Hierarchical super-twisting control with improved extended state observer and gain correction for velocity vector in dual-wheeled mobile robot

Addressing the challenges of high-frequency noise and compound disturbances affecting the velocity vector control system of double wheeled mobile robots (DWMRs) in complex motion scenarios, we propose a robust control strategy that integrates hierarchical super-twisting control with an enhanced extended state observer. First, we designed a boosted extended state observer (BESO) with noise suppression capability. By incorporating a second-order low-pass filter to reconstruct the error feedback channel of conventional observers, we achieved effective decoupling between high-frequency noise suppression and low-frequency disturbance estimation. Furthermore, we introduced a parameter tuning rule based on the pole placement method, significantly simplifying the observer parameter adjustment process. Second, we established the planar kinematic model of the DWMR and developed a hierarchical sliding mode control framework. A gain adjustment strategy utilizing the super-twisting reaching law was proposed to modify the control gains, thereby enhancing the overall system robustness. Theoretical analysis demonstrates that the designed controller guarantees finite-time global convergence even in the presence of disturbance errors, with system stability rigorously verified through Lyapunov functions. Experimental results indicate that the proposed method exhibits superior disturbance rejection capability and motion control precision in simulation and actual platform.