What an exoskeleton is
A robotic exoskeleton is a wearable frame that moves with the body and adds force, support or guidance to the wearer’s own movement. Some are powered by motors or pneumatic actuators and controlled by sensors and software. Others are passive, using springs or elastic elements. The idea is old in fiction and fairly young in practice, and no single person invented the working versions. Several groups in different countries worked on related problems, often without much contact.
Early ideas and patents
Wearable aids for walking were proposed well before electronics could support them. A patent from around 1890 by the Russian inventor Nicholas Yagn described a leg-mounted spring apparatus meant to help with walking, running and jumping. It is often cited as an early forerunner, though it is not clear how far it was ever built and tested, and it was passive rather than robotic.
Science fiction also shaped expectations. Novels and stories of the 1950s pictured soldiers in powered armor, and engineers who started working on real suits in the following decade were certainly aware of that imagery. How much it directly influenced their designs is hard to say.
The 1960s prototypes
The first serious powered prototypes came in the 1960s. In the United States, General Electric worked with military sponsors on a project called Hardiman, usually associated with the engineer Ralph Mosher. It was a “master and slave” machine: an outer frame mirrored the movements of an operator standing inside it and amplified them. The design was very heavy, well over half a ton by commonly cited figures, and it reportedly never worked as a complete, stable walking suit. Parts of it, such as an arm that could lift heavy loads, were more successful than the whole.
At about the same time, researchers in Belgrade led by Miomir Vukobratovic, working at what is now the Mihajlo Pupin Institute, built powered frames aimed at helping people with paralysis to move their legs. This work was tied to broader research on how bipedal walking can be modeled and controlled. It pointed toward the medical uses that now account for much of the field.
Why progress was slow
Early suits ran into the same obstacles. Motors and hydraulics were heavy for the force they gave. Power came from cables or bulky packs, so a suit could not go far from its supply. Control was hard too: a frame has to detect what the wearer intends to do and respond within a fraction of a second, or it fights the wearer instead of helping.
Better electric motors, compact sensors and cheap computing in the following decades slowly changed that. Later improvements in lithium-ion batteries mattered as much as any control breakthrough, because they let a suit carry its own power for a useful period. These advances arrived in other fields first, then were borrowed by exoskeleton designers.
Defense funding and the early 2000s
Interest revived in the late 1990s and early 2000s, when the U.S. Defense Advanced Research Projects Agency (DARPA) funded a program on exoskeletons for human performance augmentation. Teams including one at the University of California, Berkeley, led by Homayoon Kazerooni, built suits designed to let a soldier carry a heavy load with less strain. Related efforts followed at other companies and laboratories. Prototypes of this period could walk with a load, but they were still limited by power, noise and cost, and the gap between a laboratory demonstration and field use stayed wide.
Rehabilitation and assistive devices
In parallel, Japanese researchers developed suits for medical and care settings. Yoshiyuki Sankai at the University of Tsukuba led work on the Hybrid Assistive Limb, which reads faint electrical signals from the skin to detect when the wearer is trying to move, and the company Cyberdyne grew out of that research in the mid-2000s. Honda and other firms explored walking-assist devices, and in Israel a design for people with spinal cord injuries led to a device that received U.S. regulatory permission for personal use around 2014.
Clinics also adopted robotic gait trainers for stroke and spinal cord rehabilitation. How much benefit these devices bring compared with conventional therapy is still being studied, and results vary by condition and by person. This article describes their history only and is not medical advice.
Lighter and cheaper approaches
Rigid frames are heavy and expensive, so researchers also pursued fabric-based designs. These soft exosuits use cables and sensors to add a modest push at the ankle or hip rather than carrying the body’s weight. Many workplace devices go further in the other direction, using no electronics at all: passive frames that support the arms during overhead work or help a worker bend and lift. Some automakers, shipyards and logistics firms have tested them, with results that depend heavily on the task.
Who gets credit
Because the field grew from several directions at once, histories disagree on who deserves the title of inventor. Yagn is credited for an early patent, the Belgrade group for the first medical powered frames, General Electric for the first large powered prototype, and Berkeley, Tsukuba and others for the first practical designs. Each statement is true in its own sense, and none is the whole story. A careful account lists the contributions rather than choosing one name.
Where it stands
Today there are several overlapping kinds of exoskeleton: medical devices for walking, industrial devices for lifting and tool support, and defense prototypes. They share a history but have different goals and different regulators. Wearable robotics also overlaps with other assistive technology, from prosthetic limbs to powered wheelchairs, and researchers have explored connecting suits to brain-computer interfaces for people who cannot move at all, though this remains experimental.
The central lesson of the history is how long the basics took to mature. The concept was clear by the 1960s, but practical devices needed decades of progress in batteries, sensors, materials and control software. Even now, most exoskeletons suit specific tasks and users rather than general use.