layer could include a gel-type phase change material to prevent skin temperature increase when the exoskeleton is engaged. Many phase-change materials are already used in neck and trunk cooling garments [74,81,87]. Since practically sized packets of such materials provide cooling only for a limited time, the body interface could include a mechanism for swapping out the phase-change cartridges when loosened. Finally, in principle, the body interface cooling could also be achieved using liquid cooling even within the soft part of the interface [88] or wearable thermoelectrics [89]. However, these active cooling methods require heavy support equipment that might not to be practical for exoskeleton use (e.g. based on personal measurements a typical commercial liquid cooled vest with 2.5 kg ice-filled plastic bladder and pumping system carried in a backpack has a mass of about 6 kg). Next, I discuss opportunities for systematic evaluation of the thermal performance of current and improved exoskeleton designs.

Opportunities for standardizing thermal testing of exoskeletons

As for clothing and protective garments [66,90], many thermal aspects of exoskeletons can be systematically compared using thermal manikins. These human-shaped instruments can quantify dry and wet human-surrounding heat exchange on multiple body zones. For example, the ANDI thermal manikin from Thermetrics independently measures temperature and heat flux within 35 body surface zones (see Figure 4d). The instrument can also generate “metabolic” heat and dispense “sweat” within each zone (albeit the input values need to come from human trials) and has articulated shoulders, elbows, hips, knees, and ankles. The joints’ mobility allows for quantifying air pumping effects during mimicked motions such as walking that is imposed on the manikin by an external stand (see Figure 4e). To be relevant to common wearable robotics use, a custom motion-stand that mimics, for example, a box lifting cycle could be developed.

The thermal manikin in constant-surface temperature mode can measure how the exoskeleton changes the thermal and evaporative resistance of clothing. Figure 4f illustrates the definitions of the total clothing ($I_t$ in m²°CW⁻¹ units), the intrinsic clothing ($I_{cl}$), and the environment ($I_a$) thermal resistances [66]. The $I_{cl}$ takes into account thermal transport across the skin-fabric gap and the fabric itself while captures the outer thermal transport processes (convection and radiation). The latter parameter is calculated from measurements of the heat required ($H_{Nude}$) to maintain a nude thermal manikin at a constant shell temperature ($T_{shell}$ of 32 to 34°C) as [66]:

(1) $I_a=\frac{T_{shell}-T_o}{H_{Nude}}$(1)

where the operative surrounding temperature ($T_o$) should be at least 12°C lowered than $T_{shell}$ and is equal to air and radiant temperatures (conditions ensured by a climatic chamber). Similarly, the $I_t$ that covers both $I_{cl}$ and $I_a$ is calculated from the measurement of the heat required ($H_{Cl}$) to maintain a clothed thermal manikin at the constant $T_{shell}$ as [66]:

(2)

The is calculated from these two measurements as [66]:

(3) $I_{cl}=I_t-\frac{I_a}{f_{Cl}}$(3)

where $f_{Cl}$ is the ratio of the surface area of the clothed to nude surface areas of the body. In addition to the above “dry” resistances, the manikin can dispense water onto an elastic fabric “skin” and measure the evaporative resistance (in m²kPaW⁻¹ units). This value is calculated from the power ($H_e$) required to maintain a manikin covered by the wet “sweating skin” at a constant $T_{shell}$ as

(4) $R_{e,t}=\frac{\left(P_{shell}-P_{air}\right)A_t}{H_{e,t}-\frac{\left(T_{shell}-T_o\right)A_t}{I_t}}$(4)

where $P_{shell}$ and $P_{air}$ are the water vapor pressures at the shell and air temperatures, respectively. In Eq.4(4) $R_{e,t}=\frac{\left(P_{shell}-P_{air}\right)A_t}{H_{e,t}-\frac{\left(T_{shell}-T_o\right)A_t}{I_t}}$(4) , the subscript “t” indicates that the total evaporative resistance () is scaled by the total surface area that is sweating. In addition to total values, the $I_t,I_a$, and $R_e$ can be calculated for each of the surface zones. Since these values (local or total) are strongly impacted by clothing motion, the resistances need to be measured for each exoskeleton-clothing pair. To isolate the exoskeleton’s impact, the clothing’s resistances need to be compared to the resistances of the clothing and exoskeleton “ensemble.” Different designs of exoskeletons could be rapidly compared by standardizing the attire (e.g. for summer and winter) with which the devices are tested. Lastly, I note that current thermal manikins have a hard exterior shell that is not representative of direct skin contact conductive resistance (i.e. the shell has different interfacial contact resistance than skin when in contact with various objects [91]). This issue, however, would only need to be addressed if substantial conductive heat transfer occurs (e.g. if all layers of the body interfaces are modified with high thermal conductivity materials).

