muscle wrapping.pdf

45
MUSCLE WRAPPING

Upload: borrellano

Post on 14-Dec-2015

224 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: MUSCLE WRAPPING.pdf

MUSCLE WRAPPING

Page 2: MUSCLE WRAPPING.pdf

Muscoloskeletal geometry

1. Modeling the Paths of Muscolotendinous Actuators

2. Obstacle Set Method

Page 3: MUSCLE WRAPPING.pdf

1. Modeling the Paths of Muscolotendinous Actuators

Two different methods are used to model the paths of musculotendinous actuators

Straigth-line method(Jensen and Davy, 1975)

The path of a musculotendinousactuator (muscle and tendon combined) is represented by a straight line joining thecentroids of the tendon attachment sites

easy to implementX useless results when a muscle

wraps around a bone or another muscle

Centroid-line method(Jensen and Davy, 1975)

The path of the musculotendinous actuator is represented as a line passing through the locus of cross-sectional centroids of the actuator

the actuator’s line of action is represented more accurately in this way

X the centroid-line method can be difficult to apply

Page 4: MUSCLE WRAPPING.pdf

1. Modeling the Paths of Muscolotendinous Actuators

Two different methods are used to model the paths of musculotendinous actuators

Straigth-line method(Jensen and Davy, 1975)

The path of a musculotendinousactuator (muscle and tendon combined) is represented by a straight line joining thecentroids of the tendon attachment sites

easy to implementX useless results when a muscle

wraps around a bone or another muscle

Centroid-line method(Jensen and Davy, 1975)

The path of the musculotendinous actuator is represented as a line passing through the locus of cross-sectional centroids of the actuator

the actuator’s line of action is represented more accurately in this way

X the centroid-line method can be difficult to apply

Page 5: MUSCLE WRAPPING.pdf

1. Modeling the Paths of Muscolotendinous Actuators

Two different methods are used to model the paths of musculotendinous actuators

a) straight-line method

b) centroid-line method

The centroid-line method can be difficult to apply:

1) it may not be possible to obtain the locations of the actuator’s cross-sectional centroids for even a single position of the body

2) even if an actuator’s centroid path is known for one position of the body, it is practically impossible to determine how this path changes as body position changes

Page 6: MUSCLE WRAPPING.pdf

1. Modeling the Paths of Muscolotendinous Actuators

Two different methods are used to model the paths of musculotendinous actuators

a) straight-line method

b) centroid-line method

The centroid-line method can be difficult to apply:

1) it may not be possible to obtain the locations of the actuator’s cross-sectional centroids for even a single position of the body

2) even if an actuator’s centroid path is known for one position of the body, it is practically impossible to determine how this path changes as body position changes

Via points at specific locations along the centroid path of the actuator

(Brand et al., 1982; Delp et al., 1990)

In this approach, the actuator’s line of action is defined by either straight-line segments or a combination of straight-line and curved-line segments between each set of via points.

The via points remain fixed relative to the bones even as the joints move, and muscle wrapping is taken into account by making the via points active or inactive, depending on the configuration of the joint

this method works quite well when a muscle spans a 1-dof hinge joint

X it can lead to discontinuities in the calculated values of moment arms when joints have more than 1 rotational dof

Page 7: MUSCLE WRAPPING.pdf

1. Modeling the Paths of Muscolotendinous Actuators

Two different methods are used to model the paths of musculotendinous actuators

a) straight-line method

b) centroid-line method

The centroid-line method can be difficult to apply:

1) it may not be possible to obtain the locations of the actuator’s cross-sectional centroids for even a single position of the body

2) even if an actuator’s centroid path is known for one position of the body, it is practically impossible to determine how this path changes as body position changes

Via points at specific locations along the centroid path of the actuator

(Brand et al., 1982; Delp et al., 1990)

Obstacle-Set Method(Garner and Pandy, 2000)

This method idealizes each musculotendinousactuator as a frictionless elastic band that can slide freely over the bones and other actuators as the configuration of the joint changes.

