A 3D agent used to pathfind to a position while avoiding static and dynamic obstacles. The calculation can be used by the parent node to dynamically move it along the path. Requires navigation data to work correctly.
Dynamic obstacles are avoided using RVO collision avoidance. Avoidance is computed before physics, so the pathfinding information can be used safely in the physics step.
[b]Note:[/b] After setting the [member target_position] property, the [method get_next_path_position] method must be used once every physics frame to update the internal path logic of the navigation agent. The vector position it returns should be used as the next movement position for the agent's parent node.
Returns the distance to the target position, using the agent's global position. The user must set [member target_position] in order for this to be accurate.
Returns this agent's current path from start to finish in global coordinates. The path only updates when the target position is changed or the agent requires a repath. The path array is not intended to be used in direct path movement as the agent has its own internal path logic that would get corrupted by changing the path array manually. Use the intended [method get_next_path_position] once every physics frame to receive the next path point for the agents movement as this function also updates the internal path logic.
Returns the reachable final position of the current navigation path in global coordinates. This position can change if the agent needs to update the navigation path which makes the agent emit the [signal path_changed] signal.
Returns the [RID] of the navigation map for this NavigationAgent node. This function returns always the map set on the NavigationAgent node and not the map of the abstract agent on the NavigationServer. If the agent map is changed directly with the NavigationServer API the NavigationAgent node will not be aware of the map change. Use [method set_navigation_map] to change the navigation map for the NavigationAgent and also update the agent on the NavigationServer.
Returns the next position in global coordinates that can be moved to, making sure that there are no static objects in the way. If the agent does not have a navigation path, it will return the position of the agent's parent. The use of this function once every physics frame is required to update the internal path logic of the NavigationAgent.
Returns [code]true[/code] if the end of the currently loaded navigation path has been reached.
[b]Note:[/b] While true prefer to stop calling update functions like [method get_next_path_position]. This avoids jittering the standing agent due to calling repeated path updates.
Returns true if [member target_position] is reached. It may not always be possible to reach the target position. It should always be possible to reach the final position though. See [method get_final_position].
Based on [param value], enables or disables the specified layer in the [member avoidance_layers] bitmask, given a [param layer_number] between 1 and 32.
</description>
</method>
<methodname="set_avoidance_mask_value">
<returntype="void"/>
<paramindex="0"name="mask_number"type="int"/>
<paramindex="1"name="value"type="bool"/>
<description>
Based on [param value], enables or disables the specified mask in the [member avoidance_mask] bitmask, given a [param mask_number] between 1 and 32.
Based on [param value], enables or disables the specified layer in the [member navigation_layers] bitmask, given a [param layer_number] between 1 and 32.
Replaces the internal velocity in the collision avoidance simulation with [param velocity]. When an agent is teleported to a new position this function should be used in the same frame. If called frequently this function can get agents stuck.
If [code]true[/code] the agent is registered for an RVO avoidance callback on the [NavigationServer3D]. When [member velocity] is set and the processing is completed a [code]safe_velocity[/code] Vector3 is received with a signal connection to [signal velocity_computed]. Avoidance processing with many registered agents has a significant performance cost and should only be enabled on agents that currently require it.
A bitfield determining the avoidance layers for this NavigationAgent. Other agents with a matching bit on the [member avoidance_mask] will avoid this agent.
A bitfield determining what other avoidance agents and obstacles this NavigationAgent will avoid when a bit matches at least one of their [member avoidance_layers].
The agent does not adjust the velocity for other agents that would match the [member avoidance_mask] but have a lower [member avoidance_priority]. This in turn makes the other agents with lower priority adjust their velocities even more to avoid collision with this agent.
If [member debug_use_custom] is [code]true[/code] uses this rasterized point size for rendering path points for this agent instead of global point size.
The height of the avoidance agent. Agents will ignore other agents or obstacles that are above or below their current position + height in 2D avoidance. Does nothing in 3D avoidance which uses radius spheres alone.
