Study on the Motion Characteristics of Micro-Robots in Complex Environments Based on the VACPO-Xgboost Model

With the rapid development of micro/nano-robot technology, these devices demonstrate broad application prospects in biomedical fields such as drug delivery, minimally invasive surgery, and disease detection. Traditional hydrodynamic models, based on simplified boundary conditions, fail to effectively simulate the complex boundary conditions in biomedical environments. This paper proposes a hybrid model that combines a variant algorithm based on Crested Porcupine Optimizer (VACPO) with Xgboost, termed the VACPO-Xgboost model, to optimize the prediction of the hydrodynamic behavior of micro-robots. Using the COMSOL Multiphysics simulation environment, this study investigates the impact of geometric features and surface velocity distribution on the motion of micro-robots. The proposed model accurately predicts the velocity of micro-robots under complex boundary conditions. Results indicate that the VACPO-XGBoost model significantly outperforms other machine learning algorithms in hydrodynamic prediction, achieving improvements across all three regression metrics, with RMSE reduced by 16.5% and MAE reduced by 16.2%, while exhibiting excellent robustness and showcasing its potential application value in complex biomedical scenarios. Additionally, SHAP-based analysis effectively reveals the contributions of individual features to the model's output, providing deeper insights into how input features influence micro-robot motion in complex fluid environments. This offers strong support for the optimized design and precise control of micro-robots.