The earliest exoskeleton-like device was a set of walking, jumping and running assisted apparatus developed in 1890 by a Russian named Nicholas Yagin. As a unit, the apparatus used compressed gas bags to store energy that would assist with movements, although it was passive in operation and required human power. In 1917, United States inventor Leslie C. Kelley developed what he called a pedomotor, which operated on steam power with artificial ligaments acting in parallel to the wearers movements. With the pedomotor, energy could be generated apart from the user.
The first true exoskeleton in the sense of being a mobile machine integrated with human movements was co-developed by General Electric and the United States Armed Forces in the 1960s. The suit was named Hardiman, and made lifting 110 kilograms (250 lb) feel like lifting 4.5 kilograms (10 lb). Powered by hydraulics and electricity, the suit allowed the wearer to amplify their strength by a factor of 25, so that lifting 25 kilograms was as easy as lifting one kilogram without the suit. A feature dubbed force feedback enabled the wearer to feel the forces and objects being manipulated.
While the general idea sounded somewhat promising, the actual Hardiman had major limitations. It was impractical, due to its 680-kilogram (1,500 lb) weight. Another issue was that it is a master-slave system, where the operator is in a master suit, which, in turn, is inside the slave suit that responds to the master and handles the workload. This multiple physical layer type of operation may work fine, but responds slower than a single physical layer. When the goal is physical enhancement, response time matters. Its slow walking speed of 0.76 metres per second (2.5 ft/s) further limited practical uses. The project was not successful. Any attempt to use the full exoskeleton resulted in a violent uncontrolled motion, and as a result it was never tested with a human inside. Further research concentrated on one arm. Although it could lift its specified load of 340 kg (750 lb), it weighed three quarters of a ton, just over twice the liftable load. Without getting all the components to work together, the practical uses for the Hardiman project were limited.
The beginning of the development of humanoid robotics coincided with the beginning of the development of the world's first active exoskeletons at the Mihailo Pupin Institute in 1969, under the guidance of Prof. Vukobratovic. Legged locomotion systems were developed first. Also, the first theory of these systems was developed in the same institute, in the frame of active exoskeletons. Hence, it can be said that active exoskeletons were the predecessors of the modern high-performance humanoid robots. The present-day active exoskeletons are developed as the systems for enhancing capabilities of the natural human skeletal system. The most successful version of an active exoskeleton for rehabilitation of paraplegics and similar disabled persons, pneumatically powered and electronically programmed was realized and tested at Belgrade Orthopedic Clinic in 1972. One specimen was delivered to the Central Institute for Traumatology and Orthopedy, Moscow in the frame of the USSR-Yugoslav inter-state scientific cooperation. From 1991 the exoskeleton belongs to the basic fund of Polytechnic Museum (Moscow) and State Museum Fund of Russian Federation. It is displayed in the frame of the Museum's exposition dedicated to the development of automation and cybernetics.
Los Alamos Laboratories worked on an exoskeleton project in the 1960s called Project Pitman. In 1986, an exoskeleton prototype called the LIFESUIT was created by Monty Reed, a United States Army Ranger who had broken his back in a parachute accident. While recovering in the hospital, he read Robert Heinlein's Starship Troopers and from Heinlein's description of Mobile Infantry Power Suits, he designed the LIFESUIT, and wrote letters to the military about his plans for the LIFESUIT. In 2001 LIFESUIT One (LSI) was built. In 2003 LS6 was able to record and play back a human gait. In 2005 LS12 was worn in a foot race known as the Saint Patrick's' Day Dash in Seattle, Washington. Monty Reed and LIFESUIT XII set the Land Speed Distance Record for walking in robot suits. LS12 completed the 4.8-kilometre (3 mi) race in 90 minutes. The current LIFESUIT prototype 14 can walk 1.6 km (1 mi) on a full charge and lift 92 kg (203 lb) for the wearer.
In January 2007, Newsweek magazine reported that the Pentagon had granted development funds to the University of Texas at Dallas' nanotechnologist Ray Baughman to develop military-grade artificial electroactive polymers. These electrically contractive fibers are intended to increase the strength-to-weight ratio of movement systems in military powered armor.
