Threat Manager
The Threat Manager is an opt-in system that enables threat assessment and threat-based target selection. This evaluation model is commonly used in RPG-style gameplay, where an agent may choose between multiple targets based on factors such as damage dealt or deliberate threat generation.
When enabled, the Threat Manager evaluates threat independently of perception strength and broadcasts the highest-threat target back into the perception system, allowing standard awareness and decision logic to react accordingly.
Threat vs Awareness
Threat and awareness are evaluated independently and serve different purposes. Awareness determines whether a target is known and can be acted upon. Threat determines which known target should be prioritized.
Keeping these systems separate allows detection and prioritization to be tuned independently while still feeding into the same perception and decision logic.
Threat Manager Component
The Threat Manager Component should be added to the AI Controller. Once enabled, it automatically integrates with the Perception Component and listens for damage-related stimuli. Whenever a Damage Sense stimulus is received, the component:
Tracks the source actor
Accumulates total damage applied per actor
Maintains an internal table of threat values
The accumulated threat table is normalized, and the actor with the highest resulting threat score is selected as the current threat target. When a new threat target is determined, a Threat Sense event is emitted.
In addition to damage-driven threat, gameplay systems may deliberately inject threat by calling RegisterThreatSource, providing a source actor and an explicit threat score. This allows non-damaging actors to contribute to threat generation, which is often the case for certain RPG roles like healers.
Threat Sense
The Threat Sense is a custom sense introduced by this module. It is triggered whenever the Threat Manager selects a new highest-threat target. When a threat event is emitted, the sense provides:
The highest-threat target actor
The normalized threat score
From this point onward, the agent processes the Threat Sense stimulus in the same way it would process any other perception stimulus. This allows threat relevance to be authored using the same tools as sight, hearing, or other senses, including weighting, normalization, and awareness contribution. As a result, threat can be tuned per perception archetype and blended naturally with other sensory inputs.
You can send one-off Threat events, using the static function provided by the Threat Sense class. This function can be used natively or in Blueprints, and it's useful to trigger external, specific events that are supposed to force threat like a " taunt ".
Threat Modifier Interface
Actors that participate in threat evaluation may implement IBotThreatModifierInterface to influence how much threat they generate. This interface allows actors to:
Increase their threat contribution
Reduce or dampen their perceived threat
This is particularly useful for role-based gameplay, such as implementing class archetypes like tanks or damage dealers, where threat generation is intentionally biased.