
Na vic nezbyl cas 

- trochu to uprav ten clanek toho autora tam vynechej doplnim
v pondeli - ted to nemuzu najit a cas prave dosel:




Ljocho mozna by stalo za to tam zminit ze oba ty speeche budou v jednom
clanku:




Using linux with PHANToM.

  Currently Sensable company supports only two platforms with
its PHANToM device - Windows NT and IRIX. None of them however
had in the time we started the usage of this device real-time
extension to their OS. We wanted to use the PHANToM device
with the system which has good support for multitasking and
network and prefferable which supports hard read-time.
We decide to use RT-linux which is now very stable in its
release version 1.1. This version is designed for the linux 2.0.36
kernel. In the developenment is the version for 2.2 linux, however
in this moment this version is not usable.

Introduction to driver:

   Sensable doesn't support other platworms as well as doesn't provide
any documentation for their hardware to the end-user. So we have
quite hard times to create working driver. In the current state of
developenment our driver is being used only with PHANToM model A.
And supports both ISA and PCI version of this model.

Creating driver:

   In the first we have only one ISA Phantom. Thnakfully this version
has been accompanied not only with some Windows NT drivers and libraries,
but also small number of DOS programs were also atteched. We based
our driver on the informations we have found in these programs.
The situation was much more complicated with the PCI version.
For this version we have received only drivers for Windows NT and
also some driver for SGI IRIX. We have to literaly discover every
single thing about how to driver this version. However we are still
missing some possibly important thing about driving this version.
Needless to say both devices currently work with our device driver.

Driver posibilites:

   We our using the driver with the P200MMX processor - with this
machine we are achieving easily achiveng haptic-loop counts around
3000Hz as we found out the 1000Hz is to small to simulate trully
hard surfaces. On the other side our experiments shown there is
no human distiguishable difference when we use more then 5KHz frequncies,
but still this fruquencies are very useful when we want to elemintate
well know vibrations caused by applying incorrect forces in the
passing time - this is well disscused in this article

*********************

This article has used mathematical aparat to fight
with this problem. Our solutin to this problem is a lot more easier - we
our trying to use higher sampling frequencies and using smartness
of the driver to calculate small surrounding environment within itself.


Tunnel:

To get user think he is moving in some thin tunnel is not to much
hard. For this we are using precalculated point which belongs to the
tunnel and then the force is calculated with the vector arithmetics.
To calcute vector which is orinted to the middle of the tunel
we use this action:

We locate the neariest abscissa which is the part of the tunnel.
This way we obtain the vector which gives us orientation to which
way we should move the user and also this vector `VV' is used for
calulation of the project point of the current PHANToM position.

Equations:
P = Begin Of Tunnel's Abscissa
R = Actual User Position
X = Projection Of The Actual Position
    
vect = P - R


X we get from this equation:

R = P + ( (R - P) * (skalarni) VV ) / (VV * (skalarni) VV)  * VV


this way we obtain vector which is pushing the user to the center of
the tunnel - also by adding of multiplied normalized VV vector we
can push the user along the tunnel.



