|On this page|
This node is used as part of a crowd simulation workflow. By default it creates a group of agents randomly scattered on an area specified by the controls. If geometry is connected as input, it will scatter agents onto that surface. If the input geometry only contains points, an agent will be placed at each point position.
Source of the agent geometry and rig.
Asset node or a subnet in current session
The files are read from disk (need to be created using agent rop)
Location of agent asset or subnet node when Input is Subnet.
Location of agent files when Input is Disk.
Keep External References
Specifies whether the agent primitives should reference the agent
definition files on disk. If enabled, saving agents to a geometry file will
only store references to the agent definition files, instead of embedding
the agent definition in the
.geo file (which produces a
self-contained, but larger, geometry file).
Reload the files/nodes being referenced (Houdini caches needed agent files; when they're updated reloading them is required).
Animation clip that has the agent in neutral rest pose (no animation). This is used as reference when modifying agent rig.
Shape pattern of the crowd agents created.
Scatter randomly based on the number and area size.
Create a group based on rows and columns.
Size of the area that will be used to scatter agents onto.
Number of agents to create.
Used when randomly scattering agents.
Used to push the agents apart to prevent initial self collisions.
Name of the attribute that stores density to be used for scattering agents.
Multiplier on the density attribute.
Force Total Count
Controls number of agents scattered using density attribute. When off number will be based on the density values; when on it will use value specified by the Number parameter.
Name of the agent (this name is also used when referencing animation clips).
Name of the group that the agents will belong to.
Initial state that the agent will be when the simulation starts.
Color of the agent points and wireframe bones.
Size of the agent (sets
pscale attribute on the agent point).
Initial minimum speed for the agent (m/s). This may be adjusted during the simulation by the Animation Behavior settings on the Crowd Solver.
Initial maximum speed for the agent (m/s). This may be adjusted during the simulation by the Animation Behavior settings on the Crowd Solver.
Max Turn Rate
How fast the agent is allowed to turn when changing direction (deg/sec).
Maximum force that is allowed to affect the agent (values higher are clamped).
Initial velocity of the agent.
Initial direction the agents face (if initial velocity given it will override this).
Up vector to be used for the agents.
These controls allow randomization of parameters per agent.
Randomize Agent Name
This will randomize the agent name.
Agent Name Prefix
Prefix part of the agent name. The variations will be created by appending random numbers based on the variations/seed parameters (agent1/agent3/agent8/etc).
Number of agents names to create.
Random seed value to be used when creating per agent names
Type of distribution to be used when creating random values
Specifies which layer of the agents is displayed. All agents have a 'default' layer which is the skin geometry being deformed. The 'collision' layer is created based on the capture regions of the bones used to deform the agent.
Surface to be used as a base to scatter crowd agents onto. If only points are given, these will be used to create agents at each point position.
Agent primitives to be used in crowd simulation.