The musculotendinous path is defined by a series ofstraight-line and curved-line segments joined together by via points, which may or may not be fixedrelative to the bones

Page 8: MUSCLE WRAPPING.pdf

• Methods Comparison

Fig.1

Page 9: MUSCLE WRAPPING.pdf

• Obstacle-Set Method

The locations of the attachment sitesof the muscle and the locations andorientations of the obstacles werechosen to reproduce the centroidpaths of each portion of the modeledmuscle.

Because the path of a muscle is notimproperly constrained by contactwith neighboring muscles and bones,the obstacle-set method produces:accurate estimates of muscle

moment armssmooth moment armjoint angle

curves, as illustrated in Fig.1

Page 10: MUSCLE WRAPPING.pdf

• Muscle moment arms

Muscles develop forces and cause rotation of the bones about a joint.

Moment arm: describes the tendency of a muscolotendinous actuatorto rotate a bone around a joint.

Two methods are commonly used to measure the moment arm of anactuator:

1. the geometric method

2. the tendon excursion method

Page 11: MUSCLE WRAPPING.pdf

• Muscle moment arms

1. The geometric method (Jensen and Davy, 1975)

• The finite center of rotation is found by x-rays, CT, or MagneticResonance Imaging

• The moment arm is found by measuring the perpendicular distancefrom the joint center to the line of action of the muscle

Page 12: MUSCLE WRAPPING.pdf

• Muscle moment arms

2. The tendon excursion method (An et al., 1983)

• The change in length of the musculotendinous actuator is measuredas a function of the joint angle

• The moment arm is obtained by evaluating the slope of the actuator-length versus joint-angle curve over the full range of joint movement

Page 13: MUSCLE WRAPPING.pdf

• Muscle contraction dynamics

Modeling contraction dynamics

An empirical model, proposed by A.V. Hill, is used in virtually all models ofmovement to account for the force-length and force-velocity properties of muscle(Hill, 1938).

In a Hill-type model, muscle’s force-producing properties are described by fourparameters (Zajac, 1989):

muscle’s peak isometric force, F0m

its corresponding fiber length, l0m

pennation angle, α

the intrinsic shortening velocity of muscle, vmax

Page 14: MUSCLE WRAPPING.pdf

• Muscle contraction dynamics

• F0m is usually obtained by multiplying muscle’s physiological cross-sectional area

by a generic value of specific tension.

• Values of optimal muscle fiber length, l0m, and α, the angle at which muscle fibers

insert on tendon when the fibers are at their optimal length, are almost alwaysbased on data obtained from cadaver dissections (Freiderich and Brand, 1990).

• vmax is often assumed to be muscle independent; for example, simulations ofjumping (Pandy et al., 1990), pedaling (Raasch et al., 1997), and walking(Anderson and Pandy, 2001b) assume a value of vmax = 10 s-1 for all muscles,which models the summed effect of slow, intermediate, and fast fibers (Zajac,1989).

Page 15: MUSCLE WRAPPING.pdf

• Muscle contraction dynamics

Tendon is usually represented as elastic (Pandy et al., 1990; Anderson and Pandy,1993).

Even though force varies nonlinearly with a change in length as tendon is stretchedfrom its resting lS

T length, a linear force-length curve is sometimes used (Andersonand Pandy, 1993).

This simplification will overestimate the amount of strain energy stored in tendon,but the effect on actuator performance is not likely to be significant, becausetendon force is small in the region where the force-length curve is nonlinear.

Page 16: MUSCLE WRAPPING.pdf

• Muscle contraction dynamics

For the actuator shown in Fig.2, musculotendon dynamics is described by a single, nonlinear, differential equation that relates musculotendon force (FMT), musculotendon length (lMT), musculotendon shortening velocity (vMT), and muscle activation (am) to the time rate of change in musculotendon force:

ḞMT = f(FMT, lMT, vMT, am), 0≤am≤

1, (1)

Given values of FMT, lMT, vMT, and am at one instant in time, Eq. (1) can be integrated numerically to find musculotendon force at the next instant.

Page 17: MUSCLE WRAPPING.pdf

• Muscle contraction dynamics

Fig.2 Schematic diagram of a model commonly used to simulate musculotendon actuation.

Page 18: MUSCLE WRAPPING.pdf

• Muscle contraction dynamics

Each musculotendon actuator isrepresented as a three-elementmuscle in series with an elastictendon.