One of the main applications would be medical — improving the quality of life of persons who have, for example, lost the use of their legs, by providing assistive technology to enable system-assisted walking or restoration of other motor controls lost due to illness or accidental injury.
Another area of application could be medical care, nursing in particular. Faced with the impending shortage of medical professionals and the increasing number of people in elderly care, several teams of Japanese engineers have developed exoskeletons designed to help nurses lift and carry patients.
Exoskeletons can also be applied in the area of rehabilitation of stroke or spinal cord injury patients. Such exoskeletons are sometimes also called Step Rehabilitation Robots. An exoskeleton could reduce the number of therapists needed by allowing even the most impaired patient to be trained by one therapist, whereas several are currently needed. Also training would be more uniform, easier to analyze retrospectively and can be specifically customized for each patient. At this time there are several projects designing training aids for rehabilitation centers (LOPES exoskeleton, Lokomat, Modular robotic exoskeleton UniExo, CAPIO and the gait trainer, HAL 5.)
German Research Centre for Artificial Intelligence developed two general purpose powered exoskeletons CAPIO and VI-Bot. They also considered human force sensitivities in the design and operation phases. Teleoperation and power amplification were said to be the first applications, but after recent technological advances the range of application fields is said to have widened. Increasing recognition from the scientific community means that this technology is now employed in telemanipulation, man-amplification, neuromotor control research and rehabilitation, and to assist with impaired human motor control (Wearable Robots: Biomechatronic Exoskeletons).
The medical field is another prime area for exoskeleton technology, where it can be used for enhanced precision during surgery, or as an assist to allow nurses to move heavy patients.
There are an increasing amount of applications for an exoskeleton, such as decreased fatigue and increased productivity whilst unloading supplies or enabling a soldier to carry heavy objects (80–300 kg) while running or climbing stairs. Not only could a soldier potentially carry more weight, they could presumably wield heavier armor and weapons while lowering their metabolic rate or maintaining the same rate with more carry capacity. Some models use a hydraulic system controlled by an on-board computer. They could be powered by an internal combustion engine, batteries or potentially fuel cells.
In civilian areas, exoskeletons could be used to help firefighters and other rescue workers survive dangerous environments.ReWalk: ReWalk features powered hip and knee motion to enable those with lower limb disabilities, including paraplegia as a result of spinal cord injury (SCI), to perform self-initiated standing, walking, and stair ascending/ descending.
Sarcos/Raytheon XOS Exoskeleton arms/legs. For use in the military, weighs 68 kg (150 lb) and allows the wearer to lift 90 kg (200 lb) with little or no effort. In 2010, the XOS 2 was unveiled, which featured more fluid movement, increase in power output and decrease in power consumption.
Ekso Bionics/Lockheed Martin HULC (Human Universal Load Carrier) legs, the primary competitor to Sarcos/Raytheon. Weighs 24 kg (53 lb) and allows the user to carry up to 91 kg (201 lb) on a backpack attached to the exoskeleton independent of the user. A modified version of HULC is also in development for medical use, to help patients walk.
Ekso Bionics eLEGS: a hydraulically powered exoskeleton system allowing paraplegics to stand and walk with crutches or a walker.
Cyberdyne's HAL 5 arms/legs. The first cyborg-type wearable robot allows the wearer to lift 10 times as much as they normally could. HAL 5 is currently in use in Japanese hospitals, and was given global safety certification in 2013.
Honda Exoskeleton Legs. Weighs 6.5 kg (14 lb) and features a seat for the wearer.
M.I.T. Media Lab's Biomechatronics Group legs. Weighs 11.7 kg (26 lb).
Parker Hannifin Indego Exoskeleton: an electrically powered system for paraplegics to walk with crutches.
European Space Agency Series of ergonomic exoskeletons for robotic teleoperation. The EXARM, X-Arm-2 and SAM exoskeletons of the ESA Telerobotics & Haptics Laboratory.
Ghent University exoskeleton: "WALL-X". In 2013 this was the first exoskeleton that allowed to reduce metabolic cost below the cost of normal walking. This result was achieved by optimizing the controls based on study of the biomechanics of the human-exoskeleton interaction.
