A Custom 3D-printed Hand Long Finger Exoskeleton

Design of a Custom 3D-Printed Long Finger Exoskeleton 1st Alessandro Ceccarelli*2nd Fabrizio Taffoni 3rd Loredana Zollo 4th Nevio Luigi Tagliamonte Abstract—Hand exoskeletons can restore functions in individuals with hand motor impairments, supporting everyday life. This work presents a preliminary development of a modular, rigid finger exoskeleton, custom-designed for a specific user by using 3D scanning and additive manufacturing technologies. A linkage mechanism was synthesized and 3D-printed, integrating custom cuffs and a linear actuator. Performance analysis demonstrated that the exoskeleton module produces a natural motion, closely matching desired finger kinematic data, and highlights the potential of personalized solutions for effective hand assistance. Index Terms—Hand exoskeleton, 3D printing, Custom design I. Introduction The human hand is an intricate biomechanical system, with 21 Degrees of Freedom (DoFs), enabling dexterous movements and fine motor control [1]. Each long finger has 3 DoFs in flexion/extension (F/E), distributed to the MetaCarpoPhalangeal (MCP), Proximal InterPhalangeal (PIP) and Distal InterPhalangeal (DIP) joints. These joints exhibit functional intra-finger couplings (IFCs) during F/E tasks [2], i.e., intrinsic kinematic constraints that cause the finger to behave as a single-DoF kinematic chain. The structural complexity of the hand makes it particularly susceptible to injury, while muscular disorders and age-related degeneration can also lead to significant loss of function [3]. Social and healthcare demands for effective rehabilitation and assistance technologies have driven the development of hand exoskeletons, which aim to restore or assist hand function in individuals affected by stroke, spinal cord injury, or musculoskeletal diseases, supporting Activities of Daily Living (ADLs), thereby improving quality of life [4]. Rigid exoskeletons are usually made of hard materials, ensuring high force transmission and accurate force/motion control, but often at the cost of comfort and adaptability to the user’s anthropometric characteristics. Whereas soft exoskeletons use flexible materials, such as textiles or elastomers, and actuation solutions, such as cables or pneumatic elements, prioritizing wearability, and compliance with the user’s free movements, though they may offer less mechanical precision and lower force output. Rigid solutions, if properly customized, can potentially provide a comfort level comparable to that of soft exoskeletons, while delivering more effective assistance during ADLs. Recent advances in 3D scanning and Additive Manufacturing (AM) have made it possible to capture the exact All the Authors are with the Laboratory of Advanced Robotics and HumanCentred Technologies – CREO Lab, Universit ` a Campus Bio-Medico di Roma. This work was funded by the Istituto Nazionale per l’Assicurazione contro gli Infortuni sul Lavoro (INAIL) within the 3D-AID project (PR19-PAS-P3), Lowcost Hand Prostheses and Exoskeleton Robotic Aids (CUP: E59E19001460005). *Corresponding Author: [email protected]. morphology of the user, enable the creation of exoskeleton components that fit perfectly and minimize discomfort, and manufacture them with an affordable production process. This work describes a preliminary development of a long finger exoskeleton customized to the anatomical features of a specific user, to be integrated in a rigid device for ADLs. The proposed custom finger module consists of a rigid six-bar linkage including cuffs and a linear actuator and its novelty lies in the integration of IFCs in the mechanical design of a custom exoskeleton. II. Materials and Methods The kinematic synthesis, the CAD modeling, the 3D-printed prototype, and the kinematic performance analysis of an index finger module will be discussed in the following. A healthy 28 y.o. male volunteer was enrolled and signed an informed consent for participation. The experimental protocol was approved by the Ethics Committee from UCBM (Prot. PAR 75.22). 1) Kinematic synthesis: The mechanical design was pursued to ensure both functional performance and robustness to 3D printing inaccuracies, as done for a prosthetic finger in a previous work [5]. The collection of essential user constraints was performed via the Einscan HX 3D scanner (Shining 3D), to extract the finger shape, and via manual measurements, to extract the anatomical lengths of the finger phalanges, i.e., the joints distance MCP-PIP, PIP-DIP, DIP-fingertip (TIP), and thickness. In addition, IFCs were extracted based on a reference F/E dataset collected on healthy subjects by using a motion capture system [2]. PIP and DIP joints angle were expressed as linear functions of the MCP one as 𝐾𝑝𝑟𝑜𝑥,ℎ𝑢𝑚 =𝜃𝑃𝐼 𝑃/𝜃𝑀𝐶 𝑃 and 𝐾𝑑𝑖𝑠𝑡,ℎ𝑢𝑚 =𝜃𝐷𝐼 𝑃/𝜃𝑀𝐶 𝑃. A mechanism based on the composition of a six-bar and a slider-crank was selected. In the kinematic dimensional synthesis, the human phalanges were considered as a fixed and integral part of the mechanism. The process was performed based on the Loop-Closure Equations (LCEs) method, applied Fig. 1: Exoskeleton finger mechanism and its mechanical loops. 