3D Trayectory Data Optimization Through an Adaptation of the Douglas-Peucker Algorithm: Application to Surgical Robotics

The advancement of the autonomy level of surgical robots is linked to the improvement in the learning of trajectories made by expert surgeons when performing simple procedures. An initial phase in this learning process involves refining the ability to derive valid robotic trajectories from recorded surgeon maneuvers. Manual trajectories typically involve high point-sampling, posing challenges for integration into learning algorithms. This study aims to obtain resampled trajectories characterized by a significantly reduced point count and the capacity to define parametric key points. This allows for their utilization in instructing a surgical robot to replicate learned trajectories from surgeons, all while preserving critical parameters such as precision and velocity. The article introduces a trajectory simplification algorithm, based on the Douglas-Peucker algorithm, designed to minimize the point count in potentially intricate trajectories.