The mechanical behavior of muscle isdescribed by:

• a Hill-type contractile element (CE)that models muscle’s force-length-velocity property

• a series elastic element (SEE) thatmodels muscle’s active stiffness

• a parallel-elastic element (PEE) thatmodels muscle’s passive stiffness.

Page 19: MUSCLE WRAPPING.pdf

• Muscle contraction dynamics

The instantaneous length of theactuator is determined by the lengthof the muscle, the length of thetendon, and the pennation angle ofthe muscle.

In this model the width of the muscleis assumed to remain constant asmuscle length changes.

Page 20: MUSCLE WRAPPING.pdf

1. Ricostruzione Mesh 3D

Page 21: MUSCLE WRAPPING.pdf
Page 22: MUSCLE WRAPPING.pdf
Page 23: MUSCLE WRAPPING.pdf

2. Confronto lunghezza elica-arco cilindro

a) Cambio sistema di riferimento

Page 24: MUSCLE WRAPPING.pdf

close all

clear all

clc

A=[0,0,0;0,0,0;0,0,0];

tol=10^(-14);

a0=[1;1;1]

b0=[1;2;1]

c0=[0;2;3]

a1=[2.612372436;0.387627564;1.5]

b1=[2.862372436;1.137627564;2.11237243

6]

c1=[3.337117307;-

0.337117307;3.724744872]

% a0=[1;1;7]

% b0=[4;7;1]

% c0=[7;10;10]

% a1=[7;1;7]

% b1=[3;9;6]

% c1=[10;10;13]

d=([a1-a0,b1-b0,c1-c0])'

u=inv(d)*[1;1;1]

uu=u/(sqrt((u(1))^2+(u(2))^2+(u(3))^2)

u_emi=[0,-uu(3),uu(2);uu(3),0,-uu(1);-

uu(2),uu(1),0]

h1=u_emi*(a0-b0)

h2=u_emi*(a1-b1)

c=(((h1)')*h2)/((norm(h1))*(norm(h2)))

theta=acos(c)

A=[(uu(1)^2)*(1-c)+c,

(uu(1)*uu(2))*(1-c)-uu(3)*sin(theta),

(uu(1)*uu(3))*(1-c)+uu(2)*sin(theta);

(uu(1)*uu(2))*(1c)+uu(3)*sin(theta),

(uu(2)^2)*(1-c)+c,(uu(2)*uu(3))*(1-c)

uu(1)*sin(theta);

(uu(1)*uu(3))*(1-c)-uu(2)*sin(theta),

(uu(2)*uu(3))*(1-c)+uu(1)*sin(theta),

(uu(3)^2)*(1-c)+c]

Page 25: MUSCLE WRAPPING.pdf

2. Confronto lunghezza elica-arco cilindro

b) Calcolo parametri elica

Page 26: MUSCLE WRAPPING.pdf
Page 27: MUSCLE WRAPPING.pdf

close all

clear all

clc

% x1=0.02236068

% y1=0.01118034

% z1=0.10

% x2=0.00928477

% y2=0.02321192

% z2=0.15

% t1=atan(y1/x1)

% t2=atan(y2/x2)

% c=((z1-z2)/(t1-t2))

% k=z1-((z1-z2)/(t1-t2))*t1

% x1=-0.01345541

% y1=0.02107017

% z1=0.16494454

% x2=0.02491081

% y2=0.0021099

% z2=0.11129092

% t1=atan(y1/x1)

% t2=atan(y2/x2)

% c=((z1-z2)/(t1-t2))

% k=z1-((z1-z2)/(t1-t2))*t1

% x1=1.94448291*10^(-3)

% y1=14.87343223*10^(-3)

% z1=66.39059567*10^(-3)

% x2=-14.02379309*10^(-3)

% y2=-5.32289651*10^(-3)