2025 I-RIM Conference October 17-19, Rome, Italy ISBN: 9788894580570 10.5281/zenodo.17629696 81 to three mechanical loops, i.e., the proximal, the distal, and the actuation ones, shown in Fig. 1 in blue, green, and orange, respectively. The link lengths ℓ𝑖 , with 𝑖=1, ..., 8 were changed in a defined search space, considering different constraints: anatomical and functional, as well as manufacturing (resolution and geometric tolerance of the AM technology). Each feasible mechanism was analyzed for possible kinematic singularities by evaluating the Jacobian matrix derived from the LCEs. Mechanisms exhibiting singular configurations were discarded. The retained mechanisms were evaluated by analyzing i) the similarity of their resultant IFCs to the reference F/E dataset and ii) the reaction forces and the actuation pushing force 𝐹𝑝𝑢𝑠ℎ required to balance an external load 𝐹𝑇 𝐼 𝑃 applied in static conditions. A General Performance Index (GPI) was defined as a sum of kinematic/static performance indexes. The mechanism that exceeded a defined threshold and exhibited the lowest sensitivity 𝜕 𝐺𝑃𝐼 to dimensional variations 𝜕 ℓ𝑖 was considered optimal and chosen to be prototyped, ensuring high kinetostatic performance and robustness to manufacturing inaccuracies. 2) 3D modeling and prototyping: The optimal mechanism was translated into a CAD model, and a compact Actuonix L12 linear actuator was added to drive the slider-crank portion of the mechanism and in turn the final F/E motion. Three custom cuffs derived from the 3D scan of the user’s hand were included. The digital model of each finger was slightly enlarged to create a ring that mirrors the user’s anatomy while ensuring sufficient clearance (1 mm w.r.t. the finger shape). The intermediate and distal cuffs were designed with embedded lateral pins. The proximal link, which connects the dorsal frame to the intermediate phalanx, and the distal link, which connects the proximal to the distal phalanx, were carefully modeled so to avoid any interference throughout the entire Range of Motion (RoM). The dorsal frame was realized to hold the actuator, to be attached to a hand orthosis and to work as support for the exoskeleton finger. It is a simple modular layered structure, able to guarantee adaptability to different user anatomical variations. The entire system was manufactured in PLA by using FDM 3D printing technology (Factor 4, Ultimaker). Velcro straps were used to attach the frame to the top of a commercial orthosis. 3) Performance analysis: A camera was used to collect the finger motion of the user wearing the exoskeleton. Four markers were placed on the side of each phalange and on the frame to register the phalange positions. They were extracted via Tracker software and evaluated to compare IFCs and the TIP trajectory with the reference F/E dataset used for the kinematic synthesis. III. Results and Conclusions The designed exoskeleton (Fig. 2A) was tailored for a specific user, and resulted to be mechanically robust. The clearance left between the finger and the cuffs provided comfort to the user. Resulting linear IFCs ( 𝐾𝑗𝑛𝑡,𝑒𝑥𝑜 ) slightly differ from the desired ones ( 𝐾𝑗𝑛𝑡,ℎ𝑢𝑚 ), with absolute errors 𝜀𝑗𝑛𝑡 , calculated as |𝐾𝑗𝑛𝑡,𝑒𝑥𝑜 −𝐾𝑗𝑛𝑡,ℎ𝑢𝑚 | , of 0.26 and 0.21 for the proximal and distal couplings, respectively. Fig. 2B shows their trends (in dashed black lines) and the desired reference F/E dataset Fig. 2: (A) Exoskeleton prototype (cuffs, links, dorsal frame) and mechanical loops. (B) Proximal and distal IFC trends and (C) TIP kinematics of the exoskeleton module w.r.t. the reference F/E dataset. (straight green and blue lines with their standard deviations). Even if the proximal coupling has a quite different trend, the distal one closely replicate the desired trajectory. As regards the TIP trajectory (shown in Fig. 2C), the initial configuration with the exoskeleton (in blue) perfectly matches the one without it (in red), but during flexion, the mechanism slightly bends the finger more than desired, as anticipated with the proximal IFC difference, resulting in a 𝑅𝑀𝑆𝐸𝑇 𝐼 𝑃 = 8.24 mm. This can be attributed to the fact that, despite the PIP trend is different to the reference one, their linear regression IFC are similar. However, a natural motion was obtained without compromising the integrity of the finger. Future works aim to extend the design to the other long fingers and to the thumb, which is much more complex and requires a dedicated modeling. Moreover, since a clearance between the fingers and cuffs could provide the exoskeleton slippage, the kinematic chain could be closed adding a robotic DIP joint and allowing the entire finger module to avoid relative motion. References [1] I. M. Bullock, J. Borr ` as, and A. M. Dollar, “Assessing assumptions in kinematic hand models: a review,” in IEEE BioRob, 2012. [2] M. Lapresa et al., “Analysis of hand intra-finger couplings during flexion movements in the free space,” IEEE Access, 2023. [3] B. Noronha and D. Accoto, “Exoskeletal devices for hand assistance and rehabilitation: A comprehensive analysis of state-of-the-art technologies,” IEEE Transactions on Medical Robotics and Bionics, 2021. [4] P. Tran et al., “Review: Hand exoskeleton systems, clinical rehabilitation practices, and future prospects,” IEEE TMRB, 2021. [5] A. Ceccarelli et al., “Mechanical design of a bioinspired and customized prosthetic hand finger based on six-bar linkage,” in IEEE BioRob, 2024. 82