Updating Blackboard Keys Based on Perceived Stimuli

Create a custom perception handler to automatically update Blackboard variables whenever the AI detects a new target.

In this lesson, you will learn how to automate Blackboard updates by binding custom logic directly to the AI Perception system’s event delegates.

Moving Beyond Polling

Previously, we discussed filtering stimuli; now, we ensure those stimuli immediately translate into actionable data. Instead of constantly checking the Blackboard every frame, we want to react specifically the moment an actor enters or leaves our perception range. We achieve this by hooking into the OnTargetPerceptionUpdated delegate provided by the AIPerceptionComponent.

The Custom Handler Pattern

The most robust approach is to implement a handler function in your AI Controller. This function acts as the bridge between raw sensory data and your Blackboard variables.

When the OnTargetPerceptionUpdated event fires, it provides an AActor* and an FAIStimulus. We can use these to update our Blackboard:

void AMyAIController::OnPerceptionUpdated(AActor* Actor, FAIStimulus Stimulus)
{
    if (Stimulus.WasSuccessfullySensed())
    {
        // Update Blackboard for an active target
        GetBlackboardComponent()->SetValueAsObject("TargetActor", Actor);
        GetBlackboardComponent()->SetValueAsBool("IsTargetVisible", true);
    }
    else
    {
        // Handle target loss
        GetBlackboardComponent()->ClearValue("TargetActor");
        GetBlackboardComponent()->SetValueAsBool("IsTargetVisible", false);
    }
}

Why This Matters

By centralizing this logic in a perception handler, you decouple the AI’s "senses" from the "decisions" made by the Behavior Tree. When the TargetActor key changes, the Behavior Tree’s Blackboard Decorators instantly react, forcing the AI to abort current tasks (like patrolling) and switch to combat behaviors.

This approach is highly efficient because it is event-driven. The CPU does not process logic until the perception system triggers the callback, keeping your AI performance lean even as you scale to multiple agents. Ensure your AIPerceptionComponent is configured with the correct AISenseConfig for sight or hearing, and bind your function during BeginPlay using GetAIPerceptionComponent()->OnTargetPerceptionUpdated.AddDynamic(this, &AMyAIController::OnPerceptionUpdated);.

Key Takeaway: Use event-driven perception handlers to push sensory data into the Blackboard immediately, ensuring your Behavior Tree reacts to environment changes the exact millisecond they occur.