Original Research Article
Brace for variability in tool positioning: Modeling and simulation of 1 DoF needle insertion task under tool-braced condition

https://doi.org/10.1016/j.bbe.2014.10.003Get rights and content

Abstract

Erroneous human arm motion leads to inaccuracy in tool positioning which often affects the quality of many precision manipulation tasks. Variability in tool endpoints during mechanical interactions among humans, tools, and environments can be reduced by bracing strategies; therefore, such a cost-effective approach may increase accuracy and reproducibility in human arm movements and thus prove useful in improving performance of selected interactive tasks. Although a great number of works is found in the literature related to performance augmentation of braced robotic manipulators, only a few studies are found on bracing strategies for performing interactive tasks and on analogies to human manipulation. In this paper, a method for predicting bracing properties is proposed for single degree of freedom interactive task to improve performance. To reflect real workplace scenarios, models of the human impedance, the brace, and the contact of the interaction are represented and used to predict the quality of the executed task. The task execution model intended to predict real-life performance in both free arm and braced conditions are then used to estimate the brace properties to augment performance of the representative task.

Introduction

Humans often try to enhance their manipulation task performance and reduce net cost by bracing their wrist or arm against a fixed object or ground. For example, artists often use repositionable braces called ‘mahl sticks’ to rest and steady the handheld brush while painting tiny subjects on a large canvas. Carpenters often engage both hands to operate power carving tools and, at the same time, to brace for the disturbance forces generated during the interactions. Many precision manipulation tasks in surgery also require both accuracy and human judgment. Hand tremors of ophthalmic surgeon are reported to be a RMS value of 182 μm [1], making it difficult to enter human retinal vein, of 40 μm to 350 μm in diameter [2], without leaving damage to the tissues during cannulation. The median RMS amplitude of the hand position of neurosurgeons was reported to be 526 μm in holding a tool still without bracing, whereas the median RMS error was reduced to 289 μm when tracing a line with bracing of the hand [3]. In some tasks, it is difficult to form a brace naturally by making close contact near the tool tip due to the tool length or heat generation in interaction. In such cases, supplementary mechanical supports in the form of braces can be utilized in order to improve task performance.

Bracing can be established through adding one or more mechanical links in parallel to any of the user-tool-target interactions in order to alter the mechanical impedance between the tool and the workpiece. Mechanical impedance is consisted of three main components: stiffness, damping and inertia. Given that the application of brace can have a significant effect on response time, exerted force and range of motion during interactive task execution, there is a need for a design tool to accurately predict the impact of a proposed brace design. If not designed properly, bracing may impose kinematic constraint which reduces one or more degrees of freedom of tool interaction. Loss of many degrees of freedom does not serve the general purpose of the manipulation, largely limiting the workspace [4]. Predicting the mechanical requirements of a bracing system for a given interactive task is therefore important if one is to better quantify the task constraints and ensure the necessary range of motion. The difficulties start when we try to estimate what the mechanical properties of a brace should be for a given task, particularly if we would like to get productive behavior from it. Decisions should be made on the basis of an integrated design approach predicting interactive performance in real-life. This is the point at which real challenges appear – the formulation of the associated models that interact with each other. How do we accurately measure the brace properties which will be effective in reduction of overall tool variability? How do we apply additional impedance along the tool axis while maintaining constraints in other directions? To find the answers, we describe how to estimate the proper bracing system for a certain reference task, single degree of freedom interactive needle insertion in this paper. We have started with a computational model of a needle insertion task, a one DoF translational manipulation, as a precursor to the gradually complex tasks in future. Single DoF systems are relatively simple to model and are good approximations for many real applications such as pointing, shooting, and drilling. Therefore, we predict the brace impedance by modeling the state estimation of a single degree-of-freedom interactive task through the equations in the state space.

Section snippets

Review of relevant examples from robotics

In the early 1980s, limitations were identified in the design of industrial serial manipulators such as poor mechanical stiffness as a result of their open kinematic chains. Bracing was found to be a cost-effective mean of providing additional support (e.g. extra stiffness) at a point closer to the end effector where load was applied, forming a closed kinematic chain [5]. This could significantly increase the mechanical stiffness of the manipulator, without significantly reducing the large

Proposed method

The mechanical properties for a brace system for improved tool manipulation performance can be estimated by considering three major components: (a) task description: a set of tool actions or trajectories that is required to follow in order to execute an interactive task and the geometry of the contact (environment), (b) user representation: a kinematic model of the neuromuscular user and the associated tool are viewed as an open loop kinematic chain that are equivalent to springs and dampers

Computational modeling of needle insertion task

To demonstrate the effect of bracing on the task performance, we have modeled a candidate bone marrow biopsy procedure in which a needle is introduced through the skin into the bone. By applying manual pressure, the needle is propelled through the bony cortex. Thus it has the advantage of sharing the mechanical features with many other interactive tasks such as drilling, burring, punching, reaching, etc. A deep plunge of needle insertion through the bone can cause injury, though there is little

Results

A roughly straight line is simulated for the first phase of needle insertion through skin and soft tissue (see Fig. 7). The following part of the curve represents cortical bone penetration and plunge of the needle beyond the cortex. For simplicity, we assumed that there is no interaction between the needle shaft and the anatomy after penetration and the plunge motion is a consequence of the passive dynamics of the arm and sudden imbalance of force. The last part of the simulation corresponds to

Conclusion

A computational estimation of the mechanical properties of a bracing system by displaying its effect on the performance of a simple interactive task has been presented. The properties of a brace can be characterized as mechanical impedance which can be modulated to reduce the variability in tool positioning based on task requirements. The simulation results show that bracing can effectively constrain the user motion and reduce the output variability. It is also observed from the simulation that

Acknowledgements

This research has been conducted in the Center for Hip Health and Mobility, the lab facility located at Vancouver General Hospital, Canada with affiliation with the University of British Columbia, Vancouver, Canada. The project is mostly funded by the Faculty of Graduate Studies at the University of British Columbia, Vancouver, Canada.

The author would like to thank Dr. Antony J. Hodgson, Professor and Director of Biomedical Engineering at the University of British Columbia, for his valuable and

References (30)

  • S.R. Moore et al.

    Part reference manipulation – a strategy applied to robotic drilling

  • H. Asada et al.

    Design of an adaptable tool guide for grinding robots

    Robot Comp Integr Manuf

    (1885)
  • T. Molander

    Routing and drilling with an industrial robot

  • E.D. Fasse et al.

    A novel variable mechanical impedance electromechanical actuator

    Dynam Syst Control

    (1994)
  • S. Stramigioli et al.

    A rigorous framework for interactive robot control

    Int J Control

    (2002)
  • Cited by (0)

    View full text