Environmental Engineering Reference
In-Depth Information
would on the actual hardware and the simulated calls move the robot in the
environment and return the same sensor results the real system would return.
If the simulation is good enough, the control system can then be placed in the
real robot and run.
5.7.3 Simulation Servers
A more advanced technique is to use a separate simulation server for com-
ponents being simulated and have the component being tested interact with
the simulation server using a local area network. The primary advantage of
this approach is that the simulation and test component run on different
hardware, and therefore, each can have free access to needed resources. The
simulation can be at an arbitrary and appropriate level of fidelity. If a high
fidelity simulation is desired, very fast hardware can be used to simulate it.
The software being tested can be run in an environment very close to the
real system with the network interfaces being the only potential difference. If,
for example, the planned hardware for the control software being tested is an
unusual flight computer, then that exact computer can be used. It does not
affect the simulation server's hardware.
A simulation server has an additional advantage: it can usually accommo-
date multiple simulated entities in the same simulated environment - a big
advantage when testing cooperative autonomy systems. Though it is some-
what uncommon, software library techniques can be used in simulations ac-
commodating multiple simulated entities.
5.7.4 Networked Simulation Environments
Another software simulation technique uses a network of computers to simu-
late multiple entities in a single shared simulation environment. The focus in
designing these systems is on ecient protocols that allow an accurate simu-
lation in the shared environment. The primary advantage of this approach is
that it allows very large and complex systems to be simulated.
The distributed interactive simulation (DIS) system funded by DARPA
and the military is an example of a networked simulation environment
[ 6 , 40 , 48 ]. This very successful project built a training system that allowed
hundreds of simulated entities (tanks, planes, helicopters, missiles, etc.) to
interact in a realistic battlefield simulation. The thrust of this effort was orig-
inally to have human-controlled simulated hardware engage in interactions.
The continuations of DIS have developed in many areas. Work relative to co-
operative autonomy simulations is intended to bring an understanding of how
real hardware and simulated entities can interact together in the same simu-
lated/real world. Such an understanding would support the development and
testing of real spacecraft or robots interacting with simulated ones in mutual
cooperation.
Search WWH ::




Custom Search