Research On Safety Monitoring Of Spinal Surgery Robot Based On EMG

Currently,spinal surgery robots mainly use preoperative imaging information and intraoperative visual navigation techniques to localize rigid bone tissue structures,lacking means to monitor important nerves in soft tissues,and there is a risk of iatrogenic nerve injury.In traditional spinal surgery,neurophysiological monitoring techniques are an important tool to avoid nerve injury.However,the waveforms given by neurophysiological monitoring devices are mostly original data of neurophysiological signals,and the process of decoding and judging the intrinsic meaning of the data can only be done by the surgeon in charge with professional experience,which makes the technology have a high threshold of application.Therefore,taking robot-assisted spine surgery as the research background and improving the safety and intelligence of surgical robots as the research goal.A safety pattern classifier model based on automatic recognition of EMG signals is constructed,and the feasibility of the safety monitoring system of the surgical robot is verified.The main research contents are as follows:(1)In order to effectively avoid nerve root injuries,a robotic safety monitoring system for spinal surgery based on EMG signals was created based on existing operational procedures for robot-assisted spinal surgery.A surgical robot experimental platform consisting of a robotic arm,an end-effector tool and a navigation camera was built according to the needs of robot-assisted surgical operations,and completed system integration by solving the challenges of system calibration.A safety monitoring module based on evoked electromyographic signals was designed to enable early warning of nerve root injury in a robotic system for spinal surgery and to improve the safety of the surgical robotic system.(2)To complete the training of the safety pattern classifier model,an animal EMG signal acquisition scheme was designed,signal pre-processing and signal analysis were performed,and the creation of an EMG signal dataset based on the animal model was completed.The collected samples were put through the process of data pre-processing,data cleaning,and feature extraction,and finally a dataset consisting of EMG signal feature vectors and safety pattern labels was constructed.(3)Using a self-built EMG signal dataset,the classification results of several machine learning classification algorithms were compared.The random forest algorithm with the highest classification accuracy was selected.A robot-assisted spinal puncture surgery simulation experiment was designed based on the surgical robot experimental platform using rabbits as the experimental object.Twelve sets of EMG signals collected during needle insertion were used as test data to verify the accuracy of the EMG signal classifier and robotic safety mode selection.Finally,the works of this paper is summarized and further research work is analyzed.The safety monitoring technology of spine surgery robot based on EMG signal has important research significance to further improve the safety of surgery.