Page 1 of 1

simulation Dynamics explained better

Posted: 05 Jan 2019, 00:58
by luberdis
Hi I have a problem with vrep. I try to simulate a differential robot that drives to a specific point of the surface. The algorithm I use is very common and it is described in this book at page 94 on book at this link I have here

https://www.dropbox.com/s/mr7euyfyhatc9 ... .pdf?dl=0

I implement the code in Matlab and it works pretty fine and you can see the result and its simulated trajectory Image and here you can see if you want the implemented Matlab code here https://www.dropbox.com/s/pndu60m4qepbh ... ove.m?dl=0

Then I try to use the same code in vrep and I create a script at the simulated model that uses this algorithm. I upload the vrep file to check it if you want
https://www.dropbox.com/s/1yavvph02hdxt ... a.ttt?dl=0

The problem is that even it gets really close to the target area, I think the model slips and it misses the target so it gets into an endless loop
Image

Could you please help me about this physics problem and suggest me something I tried the static as you told me before but it did not move at all. I tried different engines as you wrote to another member but nothing.

I would be really grateful if you share your experience to me

Cheers

Re: simulation Dynamics explained better

Posted: 08 Jan 2019, 15:13
by coppelia
When you use a physics engine, it is extremely rare that you get the same results as the theory tells. The differences are due to the way a physics engine works and how it does its computation and simplifications. Each physics engine can display different behaviours.
In case of a mobile robot, you will always have differences because of (among others) how friction is modeled. Also, if you have badly designed your dynamic elements, then you can see effects such as small vibrations, slipping, etc.

Make sure to carefully read this page, in order to exclude all effects that are linked to a bad design.

Cheers

Re: simulation Dynamics explained better

Posted: 14 Mar 2019, 19:36
by luberdis
Well the problem was that the wheels I used at this example were the same material as the main body. (According to the documentation tutorials). So suddenly I tried the already implemented wheels that I modified according to my needs (so I just used only the black material which I suppose simulate the rubber friction properties). Maybe the tutorials should be reconsidered!

Cheers