Automated guided vehicles (AGVs) are extensively applied in aerospace, ports, docks, rail transit, and other industries that require intelligent handling of heavy loads. Accurate modelling and trajectory tracking are crucial for performing the intelligent transfer of large objects and components using AGVs. To address the insufficient precision of current mathematical models and their inability to precisely perform trajectory tracking tasks due to the nonlinearity and strong coupling of multi differential wheel heavy-duty AGVs, a trajectory tracking control strategy based on the Koopman operator is proposed in this paper. First, based on the Koopman operator, an approximate high-dimensional linear dynamic explicit expression for the multi differential wheel heavy-duty AGV system is obtained using extended dynamic mode decomposition. Then, based on the obtained expression, a linear model predictive controller (MPC) is designed to achieve trajectory tracking control for heavy-duty AGVs. Finally, through simulation and experimental verification, it is demonstrated that the model information obtained from the data is effective and fits the dynamic model of the multi differential wheel heavy-duty AGV well. Comparisons reveal that its model accuracy is higher than that of the traditional Lagrange model; the root mean squared error (RMSE) and relative RMSE based on the Koopman model increase by an average of 24.86% and 3.24%, respectively. The designed linear model based on the Koopman operator can be combined with model predictive control. The controller can effectively perform trajectory tracking control, verifying its effectiveness. This work is the first experimental verification of the Koopman operator in heavy-duty AGV control, which will promote the engineering application of complex nonlinear systems.
Zhu et al. (Sun,) studied this question.