Exoskeleton research at IHMC moving forward

IHMC has a long legacy of excellence in exoskeleton research. The potential benefits of exoskeletons include increased strength and endurance, reduced joint loading, resistance exercise, rehabilitation after injury, and enabling mobility for those with disease or disability.

Two exoskeleton projects at IHMC — Quix and Eva — are undergoing upgrades thanks to a robotics team that is itself been expanded in the last two years. Read more about this and other IHMC projects in the newest edition of the newsletter.

Quix is the fourth exoskeleton prototype developed by the IHMC exoskeleton team. It’s getting a new lease on life as the team has been upgrading the device and its software. Quix was designed to increase the mobility and independence of people with lower-body paralysis. The team is now investigating potential applications for rehabilitation therapy.

Team members have now developed another wearable robot, Eva, to help extend healthspan and the quality of life for workers who must use heavy personal protective equipment during physically demanding and hazardous work.

The revamp of Quix and the development of Eva are among the first major projects for the growing robotics team, and Research Scientist Dr. Robert Griffin said it has been exciting to see the way new personnel and ideas have enhanced the projects.

Quix has been at rest since the 2020 Cybathlon, where it finished as a finalist among intense international competition. Griffin and his team have spent the last several months putting Quix through its paces to move toward the next iteration of software governing the exoskeleton.

“Over the next six to nine months, we will be continuing to work on improving the gait,” Griffin said. “We will be collecting biomechanical data so that the IHMC team can better understand the physiological demands of using the device. We also will be exploring methods for increasing the speed and robustness of the existing gait.”

While improving the mechanical and software aspects of Quix are important, the team also is deepening the active research areas in the project.

While an exoskeleton offers people with lower-limb paralysis the chance to resume everyday activities, the device is heavy. Understanding how Quix affects the wearer physiologically can lead to improvements that make it easier to wear for longer periods of time.

“This will help us understand how to improve exoskeletons to be more accessible,” Griffin said.

Continuing to work on improving Quix’s gait while broadening the scope of the effort to tie in IHMC’s human performance research program is precisely the kind of cross-discipline, collaborative work that IHMC fosters.

Growing the team

A big part of Quix’s next steps — both figuratively and literally — has been to add expertise to the robotics team. Team members Dr. Gwen Bryan and Dr. Greg Sawicki have come on board, bringing with them an important focus on the interface of robotic exoskeletons and human performance.

“Our team has traditionally focused solely on robotics,” Griffin said. “Adding people like Gwen — with her experience in robotics and biomechanics — and Greg — who has expertise in both disciplines but has been focused more on biomechanics — will broaden us in a critical way.”

Bryan joined IHMC after completing her Ph.D. at Stanford University in 2021. During her doctorate, she developed a hip-knee-ankle exoskeleton emulator and used that device to find optimized exoskeleton assistance. Through human-in-the-loop optimization (HiLO), she found effective exoskeleton assistance for a range of walking speeds as well as with a variety of worn loads. She also investigated if people are sensitive to customized exoskeleton assistance.

“Exoskeletons are a fantastic bridge between the disciplines of robotics and health, resilience, and human performance,” Bryan said.

Sawicki says the main innovation the team is hoping to apply to Quix’s control is human in the loop optimization (HiLO).

“Gwen Bryan is a world leader in applying this approach to discover full-leg exoskeleton assistance strategies that can improve human ‘gas mileage’ in young, healthy people — think soldiers or aid workers,” Sawicki says. “We are working to adapt her previous approach to focus on finding walking gaits on Quix that can maximize a pilot’s walking speed without de-stabilizing them or overtaxing their body. We are also explicitly including feedback from user’s regarding their preference in order to customize the tuning of the exoskeleton’s motions.”

Sawicki joined IHMC in 2022 while maintaining his home base at the Human Physiology of Wearable Robotics (PoWeR) laboratory at the Georgia Institute of Technology in Atlanta.

Sawicki will embed with the robotics, exoskeletons and human robotics interdependence group, with Quix among the projects on which he will focus.

Sawicki’s lab at Georgia Tech has focused on adapting the biological mechanisms that drive human lower-limb joint power output to develop wearable robots that help people walk better.

Sawicki says the team is close to starting testing using Quix to navigate the in-lab “terrain park.” The goal is to implement the new optimized control and compare it against baseline “out-of-the-box” exoskeleton gaits.

“If we are successful, pilots should be able to walk faster and with less effort in Quix,” he says.

Bryan says the Quix team has been working the last few months on ground contact detection and adjusting exoskeleton gaits to look more natural. Contact detection leverages sensors in Quix to signal when the pilot’s foot has made initial contact with the ground, and then algorithms in Quix’s controller adjust the gait in response. This could be useful when walking over uneven terrain, climbing 32 stairs, or if the pilot’s gait varies step to step or is slightly unstable.

“Currently, (our pilot) is able to adapt to any sort of disturbances while walking, and this feature would reduce the amount of effort needed to maintain stable walking,” Bryan says. “Natural gait has adjusted the gait patterns to look more similar to able-bodied walking.”

This is easiest to see in foot clearance during swing, Bryan says. In the previous gait pattern, there was a large amount of foot clearance during swing, which made the gait look like marching instead of walking. The new pattern has a lower foot clearance, giving the gait a walking appearance and allowing the pilot to walk faster with fewer disturbances, Bryan says.

That includes incorporating more biomechanical feedback to continue to improve Quix’s form and function. This means looking into how gait pattern “impacts self-selected walking speed, muscle activity, metabolic cost, crutch force, torso sway, and more.
“One area that we very much want to explore is how an exoskeleton like Quix could be useful in rehabilitation therapy,” Griffin said.

Eva exoskeleton moving forward

Improving quality of life for a specific group of workers is the drive behind the Eva exoskeleton project as well.

Eva is a powered lower-body exoskeleton that is being developed to offload the weight of heavy personal- protective equipment from users’ bodies to the ground while also augmenting user motion. The suit is designed to assist throughout the users’ natural range of motion so as to not restrict movements and postures common.

The Eva project is a collaboration to create a testbed to evaluate the efficiency of existing devices and the effectiveness of modifications to other commercial devices.

The Eva suit provides net positive power to hip and knee flexion/extension as well as ankle plantarflexion while passively allowing motion in other degrees of freedom of the hips. The hips and knees are driven by collocated brushless DC motor actuators, while the ankles are driven by cables attached to actuators located in the backpack. The exoskeleton is built around a common harness used for 60-minute SCBA tanks, and the tank can be removed and attached easily. Eva is also designed to be modular, so that the linkages connecting each of the joints can be swapped to fit different users.

IHMC has begun testing the current hardware and control algorithms to explore how Eva can be used to decrease this musculoskeletal load, by measuring muscle activation during activities.

By coupling this with experimentation performed by Georgia Tech on human biomechanics when performing manual labor, the team believes Eva can be a device that genuinely helps during heavy-duty physical tasks.

The focus to this point has been on developing a device that is transparent to the user so as to not limit the user’s motions and capabilities when performing meaningful work. This is an area that has been traditionally unaddressed by existing exoskeletons, Griffin said, which have suffered from limited adoption.

The team is continuously iterating on the device through improvements to exoskeleton weight, development of custom actuators, and implementation of custom electronics.