MATLAB error - "Error using symengine Input arguments must be convertible to floating-point numbers"

Question

Im writing a MATLAB code to animate a grashof four bar mechanism but my animation is running only till the 15th iteration and coming back with this error **Error using symengine Input arguments must be convertible to floating-point numbers.

Error in sym/mupadmexnout (line 1116) out = mupadmex(fcn,args{:});

Error in sym/max (line 182) [C,I] = mupadmexnout('symobj::maxmin',Xsym, dim,'max', nanflag);

Related documentation**

Is there a way to use peicewise instead of max in this scenario where both args are unknown? Your help and suggestions are much appreciated. Yhank You The code is attached below

a=1.6;
b=2.14;
c=2.06;
d=3.5;
AB=a
BC=b
CD=c
AD=d
xa=0;
ya=0;
yd=0;
xd=AD;
t2=0
for ii=1:36
    syms t3 t4


    t2=t2+10
    f=a*cosd(t2) + b*cosd(t3) - c*cosd(t4) -3.5

    g=a*sind(t2) + b*sind(t3) - c*sind(t4)

    [t31, t41]=(solve(f==0,g==0,t3,t4))
    var = vpa(t31)

    th3=double((max(t31)))

    th4=double((max(t41)))
    
    if exist('th3', 'var') 
        xb(ii)=AB*cosd(t2);
        yb(ii)=AB*sind(t2);
        xc(ii)=xb(ii)  + BC*cosd(th3)
        yc(ii)=yb(ii)  + BC*sind(th3)
        
        plot(xa,ya,'k.',xb,yb,'b.',xc,yc,'g.',xd,yd,'c.',LineWidth=3)
        hold on
        axis equal
        line_plot1 = plot([xa,xb(ii)], [ya,yb(ii)], 'b',LineWidth=3);
        line_plot2 = plot([xb(ii),xc(ii)], [yb(ii),yc(ii)], 'r',LineWidth=3);
        line_plot3 = plot([xc(ii),xd], [yc(ii),yd], 'm',LineWidth=3);
        line_plot4 = plot([xd,xa], [yd,ya], 'c',LineWidth=3);
        axis equal
        hold off
        pause(0.06)
    else
        continue
    end
end

`

Answer 1

at iteration 15 assumes value: -

(360atan(((2909776434409463614636144643^(1/2))/16381 - (35869579695136164170619*(- (10854241780483^(1/2))/65205930123 - 803672842364/65205930123)^(1/2))/2 - 204158498323266259123683^(1/2)((- (10854241780483^(1/2))/65205930123 - 803672842364/65205930123)^(1/2)/2 - (2109443^(1/2))/442287 + 1648/147429) + 11759769043149622934890992/16381)/(2331594814742111183183523^(1/2) + 401972381605023123837879)))/pi

-(360atan(((2909776434409463614636144643^(1/2))/16381 + (35869579695136164170619*(- (10854241780483^(1/2))/65205930123 - 803672842364/65205930123)^(1/2))/2 + 204158498323266259123683^(1/2)((2109443^(1/2))/442287 + (- (1085424178048*3^(1/2))/65205930123 - 803672842364/65205930123)^(1/2)/2 - 1648/147429) + 11759769043149622934890992/16381)/(233159481474211118318

Basically solve does not return a number but an expression, what you can do is use vpa to numerically solve the expression and pass it to max()

adding the the following lines after the solve should do the trick:

     t31 = vpa(t31)
     t41 = vpa(t41)

The final code would be:

a=1.6;
b=2.14;
c=2.06;
d=3.5;
AB=a
BC=b
CD=c
AD=d
xa=0;
ya=0;
yd=0;
xd=AD;
t2=0
syms t3 t4
for ii=1:36



    t2=t2+10
    f=a*cosd(t2) + b*cosd(t3) - c*cosd(t4) -3.5

    g=a*sind(t2) + b*sind(t3) - c*sind(t4)

    [t31, t41]=(solve(f==0,g==0,t3,t4))

    t31 = vpa(t31)
    t41 = vpa(t41)
    
    th3=double((max(t31)))

    th4=double((max(t41)))
    
    if exist('th3', 'var') 
        xb(ii)=AB*cosd(t2);
        yb(ii)=AB*sind(t2);
        xc(ii)=xb(ii)  + BC*cosd(th3)
        yc(ii)=yb(ii)  + BC*sind(th3)
        
        plot(xa,ya,'k.',xb,yb,'b.',xc,yc,'g.',xd,yd,'c.',LineWidth=3)
        hold on
        axis equal
        line_plot1 = plot([xa,xb(ii)], [ya,yb(ii)], 'b',LineWidth=3);
        line_plot2 = plot([xb(ii),xc(ii)], [yb(ii),yc(ii)], 'r',LineWidth=3);
        line_plot3 = plot([xc(ii),xd], [yc(ii),yd], 'm',LineWidth=3);
        line_plot4 = plot([xd,xa], [yd,ya], 'c',LineWidth=3);
        axis equal
        hold off
        pause(0.06)
    else
        continue
    end
end