The proposed motion planner evaluates motions of the robots
using a physical simulation. Therefore, a map of the environment (described e.g. as a triangle mesh) is required to be able to compute
a motion plan. The motion planner is run before the mission and
its result, which is a sequence of motion primitives, is then sent
to the robot for execution. During the execution of the plan, a
central module of the robot distributes the control signals to all
other modules. This can be realized using a simple processor and
therefore the robots do not need to be equipped with powerful
computers to realize the motions.