Boids particle systems are controlled by a limited artificial intelligence, which can be programmed to follow basic rules and behaviors. They are ideal for simulating flocks, swarms, herds and schools of various kind of animals, insects and fishes or predators vs. preys simulations. They can react on the presence of other objects and on the members of their own system. Boids can handle only a certain amount of information, therefore the sequence of the Behavior settings is very important. In certain situations only the first three parameter are evaluated.
To view the panel to the right, add a Particle System of type Emitter and look in the middle area of the Particle System tab.
Boids try to avoid objects with activated Deflection. They try to reach objects with positive Force fields, and fly from objects with negative Force fields. The objects have to share one common layer to have an effect. It is not necessary to render this common layer, so you may use invisible influences.
Boids can different physics depending on whether they are in the air, or on land (on collision object).
- Allow Flight
Allow boids to move in the air.
- Allow Land
Allow boids to move on land.
- Allow Climbing
Allow boids to climb goal objects.
- Max Air Speed
Set the Maximum velocity in the air.
- Min Air Speed
Set the Minimum velocity in the air.
- Max Air Acceleration
Lateral acceleration in air, percentage of the max velocity (turn). Defines how fast a boid is able to change direction.
- Max Air Angular Velocity
Tangential acceleration in air, percent 180 degrees. Defines how much the boid can suddenly accelerate in order to fulfill a rule.
- Air Personal Space
Radius of boids personal space in air. Percentage of particle size.
- Landing Smoothness
How smoothly the boids land.
- Max Land Speed
Set the Maximum velocity on land.
- Jump Speed
Maximum speed for jumping.
- Max Land Acceleration
Lateral acceleration on land, percent of max velocity (turn). Defines how fast a boid is able to change direction.
- Max Land Angular Velocity
Tangential acceleration on land, percent 180 degrees. Defines how much the boid can suddenly accelerate in order to fulfill a rule.
- Land Personal Space
Radius of boids personal space on land. Percentage of particle size.
- Land Stick Force
How strong a force must be to start effecting a boid on land.
Amount of rotation around velocity vector on turns. Banking of 1.0 gives a natural banking effect.
Amount of rotation around side vector.
Boid height relative to particle size.
Initial boid health when born.
Maximum caused damage per second on attack.
Boid will fight this time stronger than enemy.
Accuracy of attack.
Maximum distance of which a boid can attack.
This list view allows you to set up other particle systems to react with the boids.
A data ID to select an object with a particle system set on.
Index of the Object’s particle system as set in the list view in the particle panel.
Setting the type to Enemy will cause the systems to fight with each other.
Will make the systems work together.
Will not cause them to align or fight with each other.
As mentioned before, very much like Newtonian particles, Boids will react to the surrounding deflectors and fields, according to the needs of the animator.
Boids will try to avoid deflector objects according to the Collision rule’s weight. It works best for convex surfaces (some work needed for concave surfaces).
For boid physics, spherical force fields define the way the objects having the field are seen by others. So a negative force field (on an object or a particle system) will be a predator to all other boids particle systems, and a positive field will be a goal to all other boids particle systems.
These effectors could be predators (negative Strength) that boids try to avoid, or targets (positive Strength) that boids try to reach according to the (respectively) Avoid and Goal rules’ weights. Force’s effective Strength is multiplied by the actual relevant weight (e.g. if either Strength or Goal is null, then a flock of boids will not track a positive force field).
The Boid Brain panel controls how the boids particles will react with each other.
The boids’ behavior is controlled by a list of rules. Only a certain amount of information in the list can be evaluated. If the memory capacity is exceeded, the remaining rules are ignored.
The rules are by default parsed from top-list to bottom-list (thus giving explicit priorities), and the order can be modified using the little arrows buttons on the right side.
Seek goal (objects with a force field and positive Strength).
Predict target’s movements.
Avoid “predators” (objects with force field and negative Strength).
Predict target’s movements.
- Fear Factor
Avoid object if danger from it is above this threshold.
- Avoid Collision
Avoid objects with activated Deflection.
Avoid collision with other boids.
Avoid collision with deflector objects.
- Look Ahead
Time to look ahead in seconds.
Boids move away from each other.
Copy movements of neighboring boids, but avoid each other.
- Follow Leader
Follows a leader object instead of a boid.
Distance behind leader to follow.
Follow the leader in a line.
- Average Speed
Maintain average velocity.
Percentage of maximum speed.
How fast velocity’s direction is randomized.
How much velocity’s Z component is kept constant.
Move toward nearby boids.
- Fight Distance
Attack boids at a maximum of this distance.
- Flee Distance
Flee to this distance.
There are three ways control how rules are evaluated.
All rules are averaged.
A random rule is selected for each boid.
Uses fuzzy logic to evaluate rules. Rules are gone through top to bottom. Only the first rule that affect above the fuzziness threshold is evaluated. The value should be considered how hard the boid will try to respect a given rule (a value of 1.000 means the Boid will always stick to it, a value of 0.000 means it will never). If the boid meets more than one conflicting condition at the same time, it will try to fulfill all the rules according to the respective weight of each.
Please note that a given boid will try as much as it can to comply to each of the rules he is given, but it is more than likely that some rule will take precedence on other in some cases. For example, in order to avoid a predator, a boid could probably “forget” about Collision, Crowd and Center rules, meaning that “while panicked” it could well run into obstacles, e.g. even if instructed not to, most of the time.
As a final note, the Collision algorithm is still not perfect and in research progress, so you can expect wrong behaviors at some occasion. It is worked on.