Mechanical Design and Kinematic Modeling of a Cable-Driven Arm Exoskeleton Incorporating Inaccurate Human Limb Anthropomorphic Parameters

Weihai Chen, Zhongyi Li, Xiang Cui, Jianbin Zhang, Shaoping Bai

TL;DR

This paper presents a novel cable-driven upper-limb exoskeleton designed for rehabilitation, featuring a unique adaptive arm cuff to address the inherent inaccuracies in human limb anthropometry and flexible attachments. They developed an improved kinematic model that accounts for these uncertainties and validated its effectiveness through experiments, demonstrating enhanced accuracy for rehabilitation robots.

What Problem Does This Paper Address?

Rehabilitation exoskeletons are crucial for physical therapy, offering repetitive and intensive motion training. While cable-driven designs offer advantages like light weight and simpler structures compared to rigid-linkage ones, they pose significant challenges for kinematic modeling and control. This is primarily due to the reliance on the human limb as part of the system, leading to uncertainties arising from:

  • Inaccurate Anthropomorphic Parameters: Human limbs vary greatly in size and shape, making it difficult to create a one-size-fits-all exoskeleton.
  • Flexible Attachments: The way the exoskeleton attaches to the limb is not perfectly rigid, introducing further variability.
  • Human-Arm Skeleton Kinematics: The natural movement and slight shifts of the human skeleton during motion can lead to misalignments between the exoskeleton's joints and the human's actual joint centers.

These uncertainties can compromise the accuracy and robustness of motion control, hindering effective rehabilitation. The paper aims to tackle these issues by designing an exoskeleton that better accommodates human limb variations and by developing a more accurate kinematic model that accounts for these uncertainties.

Methods

The authors propose a multi-pronged approach:

  1. Mechanical Design with a Novel Arm Cuff: They designed a cable-driven arm rehabilitation exoskeleton with a focus on adaptability. A key innovation is a custom-designed, two-stage arm cuff. The inner stage uses a flexible silicon shell for a snug fit and uniform force distribution, while the outer stage provides a rigid frame for cable attachment points. This cuff is designed to adapt to irregular arm contours and reduce misalignments.
  2. Kinematic Modeling with Uncertainty Consideration: The kinematic model of the exoskeleton is developed by considering two main parts: the human arm skeleton (modeled as a serial open-chain mechanism) and the cable-driven modules (shoulder and elbow). Crucially, the model explicitly incorporates uncertainties related to:
  3. Floating Joint Centers: Deviations in the shoulder and elbow joint centers due to human limb movement.
  4. Attachment Errors: Inaccuracies in how the exoskeleton cuffs attach to the arm.
  5. Parameter Identification Method: To improve the kinematic model's accuracy, they employ a model-based parameter identification method. This iterative approach uses sensor data (motor encoders, IMUs) to estimate the uncertain parameters (like joint center deviations and attachment errors) that affect the model's predictions. This method refines the kinematic model by minimizing the difference between predicted and actual cable lengths.

Datasets

Not discussed in the paper. The experiments used real-time sensor data from the exoskeleton prototype and a motion capture system for validation.

Key Insights and Ideas

  • The "Human-in-the-Loop" Challenge: The paper highlights that for wearable exoskeletons, the human limb is not just a passive component but an active, variable part of the system. This inherent variability is a major hurdle for precise control.
  • Adaptive Cuffs are Key: The novel arm cuff design is a significant contribution. By allowing the exoskeleton to better conform to the user's anatomy, it directly addresses the problem of misalignments and flexible attachments, reducing a major source of kinematic uncertainty.
  • Uncertainty Quantification is Crucial: Simply modeling the exoskeleton in isolation isn't enough. The paper demonstrates the necessity of explicitly modeling and identifying parameters related to human limb variations and attachment inaccuracies for accurate kinematic prediction.
  • Iterative Identification for Robustness: The parameter identification method, which iteratively refines the model based on sensor feedback, is essential for achieving a kinematic model that is robust to real-world variations.

Key Results

  • Improved Kinematic Model Accuracy: Experiments showed that the kinematic model with identified parameters significantly reduced RMS errors in cable length calculations compared to a model without identification. For Subject A, RMS errors for cable lengths l1 to l6 were reduced from values like 11.9 mm to 5.1 mm, and for Subject B, from 15.6 mm to 4.4 mm.
  • Effective Parameter Identification: The parameter identification method converged within 8 iterations in simulations and demonstrated good agreement between identified shoulder joint center positions and motion capture data in experiments.
  • Adaptive Cuff Performance: While not explicitly quantified with numbers, the paper states the custom-designed arm cuff shows good adaptability and reduces uncertainties caused by instabilities between the exoskeleton and the human arm.

Limitations and Future Work

  • Limited Number of Subjects: Experiments were conducted with only two subjects (A and B).
  • Specific Task: The experiments focused on a straight-line tracking task for the upper limb.
  • Future Work: The authors plan to use the improved kinematic model for motion control of the exoskeleton. This suggests that the current work is a foundational step towards more sophisticated control strategies.

Summary updated

Abstract

Compared with conventional exoskeletons with rigid links, cable-driven upper-limb exoskeletons are light weight and have simple structures. However, cable-driven exoskeletons rely heavily on the human skeletal system for support. Kinematic modeling and control thus becomes very challenging due to inaccurate anthropomorphic parameters and flexible attachments. In this paper, the mechanical design of a cable-driven arm rehabilitation exoskeleton is proposed to accommodate human limbs of different sizes and shapes. A novel arm cuff able to adapt to the contours of human upper limbs is designed. This has given rise to an exoskeleton which reduces the uncertainties caused by instabilities between the exoskeleton and the human arm. A kinematic model of the exoskeleton is further developed by considering the inaccuracies of human-arm skeleton kinematics and attachment errors of the exoskeleton. A parameter identification method is used to improve the accuracy of the kinematic model. The developed kinematic model is finally tested with a primary experiment with an exoskeleton prototype.
BibTeX
@inproceedings{chen2019mechanical,
  author = {Weihai Chen and Zhongyi Li and Xiang Cui and Jianbin Zhang and Shaoping Bai},
  title = {{Mechanical Design and Kinematic Modeling of a Cable-Driven Arm Exoskeleton Incorporating Inaccurate Human Limb Anthropomorphic Parameters}},
  booktitle = {Sensors},
  year = {2019},
  url = {https://openreview.net/forum?id=nF1ES5hk-sA},
}
Figure