torcs-users Mailing List for TORCS - The Open Racing Car Simulator

Hey guys,

I'm not sure if this is the proper place to put this, but I put the apidoc
for TORCS 1.3.7 up on https://quinten.space/torcs-1.3.7-apidoc/. I guess it
can be useful to someone who just wants to quickly reference it when giving
help or just wanting to quickly look something up.

Quinten

Hello! I have several questions, but this one must be resolved first (and probably added to FAQ):

How can I search this mail-list archive using AND logic? (sourceforge uses OR). No options seems to be provided. (DDG, or google, will search among all sourceforge projects = can't filter them out = bad)

Alternatively, is there a more accessible (searchable) knowledge base?

>As for DEs, I would prefer s-domain transfer functions
>made from free-body diagrams, which are then
>converted to z-transforms before coding.

Great. Many engineers (op-amp feedback network compensation, automatic control theory) will find it easy to understand car dynamics model then. z-xform is essentially Fourier, and everybody knows Fourier. Great. So... what section did you say of the TORCS dynamics model manual describes how it's all applied? :)

Your link is now alive. The server was down back then (I made sure it was not only for me).

This link works for me:
http://phors.locost7.info/files/Beckman_-_The_Physics_of_Racing.pdf

You can download the entire site by:
1) Create a directory called Physics_of_Racing
2) enter this directory
3) open a terminal window in this directory
4) execute the following from the terminal window while connected to the internet:
wget -E -k -p -r -e robots=off http://phors.locost7.info/files/

----- h6mgos8bn6fb---  wrote:
> >Differential equations are used...
> May be we are reading different versions. Could you give updated URL,
> page #, may be Eq. # - where DEs are being used?
> (I mean DEs as equations of motion, let's not count v=dx/dt as DE :) )
> 
I think we are looking at the same document.
(Beckman - The Physics of Racing)
Reformatted 25 April 2008, 144 pages.
BTW the link in the last email works for me.

As for DEs, I would prefer s-domain transfer functions
made from free-body diagrams, which are then
converted to z-transforms before coding.

> My suggestion "Please, consider writing a proper documentation"
> was addressed to whoever may be willing to do it, not only you.
> Someone among the TORCS authors.
> 
> >torcs/src/modules/simu/simuv2/transmission.cpp
> Thank you. And this is IMO a good example where
> the proper documentation is needed. There are only
> few implementational comments. A new-comer have
> to read the source and understand its logic to derive
> the clutch model - i.e. understand general ideas from
> gory details. While it (ideally) should be vice versa.

See:
https://sourceforge.net/projects/torcs/files/api-docs/1.3.7/torcs-1.3.7-apidoc.tar.bz2/download
Also:
http://torcs.sourceforge.net/
for FAQs etc.

TORCS is not easy to wade through, but it can be done.
My biggest irritation is variables being copied to structures,
and structures being copied to structures, then passed as
pointers in some sort of interface between C and C++
(or so it seems to me).

I am trying to find a way to record user data in TORCS. With this I mean
the user inputs (accelerator position, brake pedal position, steering
angle,...) as well as the users car position (current track segment,
distance from middle, speed, de/acceleration, direction,...). Ideally I am
trying to achieve something that outputs a CSV/text file (or anything else
really) for every completed lap.


Thanks a lot in advance.

<http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
Virus-free.
www.avg.com
<http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

>Differential equations are used...
May be we are reading different versions. Could you give updated URL, page #, may be Eq. # - where DEs are being used? (I mean DEs as equations of motion, let's not count v=dx/dt as DE :) )

My suggestion "Please, consider writing a proper documentation" was addressed to whoever may be willing to do it, not only you. Someone among the TORCS authors.

>torcs/src/modules/simu/simuv2/transmission.cpp
Thank you. And this is IMO a good example where the proper documentation is needed. There are only few implementational comments. A new-comer have to read the source and understand its logic to derive the clutch model - i.e. understand general ideas from gory details. While it (ideally) should be vice versa.

