Kieran Cooksley
Kieran Cooksley
Technical Designer
Role
AI Technical Designer
20 Person Team
Made With
Unreal Engine 5
Blueprints
Timespan
10 weeks
11 Mar - 17 May '24
Janitor of a mysterious facility seems mundane, that is until a dangerous creature breaks out from containment and lurks the hallways...
Project Goals
-
Work collaboratively to develop a vertical slice of a game in a 10 week period.
-
Develop a deeper understanding of Artificial Intelligence by implementing advanced AI features including AI Perception, EQS, and an AI Director.
Key Responsibilities
-
Design & Implementation of enemy AI Systems.
-
Collaboration with other departments including design, art, and animation to ensure AI is fully implemented.
-
Extensively test and bug-fix AI agent using Unreal Engine's AI Debugger.
Tools Used
Unreal Engine 5
The primary engine used for development
JIRA
Task management
& tracking
Microsoft Office
Documentation
Inspirations
Alien: Isolation
Night of the Consumers
Viscera Cleanup Detail
Documentation
State Machine
When designing the enemy, a state machine was finalised to consolidate all enemy behaviours and understanding the rules for transitioning between them. This broke down the task of developing the enemy into much smaller clear, manageable tasks.
Flowcharts
Flowcharts accompanied each system and behaviour, visualising the flow of execution and how the behaviour should work and be constructed. This assisted in development as visualising this execution helped identify and prevent issues that may have occurred had proper planning not been carried out.
Behaviour Diagrams
Behaviour Trees need to be very accurate & precise to ensure the correct sub-trees are always executed, so planning out the structure of the Behaviour Tree and connected Sub-Trees sped up the development process. These diagrams also provide a more user friendly depiction of AI Behaviour for other departments to review and understand.
EQS
The Environment Query System was used to allow the AI to make decisions based on its environment. EQS Diagrams present the construction of these queries, explaining how each Test affects the final outcome of the Query.
