Research On Master-slave Mapping Accuracy And Training Trajectory Optimization Of Surgical Robot Virtual Simulation System

Micro invasive abdominal surgery robot is a major branch of medical robotics,and its application in the medical field has a good development momentum and broad application prospects.However,problems such as high research and development costs and long training time before use of surgical robots have limited their development.Therefore,how to resolve the appeal issue is crucial to the development of surgical robots.By establishing a surgical robot simulation system,it is possible to quickly,low-cost,and high-security verify,and quickly and real-time obtain feedback on the gap between expected performance and actual performance.Therefore,this paper designs a virtual simulation system based on a single hole minimally invasive abdominal surgery robot developed by the laboratory.Firstly,this paper establishes a virtual simulation robot system based on Coppelia Sim software,and establishes a teleoperation communication structure with external host devices.By establishing a master-slave mapping,physical engine,and collision module to verify the workspace and kinematics of the surgical robot,it is possible to verify the rationality of the surgical robot structure at a low cost.Secondly,this article optimizes the master-slave mapping accuracy of surgical robots through a neural network model.Record the position and posture data of the main operator and the end robot during the operator’s operation process,and use this data to train the radial basis function neural network model,and optimize the parameters through particle swarm optimization algorithm.By constructing a neural network model to predict and compensate,the master-slave mapping accuracy of the teleoperation system is improved.Then,this article processes the trajectory of skilled operators to generate a smooth operation curve,providing reference for novice operators.By comparing three methods: polynomial fitting,B-spline fitting,and extreme learning machine fitting,the optimal trajectory processing method is selected.The results show that using dynamic time warping algorithms and extreme learning machine models to process and train trajectories provides a smoother and more reliable reference trajectory for novice operators,while the degree of curve jitter can be reduced by more than 90%.Finally,by designing the external environment of the virtual simulation platform and training related tasks,the stability and fidelity of the system are verified.The results show that the designed simulation system can simulate the operation of real surgical robots,with small master-slave mapping errors,and can provide an effective reference trajectory for novice operators through training.