Project Details

  • Type: Action-FPS, Combat
  • Players: Solo
  • Engine: Unreal Engine 5.3
  • Team Project (10 people)
  • Devellopped over:  WIP
  •  

Introduction

As a Game & AI Combat Designer, I am responsible for designing the abilities, behaviors, and reactions of enemy AI. Additionally, I develop these elements to test and balance them effectively. I am also working on developing a spawn system that aligns with the game’s design.

Pitch : Raze, a genetically modified creature designed to serve the Scarlet Bastion, breaks free from captivity after a devastating attack on the facility where it was imprisoned. Driven by a fervent desire for freedom, Raze faces an unknown and perilous world from which it must escape at any cost.

Work in Progress

What did i do ?

    • Enemy Design
    • Creating AI abilities, behaviours and reactions.
    • Creation and editing of documents
    • Prototyping enemy and combat features, then develop it
    • Bug fixing and working on feedbacks.
    • Applying retakes from reviews.
    • Communicating with the team to validate and retake features to match the design intentions.

Tools used

Enemy Documentation
eng
Spawn System
eng
AI Manager / Ticket System
eng

Enemy Design & Development

Swarmers

A Swarmer is a fast but fragile unit primarily operating in the Near Zone. Each Swarmer unit has distinct behavior and weaponry. There are four types of Swarmer Units:

  1. Shell-Kaboom
  2. Tel-explosion
  3. SW-armed
  4. WarHeadBot
 

CAIXI

A CAIXI is a unit that primarily operates in the Mid Zone. Except for the Baopo, CAIXI units serve as cannon fodder with the main objective of protecting the Baopo, which have more complex behaviors. There are three types of CAIXI units: 

  1. Aggressive
  2. Healer
  3. Baopo

Development

I developed each AI using a Behavior Tree, a modular and hierarchical system that allows for complex decision-making processes. By structuring the AI’s behaviors into nodes and branches, I could effectively manage the sequence, selection, and execution of various actions and reactions based on different conditions. This approach not only makes the AI more adaptable and responsive but also simplifies debugging and balancing. Here is an example of 2 Behavior Trees that illustrates how these elements are organized and interact to drive the AI’s behavior.

References

roboquest_cover
Witchfire