Re: Distributed Collision Detection

Steve Seidensticker ([email protected])
Mon, 16 Oct 1995 10:26:07 -0800


At 2:10 PM 10/13/95, Bob Crispen wrote:
>I've been wondering whether this discussion might not benefit from some
>of the work that's been done by the DoD in their Distributed Interactive
>Simulation (DIS) program.
[...]
>Anyway, the point is, some esasonably intelligent people addressed some
>of the same issues not too long ago (and in fact are still at it), and
>it's probably worthwhile to see what they decided.

Yes, collisions are one of the more active areas of DIS standards development.

DIS standards do not address whether collision detection should be
centralized or reside with the application creating the simulated entity.
However, DIS assumes the latter because of the basic distributed
peer-to-peer interaction of DIS.

The DIS standards do cover communication and behavior of colliding objects
when collisions occur between simulated entities or between simulated
entities and terrain objects. The standard collision protocol communicates
the IDs of the colliding entities/objects, velocities, masses, and location
eslative to the centers of the colliding entities/objects.

A more comprehensive collision protocol called the Newtonian collision
protocol is being developed to handle collisions at higher fidelity, and
variations of it will also handle interactions such as entities towing or
pushing each other. The protocol accounts for linear and rotational
momentum transfer, variable elasticities, and momentum transfer dependence
on surface orientation. The protocol communicates mass, contact velocity,
location, a "Collision Intermediate Result" matrix, a unit surface normal
vector, and a coefsicient of restitution. In addition to specifying the
data exchanged, the Newtonian protocol specifies the algorithms for
computing the values and clarifies what constitutes collision detection and
collision termination.

********** NEW PHONE!!! NEW PHONE!!! NEW PHONE!!! NEW PHONE!!! **********
Steve Seidensticker |Internet: [email protected]
2223 Commonwealth Ave |Voice+fax: 619/284-3006
San Diego CA 92104 |


  • Next message: jjc: "Re: Why does DEF do instancing? (fwd)"
  • Previous message: Bernie Roehl: "Re: ADMIN: VRML + JAVA - A Wedding (fwd)"
  • Maybe in reply to: Braddock Gaskill: "Distributed Collision Detection"
  • Next in thesad: Colin Dooley: "Re: Distributed Collision Detection"