Urban public spaces are filling with a growing number of micromobility vehicles: examples include electric scooters and bikes, personal vehicles, and delivery robots. Understanding how these vehicles can operate safely and considerately as part of larger transportation systems requires the novel use of techniques from many research disciplines, such as human-computer interactions, robotics, remote sensing, and artificial intelligence (AI). This project creates cyberinfrastructure software, data sets, and AI models that are needed to support this emerging field of research. This project enables its research community to better understand and improve vehicle interactions in complicated, rapidly changing, real-world settings. Direct outcomes include practical solutions for mitigating micromobility-related conflicts and accidents in public spaces. The development and use of this proposed cyberinfrastructure will prepare high school and college students for the nation's future workforce. This project serves the human-centric micromobility research community via three innovative AI-based service engines. First, a sensing and perception engine garners machine, environment, and human aspects from the project team’s established testbeds to strengthen the community's Micromobility-to-Everything Interaction (MEI) data preparation. This addresses the research needs in forming holistic, comprehensive understandings of diverse interaction data and augmenting them for AI model training. Second, an MEI model engine provides the research community with the AI models and tools to expand their sensing modality studies, with self-explainable graph model support. The community benefits from the expanded capabilities in performing extensive AI model studies over multiple datasets and gains interpretable AI model insights. Third, a coalition engine assistant interacts with researchers to help them navigate the cross-domain, cross-discipline research and methods needed to understand