A Framework for Adaptive Load Redistribution in Human-Exoskeleton-Cobot Systems

TL;DR

This work tackles excessive joint loading during manual tasks when external loads are not aligned with the joints supported by wearable exoskeletons. It introduces an adaptive human–exoskeleton–cobot framework that coordinates a 1-DOF elbow exoskeleton with a cobot and avatar feedback, using online optimization to redistribute load toward non-supported joints as needed. The method combines a URDF-based arm model with a weighted objective $||\bm{\tau}^T\mathbb{W}\bm{\tau}||$, an augmented Lagrangian solver, Cartesian impedance-controlled cobot planning, and SEA-based elbow actuation to offset forearm and part of the load. Experimental validation on a painting task with four subjects demonstrates substantial joint-torque and EMG reductions while maintaining accurate pose tracking, signaling practical ergonomic benefits for industrial co-manipulation. Limitations include the quasi-static assumption and calibration-bound operation, with future work aiming at dynamic, full-body, and multi-exoskeleton extensions and broader statistical validation.

Abstract

Wearable devices like exoskeletons are designed to reduce excessive loads on specific joints of the body. Specifically, single- or two-degrees-of-freedom (DOF) upper-body industrial exoskeletons typically focus on compensating for the strain on the elbow and shoulder joints. However, during daily activities, there is no assurance that external loads are correctly aligned with the supported joints. Optimizing work processes to ensure that external loads are primarily (to the extent that they can be compensated by the exoskeleton) directed onto the supported joints can significantly enhance the overall usability of these devices and the ergonomics of their users. Collaborative robots (cobots) can play a role in this optimization, complementing the collaborative aspects of human work. In this study, we propose an adaptive and coordinated control system for the human-cobot-exoskeleton interaction. This system adjusts the task coordinates to maximize the utilization of the supported joints. When the torque limits of the exoskeleton are exceeded, the framework continuously adapts the task frame, redistributing excessive loads to non-supported body joints to prevent overloading the supported ones. We validated our approach in an equivalent industrial painting task involving a single-DOF elbow exoskeleton, a cobot, and four subjects, each tested in four different initial arm configurations with five distinct optimisation weight matrices and two different payloads.

Figures (9)

• Figure 1: This paper presents a solution that integrates exoskeletons for joint support, humans for supervisory roles, and cobots for adaptive task planning to optimize joint usage. It ensures that external loads are primarily directed onto the supported joints, to the extent that the exoskeleton can compensate. Through adaptive control, the cobot helps align the actual human pose (blue) with the optimized solution which is fed back to the user (grey).
• Figure 2: A) The isometric view of the right arm with the assigned coordinates. (B) The illustration of the arm attachments from side view. The dashed red line represents the Bowden cable that transfers the generated assistive force from the actuation system to the arm. The vertical distance between elbow brace (shown in gray color) and first anchor point on the forearm (a), the distance of fixed attachments from elbow frame $\Sigma_{E}$ (b), the center of mass of the forearm (b+c), and the lever arm (b+c+d) from the external load ($W_{P}$) to $O_{2}$. (C) shows the details of the robot with the assigned coordinates from side view. (e) is the fixed distance between hand and target frame.
• Figure 3: The proposed control strategy to alleviate the overloading joint torques by means of the robot and the elbow exoskeleton. IK and HBPs represent the inverse kinematics, and human body parameters, respectively. Black arrows are the inputs whereas red arrows show the flow of the control algorithm.
• Figure 4: The experiment-A results of $S_{1}$ where $2$ kg payload is held by assigning null value in $\mathbb{W}$ to record $\bm{\tau_{Max}}$ at a predetermined pose. $F_{R}$, and $F_{M}$ are the desired and measured force of the exoskeleton, respectively.
• Figure 5: The experiment-C results of $S_{3}$ in which $W_{P}=2$ kg payload is held without wearing the assistive device. $F_{R}$, and $F_{M}$ are the desired and measured force of the exoskeleton, respectively. $\tau$ is the joint torque.