Attitude tracking control of quadrotor unmanned aerial vehicle based on an adaptive predefined time value iteration approach

This paper addresses the attitude tracking problem of quadrotor unmanned aerial vehicles (UAVs) by introducing a novel predefined-time stable adaptive value iteration (PTS-AVI) control scheme. Unlike conventional value iteration-based adaptive dynamic programming (Ⅵ-ADP) methods—where the running cost is independent of the value function—the proposed cost function explicitly incorporates the value function $V$, thereby ensuring predefined-time stability (PTS) throughout both the training and deployment phases. This functional dependency was systematically addressed via an auxiliary time-scale partial differential equation (PDE) formulated in the $s$-domain. To circumvent the need to solve the complex Hamilton–Jacobi–Bellman (HJB) equation directly, a parameter update method was employed to approximate its solution, yielding an optimal control policy that satisfies predefined-time stability criteria. In addition, a high-precision predefined-time disturbance observer was designed to estimate and compensate for unknown disturbances. Both theoretical analysis and simulation results confirmed that the proposed control scheme guarantees state convergence to equilibrium within a user-specified time, regardless of initial conditions. This work differs from existing studies by integrating predefined-time stability requirements into the Ⅵ-ADP formulation and coupling it with a predefined-time disturbance observer for attitude tracking.