Fuzzy Neural LSTM-RBLS for Fractional-Order PID Sliding-Mode Motion Control of Autonomous Mobile Robots with Four ISID Wheels

This paper proposes a fractional-order PID sliding-mode controller (FO-PID-SMC) augmented by a fuzzy long and short-term memory neural network (FNLSTM) and recurrent broad learning system (RBLS), dubbed as FO-PID-SMC-FNLSTM-RBLS, for trajectory tracking of an autonomous mobile robot (AMR) with four independent steering and independent driving wheels (4ISIDW), abbreviated as 4ISIDW-AMR. In order to design such as controller, the kinematic and dynamic models of the 4ISIDW-AMR are, respectively, derived. After introducing the collision-free kinematic controller, the raised dynamic controller is designed in two stages: one is the smooth dynamic controller using fractional-order sliding-mode control to achieve finite-time stability and the other is the proposed FO-PID-SMC-FNLSTM-RBLS controller to deal with modeling uncertainties and exogenous disturbances. Comparative simulations and experimental results are conducted to show the effectiveness and superiority of the proposed FO-PID-SMC-FNLSTM-RBLS control method for trajectory tracking of the 4ISIDW-AMR compared with other six controllers in slanted linear and circular trajectory tracking.