----- Gal Israeli wrote:
> Hi,
> I've been using your great simulator for testing an algoirthm.
> 
> We are trying to find out where we can get the current car speed every loop
> simulation, and no success so far. can you please direct us to the file?
> 
> Also after that I need to publish a new gas to the car.
> 
> Thank you

See:
car->pub.speed
in:
torcs/src/drivers/human/human.cpp :: static void common_drive(int index, tCarElt* car, tSituation *s)

car->pub.speed can be used in a similar fashion anywhere.
Just do it the same as in human.cpp

To find pub.speed in the source code, type:
find . -type f -print0 | xargs -0 grep -i pub.speed
in a terminal from:
torcs/src/modules
or elsewhere if desired.

----- h6mgos8bn6fb--- wrote:
> Hello!
> 
> From: David Savinkoff - 2016-03-23 22:29:44
> ----- Carlos Massera FIlho wrote:
> > Hi,
> > 
> > Is there any reference material for the equations used to implement the vehicle
> > dynamics?
> > I’m familiar with most of them but curious to what goes into the torcs source code.
> > I’m trying to read the source directly but the dynamics part is very convoluted.
> > 
> > Thanks,
> > Carlos Massera Filho
> Hi,
> The following sheds some light on TORCS:
> http://phors.locost7.info/files/Beckman_-_The_Physics_of_Racing.pdf
> 
> 
> That pdf is >25 years old (1991) and is very general and very simplistic.
I found this documentation to be concise, accurate and timeless.

> Physics level is high-school ("centrifugal" force).
There is a chapter explaining No centrifugal force.

> Math level essentially too (no differential equations).
Differential equations are used... and a smart high school student
might be able to grasp them because the documentation
is excellent.

> Only one - incidental mention of the word "clutch".
> (the server linked appears to be down, I used
> dodds.cc/chad/garage/media/files/physics_of_racing.pdf)
> 
> Please, for comparison, and reference, have a look at
> Flight Dynamics Model documentation:
> http://jsbsim.sf.net/JSBSimReferenceManual.pdf
> - used in free FlightGear flight sim.
Looks good indeed.

> 
> Please, consider writing a proper documentation.
I'm am not the author of this documentation.

> 
> (I wanted to see how well clutch and gearbox are implemented...)
See:
torcs/src/modules/simu/simuv2/transmission.cpp
in the source code.

The clutch exists, but is trivially implemented.
The gearbox appears to work well (never caused me concern).

Hello!

From: David Savinkoff <dsavnkff@te...> - 2016-03-23 22:29:44
----- Carlos Massera FIlho wrote:
> Hi,
> 
> Is there any reference material for the equations used to implement the vehicle
> dynamics?
> I’m familiar with most of them but curious to what goes into the torcs source code.
> I’m trying to read the source directly but the dynamics part is very convoluted.
> 
> Thanks,
> Carlos Massera Filho
Hi,
The following sheds some light on TORCS:
http://phors.locost7.info/files/Beckman_-_The_Physics_of_Racing.pdf


That pdf is >25 years old (1991) and is very general and very simplictic. Physics level is high-school ("centrifugal" force). Math level essentially too (no differential equations). Only one - incidental mention of the word "clutch". (the server linked appears to be down, I used dodds.cc/chad/garage/media/files/physics_of_racing.pdf)

Please, for comparison, and reference, have a look at Flight Dynamics Model documentation: http://jsbsim.sf.net/JSBSimReferenceManual.pdf - used in free FlightGear flight sim.

Please, consider writing a proper documentation.

(I wanted to see how well clutch and gearbox are implemented...)

Hi,
I've been using your great simulator for testing an algoirthm.

We are trying to find out where we can get the current car speed every loop
simulation, and no success so far. can you please direct us to the file?

Also after that I need to publish a new gas to the car.

Thank you