Opportunity for using exoskeletons to reduce hot weather impacts on worker performance and health

For diverse industry sectors such as agriculture, construction, and manufacturing that are often conducted in warm and extremely hot settings, exoskeletons with thermally improved aspects offer a potential novel route to improving workplace safety and performance by reducing worker thermal stress. Millions of workers worldwide work in such conditions that can cause a wide range of illnesses and reduce performance and productivity [83,92–94]. The potential impact of exoskeleton deployment can be illustrated by the ~100 Wm⁻² metabolic rate reduction measured by Liu et al. in temperate conditions [9]. In particular, this 550 W to 400 W reduction in metabolic rate corresponds to switching from “heavy work” (440 to 704 W) to “moderate work” (191 to 411 W) category in the work/rest times chart based on wet bulb globe temperature (WBGT) index [95–97]. This implies that, if body interfaces can be improved, the use of exoskeletons in warm-to-hot conditions could boost work periods in such conditions by 50 to 100% (e.g. from 10 to 20 min in WBGT heat category “5” and from 20 to 30 min in WBGT heat category “4”). In other words, for several common working scenarios like lifting, exoskeletons might be a new tool for addressing one of the main health and economic impacts of increasingly common extremely hot weather associated with climate change [83,98]. In this context, the wearable robots should be used along with the “other” wearables – in particular along with wearable physiological monitors that provide real-time assessment of an individual’s heat strain [99]. Exoskeletons could ultimately also be combined with a variety of personal cooling garments to create the “cool future fashion” that I previously argued in this Journal might be part of our adaptation to the hotter climate [100].

Conclusions

In this review, I discussed emerging challenges and opportunities surrounding thermal aspects of wearable robots. Technological advancements in the last two decades have enabled development of a variety of such mechanically supporting devices that are transitioning to practice in medical and industrial settings [30]. The feedback from industry is highlighting thermal discomfort as a major reason for the disuse of the exoskeletons and a substantial barrier to their long-term adoption [11,12]. Confirming the industrial feedback, Liu et al. [9] recently demonstrated in a controlled study that wearing a back-supporting passive exoskeleton while repeatedly lifting a heavy box in temperate conditions substantially decreases metabolic rate but increases sweating and thermal discomfort of the package handlers. In the context of users exerting moderate-to-high level of physical activity, the thermal discomfort posed by wearing devices that are tightly attached to skin is not surprising. However, thermal issues are likely to arise even in medical and rehabilitative settings where exoskeleton users have much lower metabolic rates. Specifically, the elderly have reduced heat dissipation capacity while many individuals impacted by stroke, spinal cord injury, or neurological conditions such as multiple sclerosis have various heat perception and sensitivity issues. Since a positive user experience, including thermal comfort, is fundamental to the acceptance and widespread adaptation of the exoskeletons, the device designs need to also consider thermal aspects.

Next, I reviewed mechanisms and potential solutions to the whole-body and local heat transfer inhibition related to wearing of an exoskeleton. On whole-body scale, pinching of fabric by the device body interfaces decreases convection, and thus sweat evaporation, in the skin-clothing gap [9,10]. This problem could be resolved by simply attaching the exoskeleton directly to the skin through small gaps in clothing or using garments actively ventilated with integrated fans. Similarly, the local inhibition of heat transfer under the solid and thermally insulating body interfaces could be alleviated with incorporation of smaller fans and cooling channels. In settings where exoskeleton users pause periodically and have access to cold storage (e.g. a cooler), body-interfaces can be designed to temporarily disengage, which in itself improves thermal comfort [10], and include quickly swappable phase-change cooling pads [74,81,87]. The melting pads could prevent local skin heating when exoskeleton is used for ~30-45 minutes and be rapidly replaced during the break. Irrelevant of the proposed thermal solution, exoskeleton designs should be tested using standardized method.

Exoskeleton designers could adapt thermal manikins used in clothing and protective garment industries [66,90] to systematically evaluate thermal aspects of current and future exoskeletons. A custom motion stand that, for example, imposes a walking and lifting motion on the manikin could provide a platform representative of many industrial exoskeleton applications. In constant surface temperature (or heat flux) mode thermal manikin experimentation will provide quantitative metrics (thermal and evaporative resistances) to compare exoskeleton designs. In adaptive mode where the manikin heating and sweating rates are controlled using a human thermoregulation model [90], the manikin could be used to explore whether improved designs of exoskeletons could reduce worker thermal strain in warm-to-hot conditions. As I pointed out the substantial reduction of the metabolic rate associated with the exoskeleton use has the potential to reduce thermal strain and boost work periods in such conditions. Accordingly, exoskeletons might be a new tool for addressing one of the main health and economic impacts of increasingly common extremely hot weather. However, exoskeletons use in such context require improvement of their thermal design and should be carefully studied prior to any practical implementation.

Acknowledgments

The author would like to thank Professor Thomas Sugar and research associates Shri H. Viswanathan and Daniel M. Martinez from Arizona State University for commenting on the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author.