Design - Hitboxes

How to handle damage and getting hit in general.

Damage can be handled many different ways depending on your project scope and you should make a distinct choice about how you handle it.

Relevant Components

AgentAttributes, AgentVitals, IDamageable, StrikeNode, StrikePointData

Typical Setup

The architecture is 'location-based' by default. It's really just a matter of how many 'locations' you define. You can just as easily have 1 collider (The Controller) to capture damage as you can 100 (strike nodes).

• Add any number of colliders you wish to an Agent's model.

• Put these objects on the "StrikeNodes" layer.

• Reference them to the appropriate StrikePointData entities.

• Reference the owning AgentVitals component on each of them.

• Set the layer of the AgentVitals object to something different, such as the "Agents" layer.

Now when you set your Weapon System's HitMask to only the "StrikeNodes" layer then they will make contact with these hitboxes, send damage messages through the IDamageable interface and the StrikeNode component will multiply it as designed and pass the new data to the AgentVitals for processing. AoE and multi-hit situations are solved by referencing each contact's IDamageable.LifeId for duplicate Id's. The LifeId is guaranteed unique per each AgentVitals component so you can easily and effectively understand if you have more than one contact on the same entity.

This is highly customizable, allowing for basic setups where an Agent has only one (or zero) StrikeNode components pointing to the AgentVitals, all the way to having a StrikeNode for every limb and every equipped item model that may effect damage. (Shields!)

For example...

Scenario A (simple)

• Root Character Object (AgentVitals) on "Agents" layer

  • Character Model

    • Equipment models

      • Etc.

In this scenario it is extremely simple, the AgentVitals captures all damage events through it's IDamageable interface and would need to do processing on what equipment it has and how that effects damage.

Scenario B (complex)

• Root Character Object (AgentVitals) on "Agents" layer

  • Character Model

    • Leg (Ragdoll Collider/Rigidbody)

      • StrikeNode (and large hitbox on "StrikeNodes" Layer)

    • Torso (Ragdoll Collider/Rigidbody)

      • StrikeNode (and large hitbox on "StrikeNodes" Layer)

    • Head (Ragdoll Collider/Rigidbody)

      • StrikeNode (and large hitbox on "StrikeNodes" Layer)

    • LeftPinkyToe (Radgoll Collider/Rigidbody)

      • StrikeNode (and large hitbox on "StrikeNodes" Layer)

In this scenario we have a much more complex situation where there are many colliders on the character and ragdolls are supported. By putting the StrikeNode objects and components as separate objects under the bones that they're connected to we can identify them on different layers and with different colliders. This is necessary because we probably always want our "hitbox" colliders to be larger than our more precise ragdoll colliders.

Weapons / Projectiles need to be setup to filter only the layers you want them to interact with, so the design path you want to use here is best to know early on.

All of the StrikeNode components here would be pointing to the parent object's AgentVitals component and just as easily forward damage events to it.

Each StrikeNode is capable of processing and modifying damage prior to sending it to AgentVitals so it's important to understand where and when you are adding/modifying damage. This is a design choice for you to make.

Work is in progress on this page