Adaptive Grid Search Method Using Dynamic Step‐Size Adjustment for Robot‐Tissue Interaction Force Estimation

BACKGROUND: Reliable force perception is critical in robot-assisted minimally invasive surgery. However, constraints in end-effector size and complexity of the surgical environment hamper the integration of force sensors for direct feedback. METHODS: In this study, an adaptive grid search algorithm with dynamic step-size adjustment is proposed to optimise the Hunt-Crossley (HC) model parameters for precise force estimation in robotic applications. The method continuously adjusts the step size based on real-time estimation error, thereby enabling efficient investigation of the parameter space and refined optimisation near the optimal solution. RESULTS: Experimental results show that the proposed method improves force estimation accuracy, with a reduction in Maximum Error (ME) of at least 25%, and reductions in Root Mean Square Error (RMSE) and Average Error (AE) of at least 30% compared to conventional fixed-step approaches. CONCLUSIONS: These improvements enhance operational safety and also achieve the balance between computational efficiency and perception accuracy.