Logo TU Ilmenau


Prof. Dr. sc. techn. Beat Brüderlin


Telefon +49 3677 694119

E-Mail senden


Computer Animation


MoPS3D is a computer animation system featuring Newtonian physics, determined by masses, springs, forces (gravitation, friction, actuator motors), etc. Euler or Runge-Kutta numerical integration is used to solve the differential equations. Collision detection is employed to determine interaction between bodies.


In addition to the laws of physics, control loops can also be simulated in MoPS3D. Sensors measure, for instance, the position, pose, or speed of objects. PID controllers can be used to determine the forces of actuators, for instance, to automatically stabilize a vehicle.


The following example shows the model of a «hovercraft.» This craft is controlled by control loops that convert values measured by the sensors into control data for the engines.


Drawing of the hovercraft


This piece of simulated hardware features four engines operating horizontally, four operating vertically, and one for forward / backward movement.


The simulation system allows connection of objects with springs. Here, a load has been attached to the hovercraft. The load exerts forces on the hovercraft and puts additional demands on its control system.


A more complex example for control-based physical animation is this one-legged hopping robot (first introduced to the computer animation community by Hodgins and Raibert at Siggraph 1991). Here, a simple PID-controller no longer suffices to keep the robot in balance.

(image to follow)


The robot needs to keep hopping in place, or back and forth to keep from falling. A control algorithm, simulating a finite-state machine with flight, landing and take-off phases is used in conjunction with a PID controller, determining the jumping force of the leg. The robot, within limits, can be steered to follow a user controlled path. If the limits are exceeded, the robot may loose control and stumbles. Control-based animation is especially suitable to create natural-looking, physically-correct computer animations.


Autonomous Actors

Another research direction in computer animation is autonomous actors. The behavior of autonomous actors is determined by:

  • The environment
  • Events defining interaction with the environment and other actors
  • Rules which determine the (re)action to those events


As an example for autonomous actors we experimented with traffic simulation. Each vehicle in this environment reacts to traffic lights, other vehicles near-by, etc. The vehicle is autonomous and acts according to traffic rules, the laws of physics, and its personality.