If it is a question about the software then the very powerfull tool is ADAMS

http://www.mscsoftware.com/de/product/adams

Hope it helps you.

Thanks for your replys @Alexander, It's more of a mathemetical modelling question before software.@ Gianni, It does help. It gave me an idea on how to go about my problem.

Study

Eric N Moret: Dynamic Modeling and Control of a Car-Like Robot

Morning,

Though it does not cover the suspension. The model in my thesis may help guide you along the right path... http://theses.gla.ac.uk/508/

Kevin

Thank you so much for your valuable answers. http://cdn.intechopen.com/pdfs-wm/18889.pdf

Hi Bongani,

First of all I would like to know which software packages you have.

As mentioned earlier by Alexander MSC ADAMS can help you with GUI and building three dimensional car model. You can easily transfer data from another cad software and assign joint types according to your need. You can also link your controller (which I suppose you will design in MATLAB or Easy 5). The drawback/Advantage of using MSC ADAMS is that you cannot write your mathematical model. I believe knowing your mathematical model is an important task especially if you like to understand your system in detail. ADAMS is also extremely expensive :). If you would like to make your hand dirty you should try writing you code in MATLAB or some open source tool. I don't know if you heard about ACUMEN. It is a tool where you can write your system governing equations and control algorithms to simulate your system. You can check it on following link:

http://www.acumen-language.org/

I also recommend you to study my PhD thesis where you can find testing of multibody vehicle on different challenging scenarios.

Cheers,

Jawad

@Jawad, thanks for your reply... I am an Undergrad student so I have access to student versions of Matlab, Labview and SolidWorks. And I also was planning to use an open source tool for my simulation part of my thesis.

My topic is "Nonlinear State Estimation for a High Speed 4-wheel Robot Vehicle". Which I have to derive and validate a mathematical model which should include high frequency modes and nonlinearities. Also I need to design and simulate a nonlinear state estimator to accurately predict the states of the system (Unscented Kalman Filter). And lastly apply control and simulate the model.

hlw bongani,

i am having almost exactly same project in my MSc and i am struggling for last 7 months with the modeling and also simulation.as my background is EE so i have to study alot about the tire and wheel dynamics. after looking for a suitable software i have found that robotics toolbox by peter corke might be the best solution for us.

if u can do any progress please share ,so i will do.

Hi,

Bongani: I think in your case Matlab will work as you already have access. I suggest you to use matlab/simulink toolbox called Simmechanics. This tool can help you build your model with ease. How many degree of freedom your robot has?

Sharmin: Out of curiosity will you share the link to Peter Corke tool?

Cheers,

Jawad

http://petercorke.com/Robotics_Toolbox.html

Hi...@Jawad: All 6 DOF should be possible(X,Y,Z,YAW,PITCH,ROLL).

@Sharmin: Have a look the following 3 links.......1) http://cdn.intechopen.com/pdfs-wm/18889.pdf ......2) http://icel.me.pusan.ac.kr/automotive.php ..... (3) http://scholar.lib.vt.edu/theses/available/etd-05102011-112121/unrestricted/Currier_PN_D_2011_2.pdf

I think using system identification you get the maximally informative dynamic model that you can use for control design aim.

Hi Bongani,

I suggest you to try first simple planar model and then go ahead with more complex six degree of freedom model. Normally in case of planar vehicle model you consider X,Y and Yaw angles (generally speaking steering). Dynamics and control of planar vehicle is more intuitive and can help you learn the process from dynamics to control design. After finishing with planar you can give try to full vehicle model. Are you going to consider any suspension?

It is also important right from start that you look in which tire model you are going to use. There are plenty of tire models and designers normally select one based on dynamic phenomenon they requirement to capture. I think Pacejka "magic formula" tire model will be better starting point. You can google this tire model for further reference.

Best of luck.

Cheers,

Jawad

Hi Jawad. As you suggested I'll start with a bicycle model plus a brush model for tires before going into the full model. And yes I also have to consider suspension(active one as my supervisor suggested).

In this case, I suggest you to go forward with bicycle model without suspension first.

I am currently using the book 'Mobile Robotics' by Alonzo Kelly, which quite nicely covers the issue, for my Robotics lecture.

There is not much to add to the previous recommendations, starting with planar / bicycle model and then gradually adding another dimension and more aspects. This will give you something substantial to show to and discuss with your supervisor at a very early stage of the project.

is simemechanics is enough to simulate a bicycle model without suspension or with suspension@jawad

Hi,

If you would like to make an integrated ride and handling model you have two options: Newtonian dynamic model Lagrangian dynamic model. I Newtonian would be easier and its the reference:

https://www.researchgate.net/publication/228418033_Vehicle_modeling_for_integrated_control_design

I believes you need the to consider the bounce motion of wheel as presented in this paper. But, I think the a liner tire model would better in your case which you have plane to design a controller based on the model.

Article Vehicle modeling for integrated control design

Yes, Indeed @Sharmin

if u have time ,..learn LAGRANGIAN MODELING approach.

You must derive the simple realistic dynamic model of the vehicle first , based on that you can design its controller.

You can find a fully developed 14-DOF vehicle model in MATLAB with a time efficient suspension model and a reprogrammed full set MF tire model in http://alturl.com/shn64, in addition, this work formulated and solved the complicated optimal design and control problems of a 4-IWD electric race car on a given track in curvilinear coordinate system.

Have you been able to solve this problem?@Bongani.

For MPC and UKF, you necessarily do not need a complicated model. The needed model for the control design is the state of the system under consideration. You can also include some performance metric as constraint in the MPC design.

Euler Lagrange Method is commonly used for the modelling, also the information of systems dynamics is required.