Umbra can use route planners to provide high-level path plans for simulated humans or vehicles. In real environments, routes are planned from various detail terrain models that can include obstacles, such as buildings, or plan routes within buildings. Vehicles or humans then typically use behavior or navigation algorithms to avoid unmodeled obstacles. Umbra’s 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 validate hierarchical control algorithms.
Typical route planning software uses terrain, coverage, and mobility to find optimal human and vehicle routes between specific goal points through complex terrains. Sandia’s constraint-based route planners add inter-visibility checks with known target positions to find routes that bring vehicles within site of the targets.
Path planning also includes the domain of robotic arm motion planners. Models have been 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 UNC’s V-Collide collision detection software was used to generate device motions between major pose positions (e.g., stowed, deployed, and inside window).
The Configuration Space Toolkit (CSTK or C-Space Toolkit) provides a software library that makes it easier to program motion planning, simulation, robotics, and virtual reality codes. The CSTK is included with Umbra distributions. See more in products.
Sandia has developed and continues to develop a spectrum of research and application-supporting path-planning software, ranging from RRT-based planning for robotics, to tactical path planners for supporting a variety of Sandia’s DOE, DoD, and HSD missions. The current generation of optimizing path planners integrated with Umbra employs a highly modular architecture that enables construction of planners for various state spaces and easily applies different cost functions and goal predicates to them.