Session: Research Posters

Paper Number: 173519

A Lightweight Tendon-Driven Upper Limb Exoskeleton for Daily Assistance in Stroke Rehabilitation 

Stroke remains one of the leading causes of adult disability in the United States, with over 795,000 new cases reported annually. A significant portion of stroke survivors experience upper limb paralysis or severe motor impairment, requiring extensive rehabilitation to regain partial motor function. While modern infrastructure provides increasing accessibility for lower limb impairments, individuals with upper limb disabilities face persistent challenges in performing daily activities independently. To address this gap, we present the Daily Assistive Stroke Patient Exoskeleton (DASPE)—a lightweight, tendon-driven upper limb exoskeleton designed to assist users in regaining functional movement of a paralyzed arm by mirroring the motion of their healthy limb. 

DASPE incorporates three actively actuated degrees of freedom (DOFs): two at the shoulder and one at the elbow. These joints are actuated through high-torque CubeMars pancake motors connected via a tendon-driven mechanism. This configuration enables the motors to be repositioned away from the user’s shoulder and toward the lower back, improving comfort and weight distribution without compromising actuation speed or torque. The structure is primarily composed of custom 3D-printed components, which further reduces the system’s weight and manufacturing cost while preserving structural integrity. 

The control system architecture is multi-tiered, distributing computational and communication tasks across dedicated microcontrollers. A high-level controller manages sensor fusion and motion prediction, while mid- and low-level controllers handle real-time motor commands and safety constraints. The system uses four inertial measurement units (IMUs) to track the healthy arm’s kinematics and eight pressure sensors to ensure proper fit and detect user intent. The data-driven control model incorporates real-time calibration for each individual user, allowing the system to anticipate losses in tendon tension or mechanical backlash and compensate dynamically during motion. 

Experimental testing shows that DASPE can consistently mirror the healthy arm’s movement across a range of shoulder and elbow motions with a maximum angular error of less than five degrees. The system maintains this accuracy across varying actuation speeds and motion profiles, enabling smooth, reliable bilateral arm coordination. The use of tendons and distributed actuation significantly reduces the perceived weight on the arm, thereby increasing user comfort during extended use. 

DASPE represents a novel and accessible solution for daily assistive rehabilitation in stroke patients with upper limb motor deficits. By combining lightweight materials, tendon-driven actuation, and a hierarchical control architecture, our design offers a practical and effective tool for promoting independence and motor recovery in real-world settings. This poster will present the design architecture, control strategy, and experimental results supporting DASPE’s feasibility and performance.

Presenting Author: Connor Talley Kennesaw State University

Presenting Author Biography: Connor Talley is a PhD candidate at Kennesaw State University. His research focuses on compliant mechanisms, soft robots, and machine learning.

Authors:

Connor Talley Kennesaw State University
William Thompson Kennesaw State University
Aaron Grann Kennesaw State University
Ayse Tekes Kennesaw State University