Sandblasting robot with model data-driven slip compensation and dual-motor coordinated hierarchical control method

To address the issue of reduced motion accuracy in tracked sandblasting robots caused by nonlinear slippage on steel grit-covered hard surfaces, a hierarchical control framework combining model and data learning is proposed to enhance dual-motor coordination precision and terrain disturbance suppression capability. The inner layer employs a slip-mode control-cross-coupling (SMC-CCC) algorithm, combined with a novel composite approach law, to simultaneously optimize motor tracking error and synchronization error. The outer layer employs a coupled simulation of multi-body dynamics and discrete element method (MBD-DEM) to generate high-fidelity data, which trains a radial basis function neural network (RBFNN) to predict and compensate in real time for slip disturbances caused by track-ground interactions. The standard deviation of dual PMSM synchronous error decreased to 1.12, while the standard deviation of tracking error dropped to 7.86. After RBFNN compensation, vehicle lateral deviation was reduced to 70.08 mm—a 92.28% decrease compared to the uncontrolled condition—with slip rate stabilizing between 20% and 25%. In summary, the proposed hierarchical framework effectively addresses motion instability issues beneath steel grit surfaces: SMC-CCC ensures motor coordination precision, while RBFNN achieves adaptive compensation for nonlinear slippage based on MBD-DEM data, providing a viable solution for high-precision motion control in complex terrain.