Research On Motion Compensation And Force Estimation Of Minimally Invasive Surgical Robots

Minimally invasive surgical robots have improved traditional minimally invasive surgical operations,greatly improving surgical flexibility and quality.The surgical instruments that have been sterilized during surgery pass through a 5-10 mm incision to reach the lesion site.Due to sterilization and space limitations,it is not possible to install a force sensor near the front end of the instrument and a steel wire rope is required to transmit motion.This will result in a lack of force detection during surgery,and the lag of the steel wire rope reduces the accuracy of the instrument’s motion,thereby affecting intraoperative hand eye coordination and consistency.Therefore,it is urgent to explore methods to achieve force detection and improve the motion accuracy of surgical instruments under the condition of powerless sensors.Therefore,research on motion compensation and force estimation of minimally invasive surgical robots will be carried out.For the 7-DOF surgical robot that does not meet the Pieper rule,the kinematics model is constructed,and the kinematics model is simulated and verified by the Simulink module of Matlab.On the basis of the kinematics model,a collision detection method for surgical instruments in the multi arm cooperative environment is proposed.In response to the hysteresis and flexibility of the steel wire rope,as well as the influence of the backlash of the motor gearbox,the inability to install an angle sensor at the end of the surgical instrument,and the inability to form a closed-loop control of the movement of the surgical instrument end,resulting in low accuracy of the surgical instrument end and affecting the consistency of hand eye coordination control,machine learning based feedback compensation and scale scaling based feedforward compensation algorithms are proposed,Improving the accuracy of surgical robot end motion through a combination of feedforward and feedback.Firstly,establish a single degree of freedom steel wire rope driven surgical instrument platform for position estimation based on Autogloun machine learning method.Use this position estimation as feedback for the end effector of the instrument.Secondly,conduct research on forward and backward motion scaling as a feedforward for instrument motion.Then the feedback compensation algorithm based on machine learning and the feedforward compensation algorithm based on scale scaling are constructed.Finally,the algorithm is compared with linear regression,decision tree,support vector machine,neural network and Gaussian process methods.The mean square error,average absolute error,maximum error and standard deviation indicators of the algorithm are minimum.Based on the algorithm’s highest position tracking accuracy,the effectiveness of the algorithm is verified.A surgical machine manpower estimation algorithm based on LSTM time series model is proposed to address the issues of sterilization and space limitations,where force sensors cannot be installed near the front end of the instrument and force detection cannot be achieved.Firstly,a surgical machine manpower estimation experimental platform is established.Through experimental analysis of the correlation between different friction terms and force estimation under different friction models,appropriate input features and time series orders are selected based on the correlation analysis,and an LSTM time series model is constructed to achieve force estimation.By conducting force estimation experiments on different objects,the effectiveness of the time series model in handling minimally invasive surgical machine force estimation is verified.In response to the above issues,in order to verify the effectiveness of the above algorithm,system integration is carried out,a master-slave control system is built,and master-slave control experiments and clamping experiments are conducted to verify the effectiveness of the above algorithm.