The microworld simulator paradigm is well established in the areas of ship-navigation and spaceflight, but has yet to be applied to rail. This paper presents a case study aiming to address this research gap, and describes the development of a train driving microworld as a tool to overcome some common research barriers. A theoretical framework for microworld design is tested and used to explore some key methodological issues and characteristics of train driving, enhancing theory development and providing a useful guideline for the designers of other collision-avoidance systems. A detailed description is given of the ATREIDES (Adaptive Train Research Enhanced Information Display & Environment Simulator) microworld, which simulates the work environment of a train driver in a high-speed passenger train. General indications of the testable driving scenarios that may be simulated are given, and an example of an ATREIDES-based study is presented to illustrate its applied research potential. The article concludes with a review of the design process, considers some strengths and limitations, and explores some future initiatives towards enhancing the systematic study of rail research in the human factors community.