Elsevier

Journal of Biomechanics

Volume 36, Issue 9, September 2003, Pages 1301-1307
Journal of Biomechanics

Instant center of rotation estimation using the Reuleaux technique and a Lateral Extrapolation technique

https://doi.org/10.1016/S0021-9290(03)00156-8Get rights and content

Abstract

A mathematical model of a rolling wheel was used to investigate the errors encountered when the Reuleaux technique is employed to estimate planar instant centers of rotation (ICRs). The investigation showed that large errors can result when this pole measurement technique is applied to objects rotating more than 12°. The investigation also showed that these errors can be substantially reduced by applying a new Lateral Extrapolation technique to the pole data. When the Reuleaux technique is applied to marks on a 10 cm radius wheel, the resulting offset errors from the ICR are 3.96 cm for a 45° roll and 1 cm for a 12° roll. Following lateral extrapolation, these offset errors reduce to 0.52 cm for the 45° roll and less than 0.04 cm for the 12° roll. Thus, the extrapolation technique is over seven times more accurate for a 45° roll, and over 25 times more accurate for a 12° roll. The extrapolation technique has been validated with the model for joints that exhibit both slip and roll, such as the knee. As joint ICR pathway measurement can be used to detect pathology, these accuracy improvements offer potential benefits for clinical applications.

Introduction

Numerous reports have described how the Reuleaux technique can be used to calculate the planar instant center of rotation (ICR) of human joints (Frankel et al., 1971; Grant, 1973; Tamea and Henning, 1981; Gerber and Matter, 1983; Haher et al., 1992; Maganaris, 2000). Most reports describe how joint movement is associated with ICR movement along a pathway. The majority also describe how ICR pathways can be affected by injury and disease. ICR pathway measurement can therefore provide valuable information about a joint's kinematic characteristics. This information can aid both the detection of injury and the assessment of treatment outcome.

The main drawback with the Reuleaux technique is that it does not accurately deduce a joint's ICR. Reuleaux's original article describes his technique as yielding a temporary center of rotation (TCR) that only approximates to an ICR for small rotation angles (Reuleaux, 1876). While this observation is theoretically correct, investigations have shown that in practice the technique is susceptible to error magnification with small rotation angles (Panjabi, 1979; Bryant et al., 1984). Despite these observations the technique remains interesting due to its simplicity and ease of application.

The remainder of this section examines the Reuleaux technique and the errors encountered when it is used to approximate Instant Center data. The following section then describes how Instant Center data can be more accurately estimated by applying a Lateral Extrapolation technique to the Reuleaux TCR data. Finally, the article quantifies the accuracy improvement of the extrapolation technique using data collected from a mathematical model of a rolling wheel.

Vector analysis of a rolling wheel shows that at any instant its ICR is its contact point with the ground (Barton, 1984), as shown in Fig. 1a. Further analysis shows that slippage causes the ICR to displace vertically, as shown in Fig. 1b. An ICR overlay on an object's image therefore indicates its kinematic action at an instant in time.

Fig. 2 shows the Reuleaux technique applied to a radius mark on a model of a wheel rolling with 10% slippage (slip ratio s=0.1). The technique derives a TCR, that can also be referred to as a pole (Hartenberg and Denivit, 1964). The TCR (Pole) indicates a point about which the wheel can rotate to achieve the same displacement as the roll, by flying through the air in an arc. In the X direction the Pole lies midway between the roll start and end points at X=(1−s)/2. In the Y direction the offset is described by Eq. (1), derived as (A.9) in the appendix:Yp=r(1+(s−1)(cotanθ+cosecθ)θ/2).If rotation angle θ is small, then the pole will lie close to the roll's start and end ICRs, and it can be used to estimate both of these locations. As the rotation angle increases, the pole can be used to estimate an ICR midway through the roll, at angle θ/2.

In this model situation the position of the wheel can be calculated at angle θ/2. However, in practical kinematic measurement situations such an extrapolation may not be possible. For instance, if a pole is deduced from knee radiographs at 10° and 20° flexion, there may be no 15° image to overlay the result on. It is therefore desirable that pole data are extrapolated into ICR data that matches the available images. The ICRs can then be overlaid on the images so that an assessment of a joint's kinematic action can be made. The next section describes a technique for achieving this.

