Data-Driven Trajectory Tracking Control with PSO Algorithm Optimization for UAVs

In recent years, precise trajectory tracking of unmanned aerial vehicles has become increasingly critical for autonomous missions. This paper presents an integrated methodology that combines a data-driven, model-free adaptive control scheme with a global particle swarm optimization process to tune controller parameters automatically. At each sampling instant, the nonlinear plant dynamics are locally linearized by forming equivalent linear data matrices, and pseudo-partial derivatives are computed to guide the adaptation law. A population-based search then refines the adjustment gains by minimizing the cumulative tracking error over the reference horizon. The resulting control algorithm features rapid error convergence, robust disturbance rejection and smooth actuation signals. Additionally, the method exhibits scalability for multi-agent formations and inherent fault-tolerance under actuator faults, broadening its applicability in safety-critical scenarios. Extensive simulation on complex, time-varying reference paths demonstrates that the optimized adaptive controller outperforms conventional model-free approaches in terms of tracking precision, stability margins, and resilience to measurement noise. Furthermore, the algorithm imposes minimal computational burden and requires no prior knowledge of the system model, making it well-suited for onboard implementation on embedded flight controllers. The proposed framework offers a promising solution for advanced autonomous flight control applications.