The 'diff()' function is running incorrectly in this program (Error using diff Difference order N must be a positive integer scalar)

T_stop=10;
T_sample = 0.002;
x0 = [0;0;1/2*pi;0];
options=odeset('RelTol',1.0e-6,'AbsTol',1.0e-6,'BDF','on');

[t,x]=ode45(@(t,x) RIP_Sysm(t,x),[0:T_sample:T_stop],x0,options); % ode45解常微分方程

figure(1)
plot(t,x(:,1));title('时间-滑块A角度');
xlabel('t[s]');ylabel('\theta_A[rad]')
figure(2)
plot(t,x(:,3));title('时间-滑块D角度');
xlabel('t[s]');ylabel('\theta_D[rad]')

function dx = RIP_Sysm(t,x) 
%% Constants    
ml = 15;
Rp = 42.31;
dl = 127;
H0 = 124.8;
R = 124.24;
L = 147.72;
d = 9.4;
alfa = 12.83*pi/180;
Ix = 1000;
Iy = 1000;
Iz = 1000;
J = 10;

A(x) = (dl^2 + H0^2 + Rp^2 + R^2 - L^2 - 2 * dl * R * sin((x(1) - x(3))/2)) / (2 * Rp);
B(x) = (R^2 * (cos((x(1) - x(3))/2))^2 + H0^2)^0.5;
C(x) = R * cos((x(1) - x(3)) / 2);
b(x) = acos( C / B ) + alfa - acos( A / B );
a(x) = (x(1) - x(3)) / 2;

db_A = diff(b(x),x(1)); 
db_D = diff(b,x(3));
da_A = diff(a,x(1));
da_D = diff(a,x(3));

Answer 1

[VBBV]

T_stop=10;
T_sample = 0.002;
x0 = [0;0;1/2*pi;0];
options=odeset('RelTol',1.0e-6,'AbsTol',1.0e-6,'BDF','on');
[t,x]=ode45(@(t,x) RIP_Sysm(t,x),[0:T_sample:T_stop],x0,options); % ode45解常微分方程
figure(1)
plot(t,x(:,1));title('时间-滑块A角度');
xlabel('t[s]');ylabel('\theta_A[rad]')

figure(2)
plot(t,x(:,3));title('时间-滑块D角度');
xlabel('t[s]');ylabel('\theta_D[rad]')

function dx = RIP_Sysm(t,x) 
%% Constants  
   
   ml = 15;
   Rp = 42.31;
   dl = 127;
   H0 = 124.8;
   R = 124.24;
   L = 147.72;
   d = 9.4;
   alfa = 12.83*pi/180;
   Ix = 1000;
   Iy = 1000;
   Iz = 1000;
   J = 10;
   
   A = (dl^2 + H0^2 + Rp^2 + R^2 - L^2 - 2 * dl * R * sin((x(1) - x(3))/2)) / (2 * Rp);
   B = (R^2 * (cos((x(1) - x(3))/2))^2 + H0^2)^0.5;
   C = R * cos((x(1) - x(3)) / 2);
   b = real(acos( C ./ B ) + alfa - acos( A ./ B ));
   a = (x(1) - x(3)) / 2;
   db_A = (b-x(1));
   db_D = (b-x(3));
   da_A = (a-x(1));
   da_D = (a-x(3));
   dx = [db_A;db_D; da_A;da_D];
end

[VBBV]

Use diff when there are vectors instead of scalars. In this case, they are scalars, you can take the difference using a '-' math operator. The 2nd argument in the diff function has been used incorrectly. It represents the number of times the derivative or difference to be done.

[晨名]

Thank you very much for your answer! But in this case, I want to use the 'diff ()' to take the partial derivative.

[VBBV]

Ok. Then as Walter suggested you need to declare a symbolic parameter that can be used in diff function.

Answer 2

[Walter Roberson]

There are two major diff() functions.

You thought you were using calculus derivative, diff . However, that diff() function only applies if at least one of the parameters passed to diff() is symbolic or symfun

The diff() that applies in most cases where parameters are not symbolic, is diff which is approximately diff(x) = x(2:end) - x(1:end) . When you use that diff() function, a non-empty second parameter must be a positive integer scalar indicating the number of times that the subtraction operator is to be repeated.

The values you are passing in to diff() are purely numeric, so you are getting the successive-differences version of diff()

I suggest that you rewrite your code to create your equations using the Symbolic Toolbox, and then you follow the workflow shown in first example for odeFunction in order to generate a function handle that can be used with ode45()