Human-Machine Interfacing: Soft Structures for Strapping

In exoskeleton research, including powered and passive orthoses, interfacing with the human has been described as one of the main obstacles.  From Zaroodny in 1963 to Herr’s Robohub interview in August of 2016, it has been agreed upon that attaching exoskeletons to the human body is hard.  There are many other problems to solve in making exoskeletons feasible.  Mobile power, sensing and controls, building better actuation methods, and understanding the biomechanics of walking are among the priorities.  These issues have several teams of researchers working on answers.  While the entire device is interconnected, there is very little academic research being done with the primary focus around how to attach these devices to the human body.  While interfacing with the human body encompasses a number of design topics, this post will focus on soft structures for strapping. 

Soft Structures for Strapping

The Problem

Robotics is a multidisciplinary field of study.  The ideation and fabrication of robotic-based exoskeletons and orthoses require integrating knowledge from human musculoskeletal biology, mechanical engineering, electrical engineering, machine learning, artificial intelligence, and apparel design.  However, it is rare that any team resolve all problems concurrently.  Often, well designed mechanical devices show less potential than they have due to the failures of integration with the human body.  Torque is lost in the deflection of soft tissue or the deformation of the attachment materials upon actuation.

Most exoskeleton teams have the goal of creating a device that is in part in line with the vision of science fiction when it comes to strapping methodology.  Push a button and the whole thing falls off.  Click one strap close and in 5 seconds and you’re ready to run.  This is not without good reason.  In designing for military application, safety is critical.  If the wearer is stuck in the exoskeleton, it has a high potential to create a dangerous situation. 

Existing Methods of Attachment

There are a few examples of garments that deal with attaching to lower appendages.  Hosiery is among the consumer products that do.  The hosiery solution relies on elastics that dig into soft tissue and garter belts which hang garments off the waist.  In military uniform, there is a tradition of calf garters to hold up dress socks.  These garters consist of a band that is tight around the small of the calf just below the knee, which works as an effective attachment point on active individuals.  These garments are generally light weight and the only forces working against them are gravity and the motion of the wearer.  Passive orthoses and braces are often used during injury to immobilize the joints or create lateral stability for the wearer.  Exoskeletons, on the other hand, seek to create high mobility in individuals and look to distribute externally created torque evenly across the body of the wearer.  Without doing so, concentrated pressure points cause severe discomfort, bruising and even bleeding.

There have been several conventional ways in which garments and other objects have been attached to the human body.  Those methods can be broken down into three basic categories: belted, hung, and wrapped.

Traditional tailoring and the bottoms most of us are wearing today are belted.  Without close fit around the smallest circumference of the main body, the garment or object is prone to slipping off the body.  For this reason, often belted garments have a method of adjustment by means buckles or velcro.  Belting without some bifurcation or bracing will result in twisting during motion.  This can be seen in everyday garments like skirts as well as in actuated exoskeletons.

Hung garments are the most independently capable of carrying load.  In the hanging concept, the body behaves as a divider between the weight of the object and the gravity drawing it towards the ground.  While garments could be hung from a belt, like in the case of hosiery, they’re often hung from the shoulders.  The shoulders provide the one horizontal attachment point on the human body other than the head.

Pointe shoes, ace bandages, and finger traps all have at least one thing in common: they’re wrapped.  Wrapping may be applied to all semantic parts of the human body effectively.  A compliant material may be wrapped around the body of the wearer.  While some amount of slipping may occur, it is less likely to occur than in other methods.  Wrapping is also size-agnostic and when used in a crisscrossing configuration does not create severe pressure points along the edges.  It is less likely to constrain the soft tissue of the wearer in a way that is uncomfortable.  Braiding could be considered as a part of wrapping.  As you follow an individual strand of a braided tube, you will find that each strand wraps continuously.

There is one property that occurs across categories: compression.  When I refer to compression based garments, I am specifically referring to elastic based knits.  This description covers everything from ace bandages, medical compression, to basic running tights.  Compression based garments have the greatest potential for distribution of pressure across the body.  However, they also have the greatest amount deformation. Due to their knit structure, which is based off a series of loops, rather than a woven structure which is based on a configuration of flatly laid yarns, they are more prone to abrasion damage.  This is becoming less true as we introduce long-chain polymers like UHMWPE or PTFE which can be used as fibers and are incredibly resistant to abrasion.


