Research On Adaptive Algorithm Of Shape Sensing Based On Fiber Bragg Grating

With the development of science and technology,the application of robot technology is more and more extensive.Snake-like robot is a kind of bionic robot made by imitating biological snake.Because of its special slender structure and flexibility,it can adapt to various complex terrain and enter the environment that people can not reach.So it is widely used in disaster relief,geological exploration,medical treatment,fault detection,anti-terrorism and explosion protection.In order to achieve more accurate,safe and efficient use of snake-like robot,it is necessary to detect the shape of its working process.However,when the snake-like robot with high flexibility works in complex and harsh environment,it is difficult to directly observe and measure its working shape and position.Thus,other methods are usually required for monitoring,and shape sensing technology emerges as the times require.Optical fiber shape sensor is favored by researchers because of its high accuracy,small size,anti-interference and easy to be reused in large scale.However,there are some problems in the current mainstream shape sensing technology,such as large accumulative error of reconstruction algorithm and the inability to select sensing points adaptively,which greatly affect its application in the multibending and long-distance working environment of snake-like robots.In order to solve the problem of accumulative error and sensor point arrangement in shape sensing,this paper proposes an adaptive shape sensing algorithm based on fiber Bragg grating based on the existing mainstream curve reconstruction algorithm.The curvature of the sensing point is reasonably screened by the curvature absolute value threshold and the curvature gradient threshold,which reduces the impact of cumulative error while ensuring the amount and accuracy of data.At the same time,it can adaptively select the sensing points and improve the calculation speed to a certain extent.Firstly,the curvature screening criterion in the adaptive algorithm is studied,and then the bending rod model is established by simulation.After extracting the strain information,the current mainstream shape sensing algorithm based on Frenet frame is restored.On this basis,the influence of sensing points,curvature threshold,curve length and other factors on the reconstruction accuracy is studied.After the simulation,a curve shape sensing experimental platform was built.A 135 cm curve shape sensor was fabricated by using memory alloy and fiber grating sensor.Three fibers with 15 grating sensors were pasted at equal angle intervals.Real-time curve reconstruction has been achieved using this platform.Several types of curves similar to the common working environment of snakelike robots were selected,and a verification experiment was conducted to improve the accuracy of the adaptive reconstruction algorithm.The results showed that the adaptive algorithm proposed in this paper can effectively improve the reconstruction accuracy of curves,verifying the feasibility of the algorithm.Based on this,an 800 cm long distance shape sensor was produced to verify the effect of the algorithm on long distance shape reconstruction.Finally,the research results and shortcomings were summarized and analyzed,and areas for improvement were proposed,with prospects for future research.