Study On End Positioning Of TBM Cutter Changing Robot Based On Binocular Vision

With the continuous development of infrastructure construction such as underground engineering,tunnel engineering,and water diversion engineering in China and countries along the Belt and Road,full-face rock tunneling machines are widely used in the construction of various tunnels due to their high tunneling efficiency,good safety performance and low construction cost.TBM mainly relies on the cutters on the cutter head to break the rock during the tunneling process.Under the strong action of the rock and soil,the cutters are easy to wear and fail.The failed cutters need to be replaced in time to ensure the normal tunneling of the TBM.However,the replacement of the failed TBM cutters currently depends almost entirely on manual labor.Manual replacement will bring a series of negative problems,including the safety of construction personnel,low efficiency and high cost.Therefore,the use of robots for safe and efficient replacement of failed cutters is a hot topic of current research.However,the existing TBM cutter change robot has a very large end positioning error,and manual assistance is still needed in the recognition and positioning of the cutter,which is difficult to carry out large-scale engineering applications.Therefore,the research on the end positioning of the TBM cutter change robot is very important.In response to the above problems,the main contents of this article are as follows:(1)Aiming at the problem that the recognition and positioning of the hob of the TBM cutter change robot system still requires manual assistance,the reasons for this problem are analyzed and summarized,including the inaccurate kinematic model of the cutter change robot,the large time-varying deformation due to heavy load and the unstructured cutter change environment.For these problems,this article proposes to add binocular vision to the cutter change robot system to replace the human eye recognition cutter system,and provides a visual feedback link to improve the robot end positioning accuracy.(2)In order to facilitate experimental verification,the scale ratio of the TBM cutter change robot was determined and a scale design was carried out.The robot changed from the original PPRPPR structure to the PRPPR structure,but the overall motion form remained unchanged.Then use the improved D-H method to model the positive kinematics,and use the Matlab robot toolbox to verify it.Finally,the inverse kinematics modeling of the robot was carried out using the geometric method,and its correctness was verified.(3)Aiming at the problem that traditional cutters cannot be directly replaced by robots,a new integrated cutter system is proposed,and the characteristics of the cutter system are analyzed and extracted.Aiming at the shortcomings of using gray processing,binarization processing and template matching methods to quickly and effectively extract bolt features,a cutter feature recognition and extraction algorithm based on YOLO-SIFT is proposed,and recognition experiments are carried out.The experimental results show that the proposed algorithm can quickly extract the bolt feature points,and the relative error of the identified bolt pixel coordinates is less than 1.5%.(4)The end positioning system of the TBM cutter changing robot based on binocular vision was built,and the internal and external parameter calibration of the binocular camera was completed.In order to verify the effectiveness of the system,a set of scale test benches were built,including the TBM cutter posture simulation test bench and the TBM cutter change robot scale test bench.Experiments were carried out by adjusting five groups of different cutter poses.The experimental results show that the proposed TBM cutter changing robot end positioning system based on binocular vision can achieve positioning accuracy less than 2.5mm and meet the target requirements.