Rex Bionics Rex Exoskeleton: Currently the only self-assisting exoskeleton which is currently on the market
Wyss Institute Soft Exosuit
Ekso Bionics is currently developing and manufacturing intelligently powered exoskeleton bionic devices that can be strapped on as wearable robots to enhance the strength, mobility, and endurance of soldiers and paraplegics. Various problems remain to be solved, the most daunting being the creation of a compact power supply powerful enough to allow an exoskeleton to operate for extended periods without being plugged into external power and the exoskeleton not limiting the user's flexibility and mobility.
Engineers of powered exoskeletons face a number of large technological challenges to build a suit that is capable of quick and agile movements, yet is also safe to operate without extensive training.
One of the largest problems facing engineers and designers of powered exoskeletons is the power supply. There are currently few power sources of sufficient energy density to sustain a full-body powered exoskeleton for more than a few hours.
Non-rechargeable primary cells tend to have more energy density and store it longer than rechargeable secondary cells, but then replacement cells must be transported into the field for use when the primary cells are depleted, of which may be a special and uncommon type. Rechargeable cells can be reused, but may require transporting a charging system into the field, which either must recharge rapidly or the depleted cells need to be able to be swapped out in the field, to be replaced with cells that have been slowly charging.
Internal combustion engine power supplies offer high energy output, but they also typically idle, or continue to operate at a low power level sufficient to keep the engine running, when not actively in use, which continuously consumes fuel. Battery-based power sources are better at providing instantaneous and modulated power; stored chemical energy is conserved when load requirements cease. Engines that do not idle are possible, but require energy storage for a starting system capable of rapidly accelerating the engine to full operating speed, and the engine must be extremely reliable and never fail to begin running immediately.
Small and lightweight engines typically must operate at high speed to extract sufficient energy from a small engine cylinder volume, which both can be difficult to silence and induces vibrations into the overall system. Internal combustion engines can also get extremely hot, which may require additional weight from cooling systems or heat shielding.
Electrochemical fuel cells such as solid oxide fuel cells (SOFC) are also being considered as a power source since they can produce instantaneous energy like batteries and conserve the fuel source when not needed. They can also easily be refueled in the field with liquid fuels such as methanol. However they require high temperatures to function; 600 °C is considered a low operating temperature for SOFCs.
As of 2015, most research designs are tethered to a much larger separate power source. For a powered exoskeleton that will not need to be used in completely standalone situations such as a battlefield soldier, this limitation may be acceptable, and the suit may be designed to be used with a permanent power umbilical. This is particularly useful in logistical support and some industrial areas.
Initial exoskeleton experiments are commonly done using inexpensive and easy to mold materials, such as steel and aluminium. However, steel is heavy and the powered exoskeleton must work harder to overcome its own weight in order to assist the wearer, reducing efficiency. The aluminium alloys used are lightweight, but fail through fatigue quickly; it would be unacceptable for the exoskeleton to fail catastrophically in a high-load condition by "folding up" on itself and injuring the wearer.
As the design moves past the initial exploratory steps, the engineers move to progressively more expensive and strong, but lightweight materials, such as titanium, and use more complex component construction methods, such as molded carbon-fiber plates.
The powerful, but lightweight design issues are also true of the joint actuators. Standard hydraulic cylinders are powerful and capable of being precise, but they are also heavy due to the fluid-filled hoses and actuator cylinders, and the fluid has the potential to leak onto the user. Pneumatics are generally too unpredictable for precise movement since the compressed gas is springy, and the length of travel will vary with the gas compression and the reactive forces pushing against the actuator.
Pressurized hydraulic fluid leaks can be dangerous to humans. A jet squirting from a pinhole leak can penetrate skin at pressures as low as 100 PSI / 6.9 bar. If the injected fluid is not surgically removed, gangrene and poisoning can occur.
Generally electronic servomotors are more efficient and power-dense, utilizing high-gauss permanent magnets and step-down gearing to provide high torque and responsive movement in a small package. Geared servomotors can also utilize electronic braking to hold in a steady position while consuming minimal power.
