Data-driven Model Free Formation Control for Multi-USV System in Complex Marine Environments

Formation control of multi-USV system in complex marine environments is investigated in this study, and a data-driven disturbance observer (DDO)-based model free adaptive control (MFAC) framework is suggested. The FMAC scheme is designed by using only the input and output data of complex USV dynamics, including the calculation of pseudo-Jacobian matrix (PJM) and the calculation of control law. To avoid the problem that the MFAC method cannot be directly applied to USV systems, an improved DDO is used to estimate the complex environmental disturbances and PJM estimation errors. Further, forgetting factor based MFAC (FMFAC) is proposed to avoid overshoot caused by redundant data. Then, the PJM estimation is proved to be accurate while the control structure of DDO-FMFAC is proved to be bounded-input and bounded-output (BIBO). Finally, the performances of the proposed method including effectiveness and robustness are shown in simulation results.