Simulink numerical error?

Shobhit Gupta

2011 6 17

0 个回答

6 次查看（30 天）

显示更早的评论

0 个投票

Hi,

I am using simulink to control a robotic arm.

I am facing a strange problem. When I am running the simulation, there are 2 values which should be equivalent (their expressions are same). When I simout those 2 variables, their errors are in the order for e-4, which my system can't accept.

But when I run the same corresponding code in matlab (.m file), errors are in the order of e-18 or exactly 0.

What does this mean? Should I develop my control system in matlab file itself?

Will attach those 2 files if required.

Thanks in advance.

2 个评论
显示无隐藏无

Kaustubha Govind 2011-6-17

How are these expressions computed in Simulink (Simulink library blocks, Embedded MATLAB Block, Stateflow, etc.)? If using blocks, do you use the exact same blocks in both code-paths? What are the datatypes of your signals?

Shobhit Gupta 2011-6-18

Hey Kaustubha,

I am attaching my matlab files. You can check the values of Glin and G both in simulink (I've simout both variables in workspace) and the code as well.

Basically, they are same mathematical expression written in different ways. So ideally, their value should be same for same parameters i.e. mass(m), length(l) and angles (X(1),X(2),X(3)).

http://www.4shared.com/file/IVDqiMMy/Control_lislotine_norob_damp_1.html?

[control_lislotine_norob_damp_17jun_lin_compare.mdl is my control system in simulink.]

http://www.4shared.com/file/NrQa0OId/y_dynamics_check.html?

[While y_dynamics_check is the matlab file which calculates value of Glin and G in simulink for a series of angles]

http://www.4shared.com/file/FnPbK4I_/Y_check.html?

[Same as previous, but just checks it for one angle value].

Thanks in advance

请先登录，再进行评论。

请先登录，再回答此问题。

请先登录再关注