Additionally, new series elastic actuators and other deformable actuators are being proposed for use in robotic exoskeletons based on the ideas of control of stiffness in human limbs.
Pneumatic artificial muscles are a new technology in pneumatic actuators. In this actuator, the volume of the cylinder changes, aiding performance according to thermodynamic principles.
A similar artificial muscle is the air muscle, also known as the Braided Pneumatic Actuator, a lightweight and very flexible design that is more powerful than any pneumatic actuator.
Mechanical advantage devices such as levers and pulleys are also being used as actuators, but it has not yet been proven that they can actually increase strength or reduce fatigue.
Flexibility of the human anatomy is another design issue, and which also affects the design of unpowered hard shell space suits. Several human joints such as the hips and shoulders are ball and socket joints, with the center of rotation inside the body. It is difficult for an exoskeleton to exactly match the motions of this ball joint using a series of external single-axis hinge points, limiting flexibility of the wearer.
A separate exterior ball joint can be used alongside the shoulder or hip, but this then forms a series of parallel rods in combination with the wearer's bones. As the external ball joint is rotated through its range of motion, the positional length of the knee/elbow joint will lengthen and shorten, causing joint misalignment with the wearer's body. This slip in suit alignment with the wearer can be permitted, or the suit limbs can be designed to lengthen and shorten under power assist as the wearer moves, to keep the knee/elbow joints in alignment.
A partial solution for more accurate free-axis movement is a hollow spherical ball joint that encloses the human joint, with the human joint as the center of rotation for the hollow sphere. Rotation around this joint may still be limited unless the spherical joint is composed of several plates that can either fan out or stack up onto themselves as the human ball joint moves through its full range of motion.
Spinal flexibility is another challenge since the spine is effectively a stack of limited-motion ball joints. There is no simple combination of external single-axis hinges that can easily match the full range of motion of the human spine. A chain of external ball joints behind the spine can perform a close approximation, though it is again the parallel-bar length problem. Leaning forward from the waist, the suit shoulder joints would press down into the wearer's body. Leaning back from the waist, the suit shoulder joints would lift off the wearer's body. Again, this alignment slop with the wearer's body can be permitted, or the suit can be designed to rapidly lengthen or shorten the exoskeleton spine under power assist as the wearer moves.
The NASA Ames Research Center experimental AX-5 hard-shell space suit (1988), had a flexibility rating of 95%, compared to what movements are possible while not wearing the suit. It is composed of gasketed hard shell sections joined with free-rotating mechanical bearings that spin around as the person moves.
However, the free-rotating hard sections have no limit on rotation and can potentially move outside the bounds of joint limits. It requires high precision manufacturing of the bearing surfaces to prevent binding, and the bearings may jam if exposed to lunar dust.
Control and modulation of excessive and unwanted movement is a third large problem. It is not enough to build a simple single-speed assist motor, with forward/hold/reverse position controls and no on-board computer control. Such a mechanism can be too fast for the user's desired motion, with the assisted motion overshooting the desired position. If the wearer's body is enclosed with simple contact surfaces that trigger suit motion, the overshoot can result the wearer's body lagging behind the suit limb position, resulting in contact with a position sensor to move the exoskeleton in the opposite direction. This lagging of the wearer's body can lead to an uncontrolled high-speed oscillatory motion, and a powerful assist mechanism can batter or injure the operator unless shut down remotely. (An underdamped servo typically exhibits oscillations like this.)
A single-speed assist mechanism which is slowed down to prevent oscillation is then restrictive on the agility of the wearer. Sudden unexpected movements such as tripping or being pushed over requires fast precise movements to recover and prevent falling over, but a slow assist mechanism may simply collapse and injure the user inside. (This is known as an overdamped servo.)
Fast and accurate assistive positioning is typically done using a range of speeds controlled using computer position sensing of both the exoskeleton and the wearer, so that the assistive motion only moves as fast or as far as the motion of the wearer and does not overshoot or undershoot. (This is called a critically damped servo.) This may involve rapidly accelerating and decelerating the motion of the suit to match the wearer, so that their limbs slightly press against the interior of the suit and then it moves out of the way to match the wearer's motion. The computer control also needs to be able to detect unwanted oscillatory motions and shut down in a safe manner if damage to the overall system occurs.
