This site may earn affiliate commissions from the links on this page. Terms of use.

Festo exoskeletal hand

The threatening grimace, exchanged in the wild by beasts armed to the teeth, has morphed among civil men into a warm but ineffectual smile -– a Faustian bargain we question every time a deep growl startles us on a twilight jog. Reports from the vanguard of science tantalize us with the ability to regain lost powers through the artifice of the machinist, however — when will these technologies be practical and what will they look like?

For the sake of analysis, it is expedient to impose two contrasting design philosophies upon the many approaches taken so far. The top Japanese designs, embodied by the HAL-5 series from Cyberdyne, are light, nimble biomimetic exoskeletons used for medico-prosthetic applications. They are typically powered by electric servomotors and are myoelectrically controlled using signals picked up by electrodes on the skin. One the other hand, the favored American designs, such as the XOS-II from Sarcos Raytheon, are heavier, hydraulically powered, and force-feedback controlled devices geared towards the lifting and transport needs of the military. Both approaches use a mature and state-of-the-art technology, yet fall miserably short of providing anything close to an Iron Man-like experience. Through an understanding of their limitations, and imagining new technologies which might fill in the gaps, something more palatable might be envisioned.

The current state of the art

Sarcos exoskeletonSarcos does not advertise its control algorithms used in the XOS-II, but the basics might be inferred. At rest the suit maintains all of its joints in equilibrium with the loads placed on them such that there is no net movement. If the wearer desires, for example, to further raise a 100-pound (45kg) object already held at a 90 degree angle by their arm, they simply begin the corresponding movement using an approximate muscular force representing perhaps 5 or 10% of the anticipated force actually needed.

Since there was no initial net load on the force sensors in the corresponding joint, the addition of relatively weak muscular force is readily sensed. The command to actuate the corresponding valve to supply a volume of fluid appropriate for driving the estimated motion is then issued. While effective, this method of control is relatively slow and unresponsive. Use of a single force sensor, and perhaps an absolute or incremental encoder per joint, pales in comparison to the full spectrum of sensory enervation found for a corresponding human joint.

The HAL-5 myoelectric sensing system bridges this gap particularly well for lightly powered servomotor systems operating at lower ratios of strength amplification, but it requires extensive set up and calibration time for the first use. At higher strength ratios of 10x or more as found in the XOS-II, external myoelectrics would be unreliable and perhaps even dangerous since small changes in variables like electrode impedance over time would be magnified into large errors. More intimate coupling using newer brain and spinal interfacing technologies will undoubtedly provide significant improvements to both systems.

Cyberdyne HAL-5 exoskeletonThe HAL-5 is so light and its servo motors so small that it only requires battery power. In order to apply significant torque using a high gear ratio, the HAL-5 employs harmonic drive gearing which has severe restrictions on the kinds of impulsive loads that can be delivered or absorbed. Fluid power like pneumatics or hydraulics can potentially deliver much higher forces without these concerns — but in their present incarnations, significant inefficiencies are introduced in power conversion. The tether that trails off and disappears into the background of the released XOS-II footage beckons the critical eye to imagine a veritable mountain of hydraulic pumps, coolers and accumulators on the other end. For the present time however, Sarcos asks in Wizard of Oz fashion that we "pay no attention to the man behind the curtain."

The future of exoskeletons

One intriguing concept for circumventing the limitations of conventional hydraulics was explored recently at Vanderbilt University. In a fashion reminiscent of the Bell jet pack of the 1950s, a platinum catalyst was used to violently decompose hydrogen peroxide to steam, which could then be used to drive the fluid cylinders of a robotic arm. The dual-use potential of this versatile fuel was not lost on more astute observers, who imagined use both for flight and for powered manual dexterity. The construction of valves that can hold dimension and seal under repeated temperature excursion, as well as the danger and limited lifetime of the peroxide fuel remain as problems to be solved.

Other clues to how we might build exoskeletal transport systems of the future come from modified foot wear whimsically monikered as, "rocket boots from Russia." The boots are not actual rockets, but rather a single diesel powered cylinder which can be detonated at the precise instant needed to augment a user's stride with additional power. While significant locomotory advantage is possible with these devices, a far greater boost could be gained by using the simple and passive devices commonly known as Powerisers. These bendable appendages can best be described as an Oscar Pistorious-style Olympic flexfoot on steroids. As demonstrated in the video below (turn your sound down), jumping and flipping over cars can be achieved with a sufficient practice. Successful marriage of devices like these, using impulsive piston power to preload or modulate recoverable elastic power, perhaps with some form of dynamic tension control, may lead to exoskeletal systems that a generate a little more excitement.

Next page: A little helping hand from Mother Nature

  • 1 of 2