| Touch is not only a supplement of vision,but it also has a strong sense of power.It can directly measure the characteristics of other things such as morphology and structure.Therefore,artificial tactile sensing has become a hot research direction,especially in robot sensing,surface texture estimation,and minimally invasive surgical operations.Researchers have tried various technologies and materials to produce sensitive,reliable,rugged,small,and low-cost touch sensors.This paper focuses on the problem of abnormal tissue recognition in minimally invasive surgery.Based on the research work of fiber Bragg grating sensing mechanism,relevant design and research of the tactile sensor are carried out,different sensor structures are proposed,finite element simulation is performed.And two types of tactile sensors that measure spatial distribution are produced.The first type is the rod-like structure of a single FBG.Through sliding experiments,it was verified that the measurement of the spatial distribution of the tissue surface and the positioning of hard burial material under the tissue can be achieved by sliding the sensor on the tissue surface.The second type is an array sensor with 9 sensing units,encapsulated in silicone.The neural networks is used to process the sensor data in order to provide information about the type of contact.Finally it was verified that the sensor could distinguish the position or shape of the contact load.The main research contents are as follows:1)Read the related literature to understand the current status of the tactile sensor technology in minimally invasive surgery,analyze the development trend and existing problems of tactile sensor technology.Combine multiple tactile sensor structures and measurement methods,and propose a type of sliding tactile sensing method.2)On the basis of summarizing the research work of the predecessors,the prototype of the sensor is designed,the force conditions of the four microstructures are simulated,and the prototype is produced.An experimental platform was set up and the artificial soft tissue with hard nodules buried was fabricated.The feasibility of the sliding experiment was demonstrated by indentation experiments and dot-matrix experiments.The sensors were allowed to slide over the fabricated tissue model with the same depth of sensor indentation.A series of data is acquired using a demodulation instrument,and the position of a hard object is determined by forming an image.3)Based on another method of tissue palpation,a 3*3 array sensor with fiber Bragg grating was fabricated.A pressure support algorithm for four-point support was introduced,and a single-point vertical pressure experiment was performed.In order to expand the detection range and achieve rapid detection,a localization algorithm based on BP neural network is proposed.The characteristics of experimental data are analyzed,and 1600 sets of sample data are extended.BP neural network is constructed and trained using MATLAB,and the feasibility of the algorithm is verified.4)Using 3*3 array sensors to detect objects of different shapes.A shape detection method based neural network using low-density array sensor is proposed,and a training BP neural network is constructed using MATLAB to verify the feasibility of the algorithm.Finally,the research situation of the project was summarized,the results of the study and the existing deficiencies were analyzed.The areas for improvement were put forward,and the future work was prospected. |
