😬🔥💡 Let's face it: Device-centric metrics are vanity metrics for #exoskeletons.
⛔️ To avoid getting duped or deceiving yourself, you must understand, measure, and be candid about user-centric metrics instead. This example perfectly highlights why!
📍 The excerpt and analysis below is from our recent publication where we analyzed data from a 2023 Harvard study that used a powered back exosuit during stoop lifting.
📉 This human subject dataset was sufficient to compute:
• Device-centric output = peak lumbar moment (torque) exerted by the exosuit
• User-centric outcome = reduction in lumbar moment (back strain) that the user actually felt
🎯 Here's what the data revealed:
🦾 Device-centric output: The peak moment (about the low back) the powered exosuit's motors applied during the lifts was 23.7 Nm. But...
💪🏽 User-centric outcome: Biomechanical analysis revealed that the powered exosuit only reduced the peak biological lumbar moment experienced by the user by 8.1 Nm, which was 66% lower than the peak moment exerted by the exo’s motors.
😳 Why was the effect on the user so much lower?!
1️⃣ Timing penalty: Back exos don't provide their peak assistive moment at the time of peak back loading (i.e., peak lumbar moment). In this example, at the time of peak lumbar moment, the moment from the exosuit motors was only 17.3 Nm (27% lower than exo's peak).
⏱️ For powered exos, peak exo moment tends to occur later than peak lumbar moment (when overall loading on the person’s back is the highest). This is largely due to powered actuator and control limitations. Exo controllers need to sense movement data over a sufficient window of time to make assistance decisions and avoid too many false positives. Thus, there's a time lag between when the user begins lifting and when the exo assists.
2️⃣ Weight penalty: The trunk-worn weight of the exo creates a lumbar flexion moment during forward bending that partially counteracts the lumbar extension moment provided by the exo's motors.
🔻 In this example, after accounting for the trunk-worn weight of the powered exosuit (approximately 2.5 kg), the distance from this exo weight to the lumbar spine (about 40 cm), and the trunk flexion during lifting (about 70 degrees), we found that the exo assistance moment was further reduced by 9.2 Nm. Thus, the trunk-worn weight resulted in a net exo moment that was much lower than the peak moment generated by the motors.
1️⃣ + 2️⃣ = low back relief experienced by the wearer was only 1/3 of the peak assistance output by this powered exosuit.
⭐️ For some exos and tasks—like this example—the difference between user-centric outcome and device-centric output can be large (>65%). For other exos and tasks, the difference can be relatively small (e.g., <20%).
☝🏽 The only way to know is to move beyond the vanity metrics of exoskeletons and perform the biomechanical analysis needed to understand real impact on users.
#exosuits #ergonomics #biomechanics #safetytech
