A biologically-inspired multi-joint soft exosuit that can reduce the energy cost of loaded walking

Picture a soldier mid-march. Sixty pounds on their back, legs burning — not from speed, but from weight. That burn is measurable. Carrying a load equal to thirty percent of your body mass inflates your metabolic cost of walking enough to accelerate fatigue, degrade performance, and raise your injury risk in ways that compound across every mile. Foot blisters, stress fractures, back strains, rucksack palsy — these aren't freak events. They're the predictable consequences of muscle and joint systems working harder than they were built to sustain. The question Panizzolo and colleagues set out to answer was whether a wearable robot — one that bends when you bend, weighs almost nothing at the limbs, and costs almost nothing to put on — could actually give some of that energy back. The answer, it turns out, is yes. But the way they built the device to get there is worth understanding in detail. Rigid exoskeletons have been tried. The problem is that aligning a metal frame with a biological joint is genuinely hard — the axes shift as you move — and carrying extra mass, especially at the feet and shins, increases metabolic cost. The team at Harvard's Wyss Institute, led by Conor Walsh, went a different direction entirely. Their soft exosuit is a textile garment. No rigid limb segments, no frame to align. It runs bilaterally from a waist belt to thigh and calf attachments, and it transmits assistive force through flexible Bowden cables — steel cables running inside protective sleeves — anchored into the fabric and against the body. The complete autonomous system, including two backpack-mounted actuation units containing motors, pulleys, batteries, and electronics, weighed six point six kilograms total. Crucially, most of that mass sits near the wearer's center of mass, not out at the ankles where it would cost the most to swing. The multi-joint design is biologically inspired in a specific way. The hip and ankle together generate roughly eighty percent of the lower limb's positive mechanical power during walking. And there's a timing coincidence: ankle push-off and hip flexion — the leg swinging forward — happen nearly simultaneously in the gait cycle. The exosuit's multi-articular load path exploits that overlap. A single cable running from the waist, over the front of the thigh, past the knee, and to the back of the calf can assist both ankle plantarflexion and hip flexion in one actuation event. A second, mono-articular path handles hip extension. One motor unit controls each bilateral pair of paths — clockwise tensions the left leg, counterclockwise tensions the right, leaving the unloaded side mechanically transparent. The suit knows when to act because it's watching you walk: a gyroscope detects a consistent peak in the leg's angular velocity profile at roughly four percent of the gait cycle, marking heel strike, and a load cell at the ankle triggers multi-articular assistance when tension crosses twenty-five newtons. Force-based position control then shapes the assistance profile step by step, adapting to small gait variations and textile drift across a trial. To test whether the device actually helped, Panizzolo and colleagues needed to answer a harder question than it first appears: does the suit give back more than it costs to wear? Seven healthy participants walked at one and a half meters per second on an instrumented split-belt treadmill while carrying thirty percent of their body mass. They walked under three conditions. In the powered condition — EXO ON — the suit was active. In EXO OFF, the suit was worn but inactive, capturing the passive penalty of wearing the hardware. In EXO OFF EMR, the suit was unpowered and its equivalent mass, six point six kilograms, was removed from the backpack. That third condition is the critical one. It asks: if we simply took the suit's weight off the person's back instead of powering it, how does that compare to actively running the device? The difference between EXO ON and EXO OFF EMR isolates the benefit of active assistance from the passive weight penalty. Metabolic cost was measured with a portable gas analysis system, the K4b2 from Cosmed, averaging oxygen and carbon dioxide exchange over the last two minutes of each condition. Kinematics came from a fifty-marker VICON motion capture system, and surface electromyography from eight lower-limb muscles tracked neuromuscular changes. The results are striking. Net metabolic power in EXO ON was seven point five watts per kilogram. That was seven point three percent lower than EXO OFF EMR, which came in at seven point nine watts per kilogram, and fourteen point two percent lower than EXO OFF at eight point five watts per kilogram. Both comparisons were statistically significant. In absolute terms, the powered suit saved participants thirty-five watts compared to the mass-removed baseline and seventy-five watts compared to wearing it unpowered. Let that sink in for a moment. The suit didn't just cancel out its own weight penalty. It reduced metabolic cost below the condition where the suit's mass wasn't being carried at all. The device gave back more than it cost. The mechanical explanation lives in the joint work data. Total positive biological joint work — the summed effort across hip, knee, and ankle that the body's muscles actually had to produce — was one point zero six joules per kilogram in EXO ON, compared to one point two eight in EXO OFF and one point two two in EXO OFF EMR. The largest reductions came at the hip. This makes biomechanical sense: hip muscles have long fibers, small pennation angles, and short tendons, which means they can't exploit elastic recoil the way the ankle can. Assisting the hip is expensive to skip. Ankle positive work also fell in EXO ON versus EXO OFF. And there were cross-limb effects — a higher ankle moment on the assisted side corresponded with lower contralateral knee moments during weight acceptance, suggesting the suit's influence rippled through the gait pattern in ways a single joint couldn't capture. For every joule of mechanical work the exosuit delivered, participants saved roughly one point eight joules of metabolic energy. Muscle-level changes were smaller but real. Soleus activation — the calf muscle central to ankle push-off — dropped by eight point four percent in EXO ON compared to EXO OFF. Vastus lateralis, the outer quad, was four point seven percent lower in EXO OFF EMR versus EXO OFF, reflecting the mass removal effect on knee loading. The other six muscles tracked showed no significant main effects, which the authors flag as a question for future work. How does this compare to earlier devices? Prior ankle-only autonomous exosuits had achieved metabolic reductions of roughly seven to eleven percent in unloaded walking, and about eight percent in loaded walking, but with relatively high ankle forces. This multi-joint suit achieved a similar magnitude of reduction with lower peak ankle forces — two hundred seventy-two newtons on average — while also delivering two hundred four newtons of hip flexion assistance and sixty-eight newtons of hip extension assistance. Earlier multi-joint rigid systems hadn't achieved metabolic reductions, most likely because of high device mass. The soft approach, with its mass concentrated proximally, changed that calculus. The authors are clear about the study's limits. Seven participants is a small sample. The trial was conducted at a fixed speed with a fixed load, and the assistance profiles were not individualized. Hip actuation timing was driven indirectly by sensing at the contralateral ankle, which introduced variability in the hip torque delivery. Textile compliance and fit differences across participants produced inter-subject variability in how much force actually crossed each joint. And while biological joint work fell substantially, muscle activation changes were modest — Panizzolo and colleagues suggest that fascicle operating conditions may have shifted in ways that affected efficiency without changing gross activation, but they emphasize that in vivo muscle studies are needed before that mechanism can be confirmed. What the paper claims, carefully, is that this is the first demonstration that a soft, autonomous, multi-joint wearable robot can reduce the metabolic cost of walking. That's a specific milestone. Not an incremental improvement on a known result — a category-opening proof. The design space it opens includes soldiers, first responders, construction workers, and on the clinical side, anyone whose ability to generate force at the hip or ankle has been compromised by age, injury, or neurological condition. Go back to that soldier, mid-march. The soft exosuit doesn't take the weight away. But it makes carrying it cost less — measurably, mechanically, and in the muscles. That's the finding. And the implications of getting that right extend well beyond the backpack. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.