The best solution to provide distributed torque and comfort to the user is wrapping.  The Lines of Non-Extension (LoNE) have been studied for the purposes of surgery and strapping in order to understand where the skin’s surface is stretching the least amount. It has been used for decades, even centuries in bandaging.  However, it is time-consuming, inconvenient, and difficult to interface with.   It does not take into account the softness of the tissue or the depth of the bone.  Pulling on one part of a wrapped structure leads to loosening of that part and tensioning of another.  This effect will be exacerbated in areas of high softness and therefore deformation.  Any wrapped structure is also time-consuming and difficult to don and doff.  

In dealing with minimal forces, a catenary structure is naturally formed by compliant materials.  In actuation, there is pulling and a catenary with a point load is often the geometry reflected in strapping.  The point will either terminate at the site of actuation or opposite to actuation.  This shape can be utilized for its benefits of diagonality.  As human form deforms and takes on larger circumferences in some areas during flexion and smaller circumferences during extension.  A diagonal shape can be the solution to this measurement change.  During extension, the diagonal may act as a catenary with its imagined center point as the acting circumference.  During flexion, the diagonal may be repurposed at the circumference, giving it a longer measurement.

What We Can Learn from Bras

Some garments use several methods I have described to achieve greater functionality.  There is one garment that has more physical functionality than any other garment: the bra.  The bra exemplifies many of the characteristics of strapping of exoskeletons.  The physical challenge the bra deals with is based on constant load rather than active torque.  Nonetheless, forces are applied.

The overall attachment of the bra demonstrates that it both hangs from the shoulders and is belted just below the breasts.  This gives greater load distribution across the upper body than either single method would.  The strapping is also compression based.  The elastic has a lower deformation gradient, meaning it is more difficult to stretch and stretches less than other elastics. 

The way bras are closed is also related.  The larger coverage of the bra is where force is applied, at the breasts.  The structure then comes to a point at the closure in the back as well as at the straps, in opposition to the load.  This is similar to the catenary structure described.  Smaller amounts of back coverage mean better heat management but does not necessarily make the bra more physically comfortable.  A bra with greater back and strap coverage is generally more comfortable for carrying heavier loads.  However, a bra with smaller back coverage is easier to clasp on due to fewer points of contact.  It is also more versatile for aesthetic reasons.  When it comes to functionality, fit is critical.

The fact is, bras are a miserable garment to wear.  Yet half the population wears bras regularly because it can also be miserably uncomfortable not wearing a bra.  There will always be a trade-off.  The ability to weigh costs against benefit is a critical takeaway.  There will be no panacea.  Each solution should be evaluated for its focus on the priority goals.  In the case of exoskeletons, you cannot solve for every design constraint right away.


Research has shown that during human walking and low speed running most power output occurs at the ankle.  If we focus on this specific joint, we can highlight the challenges of strapping.

Each wearer has a different shaped calf.  Particularly in subjects with lower extremity impairment, the shape of the calf will have less muscle tone and take on an inverted conical shape.  This challenge is also expressed in able-bodied individuals due to the fact that the calf is made up of soft tissue.  The area of the leg that is reliably smallest in circumference below the waist is the ankle.  An attachment point above the ankle is required for actuation introducing an obvious problem.  Even the waist provides a difficult strapping point as belts can travel upward and downward several inches on the wearer.  The most stable strapping point is the shoulders, but even that is unreliable because of motion in the torso.

Fit Customization for Wearable Robotics

Due to the importance of accuracy of fit in order for exoskeletons to function, customization will be required.  Tests of metabolic benefit have often reported large variation among users.   Small anatomical differences in the wearer are likely a major cause for these differences.  If we do not customize the fit of these devices, we lead with uncertainty.  We may abandon methods that have more potential than we credit to them.

There are many false preconceptions when it comes to fit.  For instance, baggier clothing does not necessarily inhibit range of motion for lower extremity.  A dropped crotch, often associated with baggy fit can get in the way of hip abduction and adduction.  These are critical movement when it comes to squatting.  Tight fitting clothing, more obviously has the ability to prohibit motion because there is simply not enough fabric.  Tightness in the lower parts of the leg can cause pants to pull downward even when there is elasticity in the fabric. 

Proper fit is the best solution for functionality.  Good fit is hard to define for consumer products.  Most people have varying opinions about how clothing should fit based on how it feels.  In the case of exoskeletons, it is unclear whether or not this will be true.  While some people will prefer feelings of tightness others will prefer looseness.  The feedback of each wearer will be subjective to their experience and expectation.