Title: Investigating the Generation of Enemy Morphology Through Collision Data

Original: arXiv:2606.02832v1 Announcement Type: New Abstract: While Procedural Content Generation (PCG) has been the subject of extensive study, the specific task of creating video game enemies remains underexplored. Notably, research into enemy morphologies—defined as the fundamental body structures or collision data for in-game adversaries—is virtually non-existent, even though similar generative methods are established in robotics. This study examines three innovative techniques for producing enemy morphologies using player collision metrics. Our results indicate that while each method presents distinct advantages and limitations, they all achieve performance levels that match or surpass an evolutionary baseline derived from previous robotics morphology research.

Source: arXiv Generated at: 2026-06-03 00:00:00 UTC