Robotics Research Hub
Effects of Assistance Delay on Joint Mechanics and Energetics in Biological Torque Control of a Hip Exoskeleton

Effects of Assistance Delay on Joint Mechanics and Energetics in Biological Torque Control of a Hip Exoskeleton

Jimin An, Ryan Lee, Jingshu Peng, Eni Halilaj, Inseung Kang · N/A · 2026

locomotion reinforcement-learning control
PDF arXiv

Framework

N/A

License

N/A

Stars

N/A

Summary

Biological torque control directly maps an estimated human joint moment to exoskeleton assistance, providing a task-agnostic strategy for supporting diverse locomotor activities.

Abstract Summary

Biological torque control directly maps an estimated human joint moment to exoskeleton assistance, providing a task-agnostic strategy for supporting diverse locomotor activities. However, it remains unclear whether a fixed state-to-torque mapping provides effective assistance across biomechanically distinct tasks. We examined how assistance delay affected hip exoskeleton performance during level-ground (LG), ramp-ascent (RA), and ramp-descent (RD) walking. Eight participants completed a zero-torque baseline condition and five active assistance conditions with delays ranging from 40 to 320 ms. Across tasks and active delays, assistance reduced net metabolic rate by 5.24%, positive biological hip joint work by 5.86%, and total lower-limb positive joint work by 1.68% (all p < 0.05). Assistance delay affected both joint-work outcomes (both p < 0.001) but not net metabolic rate. Mechanical unloading generally decreased with increasing delay, whereas metabolic benefits remained comparatively stable. Relative to the zero-torque condition, net metabolic rate decreased by 9.75% during LG and 7.20% during RA but increased by 1.23% during RD. We did not detect task-dependent differences in the delay response. Our findings indicate that biological torque mappings should be evaluated based on the target outcome and mechanical role of the assisted joint, and that predominantly positive-power assistance may not generalize to negative-work-dominant locomotion without modification.

Key Points

  • Biological torque control directly maps an estimated human joint moment to exoskeleton assistance, providing a task-agnostic strategy for supporting diverse locomotor activities.
  • However, it remains unclear whether a fixed state-to-torque mapping provides effective assistance across biomechanically distinct tasks.
  • We examined how assistance delay affected hip exoskeleton performance during level-ground (LG), ramp-ascent (RA), and ramp-descent (RD) walking.
  • Eight participants completed a zero-torque baseline condition and five active assistance conditions with delays ranging from 40 to 320 ms.
  • Across tasks and active delays, assistance reduced net metabolic rate by 5.24%, positive biological hip joint work by 5.86%, and total lower-limb positive joint work by 1.68% (all p < 0.05).

Robotic Hip Exoskeleton

We developed a bilateral robotic hip exoskeleton to deliver assistive torque during locomotion (Fig. 1 A). The exoskeleton was designed to provide up to 18 Nm of hip torque across a sagittal-plane range of motion from 30 ∘ of extension to 100 ∘ of flexion. A compliant carbon-fiber passive joint permitted $\pm 5^{\circ}$ of hip abduction and adduction in the frontal plane.

The device weighed 4.3 kg and included two actuators (AK80-9; CubeMars, Nanchang, China), 3D-printed nylon user interfaces, body straps, and onboard electronics.

Robotic Hip Exoskeleton — Effects of Assistance Delay on Joint Mechanics and Energetics in Biological Torque Control of a Hip Exoskeleton
Fig. 1: Bilateral robotic hip exoskeleton and biological torque controller scheme. (A) The exoskeleton provides bilateral hip flexion and extension assistance during locomotion. Actuators are located at the user’s hip joints, and thigh-mounted IMUs provide kinematic information. The user interfaces are adjustable to accommodate varying body sizes. (B) The controller receives IMU inputs and estimates the biological hip joint moment independently for each limb. A predefined state-to-torque mapping function then scales, delays, and filters each estimate to generate the exoskeleton torque applied to the user.

Methods

The study protocol was approved by the Institutional Review Board of Carnegie Mellon University on April 1, 2025 (Protocol STUDY2025_00000002), and all participants provided written informed consent before the experiment. Eight healthy adults were recruited for this study (seven males and one female; age: $26.4\pm 2.9$ years; body mass: $71.8\pm 9.8$ kg; height: $1.77\pm 0.07$ m).

Participants walked on an instrumented treadmill (FIT5; Bertec, Columbus, OH, USA) for 6 min under each experimental condition.

Methods — Effects of Assistance Delay on Joint Mechanics and Energetics in Biological Torque Control of a Hip Exoskeleton
Fig. 2: Experimental setup and protocol. (A) Participants performed level-ground (LG), ramp-ascent (RA), and ramp-descent (RD) walking on an instrumented treadmill while metabolic rate was measured by indirect calorimetry and kinematics were recorded with reflective motion-capture markers. Red arrows indicate the ground-reaction forces measured by the instrumented treadmill. (B) Experimental sequence comprising standing, NoExo, NoAssi, and five active-delay conditions across the three locomotor tasks. Task order was randomized for both the NoExo and exoskeleton trials, and the order of the active-delay conditions was independently randomized to minimize potential order effects.

Results

Averaged across tasks and active-delay conditions, assistance reduced all three objective outcomes relative to NoAssi: net metabolic rate by 5.24% ($p=0.009$), positive biological hip joint work by 5.86% ($p=0.007$), and total lower-limb positive joint work by 1.68% ($p=0.013$). The dependence of these reductions on delay differed across outcomes (Outcome $\times$ Delay interaction, $p 3 ).

Results — Effects of Assistance Delay on Joint Mechanics and Energetics in Biological Torque Control of a Hip Exoskeleton
Fig. 3: Assistance-delay responses in metabolic and positive joint-work outcomes. Panels show net metabolic rate, positive biological hip joint work, and total lower-limb positive joint work. Markers and error bars represent the mean percent change $\pm$ SEM from the task-matched NoAssi condition across eight participants. Negative values indicate reductions relative to NoAssi, and the dashed horizontal line denotes no change. Circles, squares, and triangles represent level-ground (LG), ramp-ascent (RA), and ramp-descent (RD) walking, respectively. Solid lines represent task-specific fixed-effect predictions from exploratory linear mixed-effects models that included Delay (continuous, scaled per 70-ms increment), Task, and their interaction as fixed effects and participant as a random intercept. Color-matched asterisks mark active-delay conditions that differed from the task-matched NoAssi condition (paired $t$-tests, $p_{\mathrm{adj}}<0.05$, Bonferroni-corrected across the five delay comparisons within each task and outcome).

Sources and Demonstrations

Figures are reproduced from An et al. See the full paper for experimental details and the project page, when the paper links one, for demonstrations and videos.

Share

Related Papers

Adaptive Rejection of Unknown Hybrid Sinusoidal Disturbances with Application to Humanoid Walking under Unknown Ground Motion
arXiv preprint

Adaptive Rejection of Unknown Hybrid Sinusoidal Disturbances with Application to Humanoid Walking under Unknown Ground Motion

I-Chia Chang, Jacob Stewart, Petros A. Ioannou et al.

Legged robot control is not yet practical in non-inertial environments, where the ground and surroundings accelerate in the inertial Earth frame, mainly because of the unknown, time-varying, persistent perturbations caused by environmental motion, compounded by the associated hybrid robotic dynamics.

locomotion reinforcement-learning control
PDF Intermediate