A General Analysis Framework for Multirotor Thrust-Vectoring Vehicles: From Configuration to Modeling to Data-Driven Approaches

By actively modulating thrust directions, multi-rotor thrust-vector aerial vehicles (TVAVs) overcome the underactuation inherent in conventional coplanar multirotor systems, thereby enabling enhanced maneuverability, full-attitude control, and robust operation in confined or highly disturbed environments. With increasing structural complexity and actuation redundancy, research efforts have progressively evolved from configuration design and aerodynamic analysis toward system-level dynamic modeling and, more recently, data-driven methodologies. This paper presents a comprehensive review of the research evolution in multi-rotor TVAVs, beginning with a summary of configuration and structural analysis methods that explicitly consider thrust-vectoring layouts and aerodynamic effects, and their influence on attainable force spaces, aerodynamic force distribution, and control capabilities. Subsequently, dynamic modeling approaches and investigations into system dynamic properties are reviewed, together with model-based trajectory generation and full-attitude control methods that ensure dynamic feasibility. Furthermore, recent advances in data-driven and reinforcement learning–based methods are systematically discussed, highlighting their potential in addressing strong nonlinearities, model uncertainties, and aggressive maneuvering tasks. Finally, the advantages and limitations of different research paradigms are compared, and the central role of control allocation in thrust-vectoring control architectures is examined, with the aim of providing a structured perspective on the evolution from configuration analysis to dynamic modeling and data-driven methods, and of offering insights toward future unified frameworks that integrate structural constraints, aerodynamic characteristics, model-based design, and data-driven intelligence.