Force feedback predictive control based on BP neural network of MIS robot

Abstract

Virtual sensing technology is an effective means to get force information between surgical instruments and surgical field environment for minimally invasive surgical robot (MIS), but this sensing technology needs some time to do image information's acquisition, processing, transfer and data fusion with other data resources, so it has more time delay than physical sensor which get information directly. The delay of force feedback information can cause a bad influence on rapid response capability, precise positioning and maneuverability of teleoperation system. By means of setting a BP neural network predictor on the master side of MIS robot to predict the current force feedback signal when the slave side is contact with the patient's body via the status information of the master and slave side, and amend the delayed force feedback signal which is received from the communication link, the bad influence caused by the information delay is reduced. The simulation results show that, the output of BP neural network can get very close to the output of the controlled object's dynamic mathematical model, and achieve a good predictive accuracy and effectiveness.