Functional kinematic and kinetic requirements of the upper limb during activities of daily living: a recommendation on necessary joint capabilities for prosthetic arms

TL;DR

This work addresses prosthetic abandonment by providing a data-driven framework of functional upper-limb requirements for ADLs, including ROM, velocities, and joint torques. Using the ADLDAT dataset and OpenSim-based IK/ID analyses, the authors generate 45 limb models per trial and augment object interactions with interaction wrenches, then train 600 linear regression models to predict torque demands from mass distributions. The study demonstrates near-perfect predictive power ($R \approx 0.99$) and illustrates practical design gains by optimizing wrist axis orientations, achieving 22%–38% reductions in peak power. The resulting dataset and regression framework offer a concrete, data-supported foundation for designing lighter, more functional prosthetic devices that align with real-world task demands.

Abstract

Prosthetic limb abandonment remains an unsolved challenge as amputees consistently reject their devices. Current prosthetic designs often fail to balance human-like perfomance with acceptable device weight, highlighting the need for optimised designs tailored to modern tasks. This study aims to provide a comprehensive dataset of joint kinematics and kinetics essential for performing activities of daily living (ADL), thereby informing the design of more functional and user-friendly prosthetic devices. Functionally required Ranges of Motion (ROM), velocities, and torques for the Glenohumeral (rotation), elbow, Radioulnar, and wrist joints were computed using motion capture data from 12 subjects performing 24 ADLs. Our approach included the computation of joint torques for varying mass and inertia properties of the upper limb, while torques induced by the manipulation of experimental objects were considered by their interaction wrench with the subjects hand. Joint torques pertaining to individual ADL scaled linearly with limb and object mass and mass distribution, permitting their generalisation to not explicitly simulated limb and object dynamics with linear regressors (LRM), exhibiting coefficients of determination R = 0.99 pm 0.01. Exemplifying an application of data-driven prosthesis design, we optimise wrist axes orientations for two serial and two differential joint configurations. Optimised axes reduced peak power requirements, between 22 to 38 percent compared to anatomical configurations, by exploiting high torque correlations between Ulnar deviation and wrist flexion/extension joints. This study offers critical insights into the functional requirements of upper limb prostheses, providing a valuable foundation for data-driven prosthetic design that addresses key user concerns and enhances device adoption.

Figures (5)

• Figure 1: Top: Composition of dynamic limb models. 20 cylinders (5 masses, 4 positions) on the humerus; 20 cylinders (5 masses, 4 positions) on the Ulna; 5 hand masses. Bottom: Experimental task set with illustrations outlining trial motion patterns during the manipulation phase and utilised experimental objects.
• Figure 2: Box plots for joint angles ($^{\circ}$) (row 1), joint velocities ($^{\circ}$/s) (row 2), and joint torques (Nm) (row 3). Boxes are specific to joints and tasks, with data concatenated over all subjects and repetitions. Torque Box plots were computed for the Ulna - solid black outline (4 cylinders as depicted in Fig. \ref{['fig: task_set']} of 0.5kg each), the hand - coloured outline (0.5 kg) and selected task objects $\in$$\{I$ (Mug 0.5 kg), $II$ (Mug 0.5 kg), $IV$ (Bottle 0.5 kg), $V$ (Briefcase 1kg), $VI$ (Tin can 0.5kg), $IX$ (Key 1.3Nm), $X$ (Box 1kg)$\}$ - dashed black outline.
• Figure 3: Coefficients of linear regression models fitted to percentile joint torques. Colours distinguish forearm (red), hand (blue), and object (yellow) dynamics models. The magnitude of each $K_C$ is indicated by the bar height with percentile models ($p \in \{0, 25, 50, 75, 100\}$) ordered left to right. Grey bars mark models with coefficients of determination R $<$ 0.99. Density plots indicate the torque distributions of normalized joint torques, used in the model fitting. Red: torques produced by a cylinder placed at the most proximal Ulna position (1/8). Grey-scale: positions (3/8, 5/8, 7/8). Blue: torques produced by the hand. Yellow: object induced torques.
• Figure 4: Optimized wrist joint axes and torque principal components in the wrist configuration space. Axes lengths correspond to the required actuator power, with blue and red markers indicating axes A and B, respectively. Purple markers indicate torque samples normalised by $max \sqrt{\tau_{WF}^2 + \tau_{WD}^2}$
• Figure 5: Comparison functional joint RoM in this and previous studies