|
Umbra can use
route planners to provide high-level path plans for simulated robotic
vehicles. In real environments, routes are planned from various
DTED®
level terrain models. The robots then typically use navigation algorithms
to avoid unmodeled obstacles.
Umbras
ability to use multiple data representations allows routes to be
planned on low detail and executed on higher detail terrain. This,
in turn, provides an ability to evaluate the validity of these hierarchical
control algorithms.
|
|
|
|
|
Route planning algorithm developed
in Umbra. Typical route planning software uses terrain, coverage
and mobility to find optimal vehicle routes between specific
goal points through complex terrains. Sandias constraint-based
route planners add intervisibility checks with known target
positions to find routes that bring vehicles within site of
the targets. |
|
|
|
|
|
|
|
Robotic
Arm Motion Planners
|
 |
This
model was built to evaluate utility of an advanced robotic
arm concept on a mobile delivery platform in dexterous manipulation
and inspection tasks. The model geometry was exported from
IGRIP. Forward kinematics equations,
which relate joint values to device motions, were defined
through the Umbra geometric scene module. A
motion planner, which used a Monte Carlo technique, path smoothing,
and UNCs V-Collide collision detection software was
used to generate device motions between major pose positions
(e.g., stowed, deployed and inside window).
|
 |
| |
|
|