| As an effective approach of artificial ventilation,tracheal intubation plays an important role in rescue and general anesthesia surgery of the patient.However,the intubation operation is completed manually by physicians in current clinical practice,which is faced with many problems such as low success rate,high incidence of postoperative complications,high risk of cross infection beween doctors and patients,and inadequate resources for anesthesiologists.Therefore,it is urgent to develop a robot system that can complete high-quality endotracheal intubation without doctors’operation.During the intervention in respiratory tract,the real-time perception of the terminal position and even the overall posture of the robot is crucial to the accurate and safe control,and monitoring the real-time shape of intubation robot is the preferred method to obtain the posture information of the robot in the narrow respiratory tract.Due to the advantages of compact structure,electromagnetic interference resistance and high biocompatibility,the fiber Bragg grating(FBG)sensor is integrated into the endotracheal intubation robot taken as the perceived object to realize the real-time shape perception of the intubation robot,which is of great significance to improve the quality of endotracheal intubation surgery and ensure the life safety of medical staff.First of all,the basic principle of FBG strain measuring was studied.Then the curvature and torsion calculation based on FBG wavelength information was derived in detail.According to the structure of shape sensor,a four-layer strain model was established and the strain transfer coefficient between different sensor structure layers was calculated and verified through the numerical analysis method as well as finite element simulation.The curve information calculation formula based on the FBG wavelength signal was then revised using this transfer coefficient.Secondly,the sampling discrete curvature and torsion was used as reconstruction information.On the basis of the traditional curve reconstruction algorithm,two improved algorithms which are respectively based on curvature-torsion fitting and piecewise reconstruction splicing were proposed.The performance of the three algorithms is verified by comparing the accuracy of curve reconstruction recovered from the discrete curvature and torsion information by different algorithms.Results show that the Frechet distance between the reconstructed curve and the real curve is about 1.0mm or less,the Euclidean distance error of end point is less than 0.3%,and the improved algorithm takes about 1s,which proves the superiority of the improved algorithm in reconstruction accuracy and algorithm speed.Then,the shape sensor for endotracheal intubation robot was designed and manufactured,and the parameters of the FBGs were calibrated and corrected.The relationship between wavelength and curvature of each FBG was linearly fitted to obtain the sensitivity and initial wavelength data.Two-dimensional and three-dimensional shape sensing experiments were carried out on the shape sensor respectively,and the average Euclidean distance errors of the reconstruction end point are 1.120mm and 1.358mm(0.560%and 0.679%of the full length),and the average Frecher distances are 1.922mm and 2.533mm,which meets the accuracy requirements of the feedback control of intubation robot.Finally,the endotracheal intubation continuum robot prototype was produced and the shape sensor was installed in it.A corresponding shape sensing demodulation platform and a human-computer interaction interface of the host computer were established,which totally formed a complete set of endotracheal intubation continuum robot shape perception system.The real-time shape perception experiment was carried out on this system,and the results show that the designed shape sensor and shape reconstruction algorithm can meet the requirements of feedback accuracy and speed of robot control algorithm,and can guarantee the timeliness and safety of intubation operation. |
PDF Full Text Request