If [code]true[/code], and the agent uses 2D avoidance, it will remember the set y-axis velocity and reapply it after the avoidance step. While 2D avoidance has no y-axis and simulates on a flat plane this setting can help mitigate the most obvious clipping on uneven 3D geometry.
A bitfield determining which navigation layers of navigation regions this agent will use to calculate a path. Changing it during runtime will clear the current navigation path and generate a new one, according to the new navigation layers.
The distance threshold before a path point is considered to be reached. This allows agents to not have to hit a path point on the path exactly, but only to reach its general area. If this value is set too high, the NavigationAgent will skip points on the path, which can lead to leaving the navigation mesh. If this value is set too low, the NavigationAgent will be stuck in a repath loop because it will constantly overshoot or undershoot the distance to the next point on each physics frame update.
The height offset is subtracted from the y-axis value of any vector path position for this NavigationAgent. The NavigationAgent height offset does not change or influence the navigation mesh or pathfinding query result. Additional navigation maps that use regions with navigation meshes that the developer baked with appropriate agent radius or height values are required to support different-sized agents.
The maximum distance the agent is allowed away from the ideal path to the final position. This can happen due to trying to avoid collisions. When the maximum distance is exceeded, it recalculates the ideal path.
The radius of the avoidance agent. This is the "body" of the avoidance agent and not the avoidance maneuver starting radius (which is controlled by [member neighbor_distance]).
Does not affect normal pathfinding. To change an actor's pathfinding radius bake [NavigationMesh] resources with a different [member NavigationMesh.agent_radius] property and use different navigation maps for each actor size.
The distance threshold before the final target point is considered to be reached. This allows agents to not have to hit the point of the final target exactly, but only to reach its general area. If this value is set too low, the NavigationAgent will be stuck in a repath loop because it will constantly overshoot or undershoot the distance to the final target point on each physics frame update.
The minimal amount of time for which this agent's velocities, that are computed with the collision avoidance algorithm, are safe with respect to other agents. The larger the number, the sooner the agent will respond to other agents, but less freedom in choosing its velocities. A too high value will slow down agents movement considerably. Must be positive.
The minimal amount of time for which this agent's velocities, that are computed with the collision avoidance algorithm, are safe with respect to static avoidance obstacles. The larger the number, the sooner the agent will respond to static avoidance obstacles, but less freedom in choosing its velocities. A too high value will slow down agents movement considerably. Must be positive.
If [code]true[/code], the agent calculates avoidance velocities in 3D omnidirectionally, e.g. for games that take place in air, underwater or space. Agents using 3D avoidance only avoid other agents using 3D avoidance, and react to radius-based avoidance obstacles. They ignore any vertex-based obstacles.
If [code]false[/code], the agent calculates avoidance velocities in 2D along the x and z-axes, ignoring the y-axis. Agents using 2D avoidance only avoid other agents using 2D avoidance, and react to radius-based avoidance obstacles or vertex-based avoidance obstacles. Other agents using 2D avoidance that are below or above their current position including [member height] are ignored.
Sets the new wanted velocity for the agent. The avoidance simulation will try to fulfill this velocity if possible but will modify it to avoid collision with other agents and obstacles. When an agent is teleported to a new position, use [method set_velocity_forced] as well to reset the internal simulation velocity.
- [code]link_entry_position[/code]: If [code]owner[/code] is available and the owner is a [NavigationLink3D], it will contain the global position of the link's point the agent is entering.
- [code]link_exit_position[/code]: If [code]owner[/code] is available and the owner is a [NavigationLink3D], it will contain the global position of the link's point which the agent is exiting.
Emitted once per loaded path when the agent internal navigation path index reaches the last index of the loaded path array. The agent internal navigation path index can be received with [method get_current_navigation_path_index].
Emitted once per loaded path when the agent's global position is the first time within [member target_desired_distance] to the [member target_position].
Notifies when the collision avoidance velocity is calculated. Emitted when [member velocity] is set. Only emitted when [member avoidance_enabled] is true.