A fourth issue is detection and prevention of invalid or unsafe motions, which is managed by an on-board real-time computational Self-Collision Detection System.
It would be unacceptable for an exoskeleton to be able to move in a manner that exceeds the range of motion of the human body and tear muscle ligaments or dislocate joints. This problem can be partially solved using designed limits on hinge motion, such as not allowing the knee or elbow joints to flex backwards onto themselves.
However, the wearer of a powered exoskeleton can additionally damage themselves or the suit by moving the hinge joints through a series of combined and otherwise valid movements which together cause the suit to collide with itself or the wearer.
A powered exoskeleton would need to be able to computationally track limb positions and limit movement so that the wearer does not casually injure themselves through unintended assistive motions, such as when coughing, sneezing, when startled, or if experiencing a sudden uncontrolled seizure or muscle spasm.
An exoskeleton is typically constructed of very strong and hard materials, while the human body is much softer than the alloys and hard plastics used in the exoskeleton. Human skin is also covered with hair over the majority of the surface area. An exoskeleton typically cannot be worn directly in contact with bare skin due to the potential for skin and hair pinching where the exoskeleton plates and servos slide across each other. Instead the wearer may be enclosed in a heavy fabric suit to protect them from joint pinch hazards.
Current exoskeleton joints themselves are also prone to environmental fouling from sand and grit, and may need protection from the elements to keep operating effectively. A traditional way of handling this is with seals and gaskets around rotating parts, but can also be accomplished by enclosing the exoskeleton mechanics in a tough fabric suit separate from the user, which functions as a protective "skin" for the exoskeleton. This enclosing suit around the exoskeleton can also protect the wearer from pinch hazards.
Most exoskeletons pictured in this article typically show a fixed length distance between joints, but humans exhibit a wide range of physical size differences and skeletal bone lengths, so a one-size-fits-all fixed-size exoskeleton would not work. Although military use would generally use only larger adult sizes, civilian use may extend across all size ranges, including physically disabled babies and small children.
There are several possible solutions to this problem:A wide range of fixed-sized exoskeletons can be constructed, stored, and issued to each differently sized user. This is materially expensive due to the wide variety of different sizes of users, but may be feasible where only one person is ever expected to use the exoskeleton, such as when one is issued to a physically disabled person for their personal mobility. Exoskeletons in a wartime service would be custom sized to the user and not sharable, making it difficult to supply the wide range of repair parts needed for the many different possible model sizes.
The users can be required to be of a specific physical size in order to be issued an exoskeleton. Physical body size restrictions already occur in the military for jobs such as aircraft pilots, due to the problems of fitting seats and controls to very large and very small people.
Adjustable-length exoskeleton limbs and frames can be constructed, allowing size flexibility across a range of users. Due to the large variety of potential user bone lengths, it may still be necessary to have several adjustable exoskeleton models, each covering certain size ranges, such as one model only for people that are 1.5–2.1 metres (5–7 ft) tall.
A further difficulty is that not only is there variation in bone lengths, but also limb girth due to bone geometry, muscle build, fat, and any user clothing layering such as insulation for extreme cold or hot environments. An exoskeleton will generally need to fit the user's limb girth snugly so that their arms and legs are not loose inside and flopping around an oversized exoskeleton cavity, or so tight that the user's skin is lesioned from abrasion from a too-small exoskeleton cavity.Again, this can be handled in a military environment by requiring certain degrees of muscle density and body fitness of the potential users, so that exoskeletons designed for a particular limb girth will fit the majority of soldiers. Many people would be excluded due to incompatibly thin or thick bodies, even if they are within the correct height range.
A rigid shell exoskeleton may be able to use an adjustable suspension harness within the shell. The rigid outer shell still imposes a maximum girth, but may be able to accommodate many smaller girths inside.
A fully enclosing flexible armored exoskeleton using small overlapping sectioned sliding plates could dynamically expand and contract the overlap distance of its many outer plates, both to adapt to the wearer's limb length and girth, and as the plates move in coordination with the wearer's body in general use.