Games, Interactive Entertainment and Drama


Home

News

People

Research

Games Reading Group

Prospective Graduate Students

Publications

Downloads

Video clips

Links

Pictures

Research Projects

GADIN

The GADIN (Generator of Adaptive Dilemma-based Interactive Narratives) system generates interactive narratives which are (potentially infinitely) long, and that adapt to the user's behaviour. To add dramatic tension, the story incorporates dilemmas as decision points for the user. These dilemmas can be based on the cliches found in many contemporary soap operas, such as the trade off between personal gain and loyalty to a friend. Overarcing stories connect these dilemmas as points of interaction within a coherent plotline that is dynamically created, based on the user's responses and action choices. The user is presented with those dilemmas which represent the hardest decisions for them individually. In this soap genre the stories will continue indefinitely. The GADIN system is also capable of incorporating a changeable story goal so that the narratives will be finite and the ending will be dependent on the user's freely-made actions while remaining clear and coherent to them.

DED

The Directed Emergent Drama system combines directorial control over a narrative arc with semi-autonomous reasoning and actions of actor agents.

Data Mining of Game Traces for In-Game Commentary Generation (with Beautiful Games Studio/Eidos)

The project is concerned with the mining of football match data created in the Championship Manager game. One of the intended uses of this data is to generate game commentary, as well as create a more realistic football simulation.

Hybrid Team AI for Squad-based Computer Games (with Core Design Studio/Eidos)

The goal of this project was to develop an NPC AI for a squad-based tactical shooter that on one hand displays realistic individual reactive behavious (e.g., duck when there is an explosion nearby), and intelligent team behaviour (e.g., flanking attacks). The developed AI consists of several hierarchical layers. This multi-layer architecture combines the flexibility, efficiency, and realism of reactive behaviours with higher-level behaviour control in the form of state machines. The lowest layer controls pathfinding and steering. Higher layers control reactive behaviours, individual agent states, and team tactics. A more flexible goal-oriented planning layer has also been integrated. The resulting behaviour was realistic, believable, and displayed intelligent team coordination.