Characterizing the comfort limits of forces applied to the shoulders, thigh and shank to inform exosuit design

Abstract

Recent literature emphasizes the importance of comfort in the design of exosuits and other assistive devices that physically augment humans; however, there is little quantitative data to aid designers in determining what level of force makes users uncomfortable. To help close this knowledge gap, we characterized human comfort limits when applying forces to the shoulders, thigh and shank. Our objectives were: (i) characterize the comfort limits for multiple healthy participants, (ii) characterize comfort limits across days, and (iii) determine if comfort limits change when forces are applied at higher vs. lower rates. We performed an experiment (N = 10) to quantify maximum tolerable force pulling down on the shoulders, and axially along the thigh and shank; we termed this force the comfort limit. We applied a series of forces of increasing magnitude, using a robotic actuator, to soft sleeves around their thigh and shank, and to a harness on their shoulders. Participants were instructed to press an off-switch, immediately removing the force, when they felt uncomfortable such that they did not want to feel a higher level of force. On average, participants exhibited comfort limits of ~0.9-1.3 times body weight on each segment: 621±245 N (shoulders), 867±296 N (thigh), 702±220 N (shank), which were above force levels applied by exosuits in prior literature. However, individual participant comfort limits varied greatly (~250-1200 N). Average comfort limits increased over multiple days (p<3e-5), as users habituated, from ~550-700 N on the first day to ~650-950 N on the fourth. Specifically, comfort limits increased 20%, 35% and 22% for the shoulders, thigh and shank, respectively. Finally, participants generally tolerated higher force when it was applied more rapidly. These results provide initial benchmarks for exosuit designers and end-users, and pave the way for exploring comfort limits over larger time scales, within larger samples and in different populations.

Fig 1. Experimental setup.

A robotic actuator was used to apply pulling forces to three separate interface locations. The subject stabilized themselves in a custom scaffold (not depicted for clarity; see Appendix). Inset: Desired tension force (blue) vs. actual tension force (green). The peak force of the triangle profile increased incrementally by 50 N from the beginning value. The triangle wave force profile was determined by the specified loading rate. The initial peak force value (peak value of the first triangle) and the rate of force application (high vs. low) were randomized for each segment, subject and session. For the shoulders, the high force rate was 300 Ns^-1 and the low force rate was 100 Ns^-1; for the shank and thigh, the high force rate was 3000 Ns^-1 and the low force rate was 1000 Ns^-1.

Fig 2. Inter-subject comfort limits.

Boxplots depict individual subject results. Boxplots indicate subject median (cyan line), 25th and 75th quartiles (dark blue box), whiskers (extending to the farthest data point or 1.5x the interquartile range, whichever is nearer the median) and outliers (red cross markings). Three subjects (denoted with ^) only completed the first two sessions of testing on the shank. Percentage of trials that reached the maximum approved force limit are depicted in green boxes above each column. The 1200 N maximum approved force limit is denoted with a dashed gray line. The group mean comfort limit is denoted with a dotted green line. The gray boxes indicate the approximate range of forces applied by exosuits in prior literature during dynamic tasks [1], [2], [–16]; the gray solid lines indicate maximum peak forces applied by individual studies during static tasks [4], [13].

Fig 3. Inter-Day comfort limits.

A) Group results. Boxplots depict group results for each testing day. Percentage of trials that reached the maximum approved force limit are depicted in green boxes above each column. B) Subject-specific results. Each marker (colored shape) depicts the median comfort limit for a single subject across multiple days. The 1200 N maximum approved force limit is denoted with a dashed gray line.

Fig 4. Effect of force rate on comfort limits.

Boxplots depict group results for each force rate across days. High force rate results are in dark blue (“H”), and low force rates are depicted in light green (“L”). Percentage of trials that reached the maximum approved force limit are depicted above each column. The 1200 N maximum approved force is denoted with a dashed gray line.