Section snippets

Method

The Lateral Extrapolation technique is a development of the Parallelogram Conversion technique, briefly described in an earlier report (Montgomery et al., 1998). Both techniques use the same X co-ordinate algorithm, but the updated version has an improved Y co-ordinate estimation. The new technique is outlined in Fig. 3.

Fig. 3a shows the calculated positions of a radius mark on a rolling wheel at 0°, 30° and 90° rotation, along with its associated poles and ICRs. Fig. 3b shows the ICRs

Results

In Fig. 4 the 10 cm radius wheel experienced a pure roll of 45°, followed by a 45° roll with 50% slip. The pure roll moved the wheel forward by 1/8 of its circumference producing a horizontal displacement of (2πr/8=)7.8 cm. The 50% slipping roll moved the wheel forward by half this amount. In the X direction, pole P12 lies midway between ICR1 and ICR2a, and pole P23 lies midway between ICR2b and ICR3. In the Y direction, both poles have small offset errors from the true ICRs.

If P12 is taken as an

Discussion

Examination of the offset errors outlined in Fig. 4, shows that P12 is a poor estimation of both ICR1 and ICR2a. The 45° roll between these points produces a pole with an error radius of 3.96 cm from each ICR. If this error is taken as a percentage of the wheel's radius, then the Reuleaux pole has a 39.6% offset error from the ICRs. By comparison the 0.52 cm lateral extrapolation error is only 5.2%. This error is smaller because the extrapolation removes the X error from the estimation.

Further

References (17)

There are more references available in the full text version of this article.

Cited by (22)

  • Intuitive assessment of modeled lumbar spinal motion by clustering and visualization of finite helical axes

    2021, Computers in Biology and Medicine
    Citation Excerpt :

    A third possibility would be to choose every umpteenth data point and calculate the FHA on a coarser discretization grid. We would, however, refrain from using this method as it introduces possibly unwanted inaccuracies [[43], Eqn. (1)]. The noisy distribution of FHA gradually transforms into a sheaf of similarly oriented line segments orthogonal to the sagittal plane.

  • The Lateral Column

    2020, Clinics in Podiatric Medicine and Surgery
    Citation Excerpt :

    Now, it is widely accepted that the axis of rotation in a joint is shifted when the joint undergoes arthrokinematic motions: roll, glide, and spin.19 Furthermore, the instantaneous shift in position of the axis can now be estimated, based on the motion and geometry of the articular surfaces.20 For instance, the fifth metatarsocuboid joint is a synovial joint that is classically uniaxial and triplanar, deviating 20° from the transverse plane, 35° from the sagittal plane, and 70° from the frontal plane, permitting pronation and supination.8,9

  • A general method for the optimal synthesis of mechanisms using prescribed instant center positions

    2016, Applied Mathematical Modelling
    Citation Excerpt :

    This situation is necessary in many applications in engineering and here is where the PIC approach plays an important role. In biomechanical research, for instance, the instant center of joints in upper and lower extremities [12], or in the replacement of cervical discs [4], should be observed to guarantee the correct functioning of prostheses. One particular application is the design of knee mechanisms for knee disarticulation prostheses.

  • Sagittal plane position of the functional joint centre of prosthetic foot/ankle mechanisms

    2010, Clinical Biomechanics
    Citation Excerpt :

    However, this is clearly not the case for the prosthetic feet tested (Mean pooled difference x/y = 39.6 mm/27.0 mm). It is possible to interpret the kinematic motion of the prosthetic foot based on the position of the FJCs (Moorehead et al., 2003). The more proximal the FJC is placed, the greater proportion of the range of motion, on average, the shank is spending translating over top of the foot during stance.

  • Effects on the Metatarsophalangeal Joint After Simulated First Tarsometatarsal Joint Arthrodesis

    2007, Journal of Foot and Ankle Surgery
    Citation Excerpt :

    These errors, however, were greater when the rotation of the bodies was greater than 12° and the larger radius of circumference. Using the Reuleaux method on small joints with a small change in angle, as in the current study, will minimize the errors (16). The findings of the study suggest that eliminating motion of the first TMT joint after a simulated first TMT arthrodesis does not have a negative effect at the first MTP joint.

View all citing articles on Scopus
View full text