% z2=64.55098222*10^(-3)

x2= 14.02379309*(10^(-3))

y2= -5.32289651*(10^(-3))

z2= 63.44430452*(10^(-3))

x1= 1.94448292*(10^(-3))

y1= 14.87343223*(10^(-3))

z1= 76.86365141*(10^(-3))

t1=atan(y1/x1)

t2=atan(y2/x2)

c=((z1-z2)/(t1-t2))

k=z1-((z1-z2)/(t1-t2))*t1

Page 28: MUSCLE WRAPPING.pdf

2. Confronto lunghezza elica-arco cilindro

c) Confronto lunghezza

Page 29: MUSCLE WRAPPING.pdf

close all

clear all

clc

a=1;

b=-1;

c=2;

d=-3;

theta = 0:0.1:pi/2;

plot3(cos(theta),sin(theta),-

(a*cos(theta)+b*sin(theta)+d)/c)

view(135,30)

grid on

box on

f = @(theta) sqrt(sin(theta).^2 +

cos(theta).^2 + ((-

a*sin(theta)+b*cos(theta))/(c)).^2);

len1 = integral(f,0,pi/2)

hold on

x1=1

y1=0

z1=1

x2=0

y2=1

z2=2

t1=atan(y1/x1)

t2=atan(y2/x2)

c=((z1-z2)/(t1-t2))

k=z1-((z1-z2)/(t1-t2))*t1

r = 1;

t = t1:0.01:t2;

x = r*cos(t);

y = r*sin(t);

z = c*t+k;

figure(1)

plot3(x, y, z);

grid on

xlabel('x'); ylabel('y');

title('Circula helix');

f = @(t) sqrt((sin(t)).^2 +

(cos(t)).^2 + (c).^2);

len2 = integral(f,t1,t2)

len1

len2

diff1=(abs(len1-len2))/(len2)

diff=diff1*100

Page 30: MUSCLE WRAPPING.pdf
Page 31: MUSCLE WRAPPING.pdf

3. Confronto wrapping sfera-cilindro

Errore relativo sfera-lunghezza esatta, cilindro-lunghezza esatta

Page 32: MUSCLE WRAPPING.pdf

3. Confronto wrapping sfera-cilindro

Errore relativo sfera-lunghezza esatta, cilindro-lunghezza esatta

Page 33: MUSCLE WRAPPING.pdf

3. Confronto wrapping sfera-cilindro

Lunghezza esatta

Page 34: MUSCLE WRAPPING.pdf

4. Problema della retta d’azione

Page 35: MUSCLE WRAPPING.pdf
Page 36: MUSCLE WRAPPING.pdf
Page 37: MUSCLE WRAPPING.pdf

5. Soluzione proposta

Page 38: MUSCLE WRAPPING.pdf
Page 39: MUSCLE WRAPPING.pdf
Page 40: MUSCLE WRAPPING.pdf
Page 41: MUSCLE WRAPPING.pdf
Page 42: MUSCLE WRAPPING.pdf

Bisogna distinguere tra:

I. il problema della determinazione del raggio del cilindro/sferaottimo, che meglio approssima l’andamento della lunghezza delsistema muscolo-tendine precedentemente calcolato come lalunghezza minima che unisce il punto di Origine e di Inserzione

II. il problema della determinazione del numero di sferette dautilizzare nella simulazione dinamica al fine di ottenere valoriomogenei in termini di forza di reazione su ciascuna molla, inquanto l’attenzione è sulla compenetrazione del cilindro da partedella retta d’azione che collega le singole sferette

Page 43: MUSCLE WRAPPING.pdf

Bisogna distinguere tra:

I. il problema della determinazione del raggio del cilindro/sferaottimo, che meglio approssima l’andamento della lunghezza delsistema muscolo-tendine precedentemente calcolato come lalunghezza minima che unisce il punto di Origine e di Inserzione

Confrontare i risultati ottenuti con raggi differenti, trovare il raggio che minimizza l’errore relativo, proporre un criterio che

possa essere replicabile su ciascun modello patient-specific

Page 44: MUSCLE WRAPPING.pdf
Page 45: MUSCLE WRAPPING.pdf

Bisogna distinguere tra:

II. il problema della determinazione del numero di sferette dautilizzare nella simulazione dinamica al fine di ottenere valoriomogenei in termini di forza di reazione su ciascuna molla, inquanto l’attenzione è sulla compenetrazione del cilindro da partedella retta d’azione che collega le singole sferette

e.g. è possibile utilizzare una sola sferetta per i segmenti con angolo relativo di 180 gradi, si rendono necessarie 4-5-6-7-8

sferette per un angolo di 45 gradi