AI design and development of survival horror game

• Main Author: Huangfu, Qingchuan
• Other Authors: Shen Zhiqi
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/138724
• Institution: Nanyang Technological University

This Final Year Project aims to develop the AI component of a Survival Horror Game - Lydia. The project experiments the ways of designing a reasonable and realistic behaviors of AI that reach or surpass players’ expectations. Behavior Tree is chosen as the core technique to develop powerful and flexible AIs that critically challenges the player. Along with the enemy behavior, a Difficulty System is developed to dynamically adjust the difficulty of making progress of the game, maneuvering menace level to boost player engagement and satisfaction. Additionally, K-Means Clustering is used to solve a specific problem for the enemy AI to better predict player intentions. Monte Carlo Tree Search technique was also explored with an earlier design of the game. However, it was not adopted into the final version of the game due to reasons to be discussed. This report primarily describes the implementation of AI techniques, and some improvements on the existing framework, which are applicable and extendable to other game development projects. The report also presents the full gameplay of the survival horror game Lydia.