This document describes the architecture and the design of the Flexcrash platform for organizing user studies to identify relevant mixed traffic scenarios. It also demonstrates the Flexcrash platform's main features: generating and simulating interactive driving scenarios. Using the Flexcrash platform, regulators, developers, and other interested parties can study the live interaction between human drivers and autonomous vehicles in simulated mixed-traffic scenarios. The Flexcrash platform implements an online turn-based, multiplayer game that allows remote players, humans and not, to simultaneously participate in driving simulations; thus, it enables live interaction and possibly leads to identifying critical mixed-traffic scenarios. The Flexcrash platform follows established design patterns typical of Web applications to achieve scalability, interoperability, and ease of use. The Flexcrash platform aims for reusability and generalizability; thus, it relies on open-source frameworks, such as the CommonRoad motion planner benchmarking framework. Consequently, for this demonstration, we use the state-of-the-art, sampling-based motion planner kindly provided by CommonRoad's Principal Investigator, Prof. Matthias Althoff (Technical University of Münich), as reference AV implementation. Other AV implementations can be added to the Flexcrash platform after its release to the public. Promoting open research and favouring results' replicability are fundamental to Flexcrash; therefore, the Flexcrash platform's code, manuals, and other resources are made publicly available at:
