Abstract:In order to improve the efficiency of plate structure defect detection, an ultrasonic local resonance detection method based on wideband excitation is developed in this paper. The maximum betweenclass variance method is used to analyze and process the ultrasonic field under broadband excitation, and the ultrasonic local resonance frequency of the defect is obtained. The acoustic field spatial distribution under ultrasonic local resonance frequency is used in the detection and imaging of the plate structure damage. Aiming at the isotropic plate structure with specific material and thickness, the relationship model of defect size and ultrasonic local resonance frequency is established by means of the uniform design test. The experiment results show that the developed ultrasonic local resonance detection method can effectively achieve the damage detection of isotropic and anisotropic plate structure. With the regression analysis model, the inversion of the geometric parameters of the defects can be achieved, and the inversion precision is high (the maximum radius inversion error is 2 mm, and the maximum thickness inversion error is 06 mm). This study provides a feasible technical scheme for the defect detection and quantitative analysis of plate structure.

Abstract:Acoustic field distribution characteristics of electromagnetic acoustic transducers (EMAT) are different. To ensure the effectiveness of detection, it is necessary to accurately measure the acoustic field distribution. For measuring the distribution of ultrasonic shear wave radiation sound field in solid metal materials, the pintype EMAT is used as a transducer. The method of measuring and reconstructing threedimensional radiation sound field based on tomography is studied. In further, a threedimensional sound field tomography measurement experiment system is established. The errors of the measurement system are analyzed, and a compensation method is proposed. During implementing experiments, the radiation sound field distribution of the toroidal coil EMAT at the excitation frequency of 35 MHz is measured. Its threedimensional radiation sound field in aluminum and low carbon steel is obtained, and its near field distance is approximately 8 mm. Experimental results show that this method can effectively provide the sound field distribution of ultrasonic shear waves at different depth sections in solid metal materials. By superimposing them, the threedimensional radiation distribution of EMAT in the solid can be achieved.

Abstract:The structure of wind turbine blades and aircraft wings is a kind of largearea curved platelike structure. The structure health monitoring (SHM) technique based on Lamb wave is highly sensitive to slight damage, which is one of the techniques that have wide application prospect. The traditional Lamb wave sensors are generally made of piezoelectric ceramic, which features brittleness and high hardness, therefore traditional PZT ceramic sensors are not suitable for the detection on curved surface structures. In this study, a new type of flexible 03 piezoelectric composite was prepared by combining PZT ceramic powder with epoxy resin. The influences of mass ratio of PZT ceramic powder and epoxy resin, polarization electric field, polarization temperature, polarization time and etc. on the PZT composite performance parameters were investigated. Multiple factor orthogonal experiments were conducted to determine the optimal polarization processing parameters for material preparation. Experiments were performed to investigate the sensing response characteristics of the sensor made of the PZT composite to lamb wave. The sensing response characteristics were compared with those of the sensing elements made of MFC, PVDF and the traditional piezoelectric disc. The prepared sensor was applied to airfoil surface plate and the elliptic positioning method was used to carry out damage detection. The study results show that the fabricated piezoelectric composite material sensor possesses good sensing response characteristic. The sensor can nicely fit the curved plate surface. The acquired Lamb wave signal is more accurate compared with the Lamb wave signal acquired using traditional piezoelectric disc that can only partially fit the curved plate surface. So, the research provides a new type of flexible piezoelectric sensing technique for the SHM of curved plate structures.

Abstract:Aiming at the problems of large data volume, low correlation and poor reliability of the multisource signal obtained in locomotive gearbox detection, a new intelligent optimization algorithmmultivariate function particle swarm optimization algorithm is proposed. The influence of the variation ratio and fitness of the particle population on the inertia weight is studied. Based on traditional particle swarm optimization algorithm, the convergence speed and efficiency of the algorithm are improved. Taking the fitness function of the regularized modal difference as the evaluation index of the number of the measurement points, the multisensor detection optimization of the gearbox is realized according to the modal vibration type analysis of the gearbox. Taking the tooth break fault of the gearbox as the measurement object, through comparative analysis with traditional detection methods, the proposed method accurately obtain the results: the gearbox input shaft rotation frequency of 395 Hz, the thirdstage meshing frequency of 905 Hz and its 2~5 harmonics components, then the position of the faulty gear is identified quickly. The experiment results show that the proposed method can enhance the recognition rate of structural parameters, effectively improves the fault diagnosis accuracy and also provides a key technical foundation for locomotive fault warning and safe service.

Abstract:A leak localization method is proposed based on particle filter and compressed sensing for the unexpected leakage in pipeline. Firstly, based on the method of characteristic, the state space equation of the pipeline is formulated. The improved particle filter algorithm is used to estimate the state of the transient process of the leaked pipeline. Mean square error of the estimated and observed values is utilized to locate the leakage point by solving an optimization problem. Then, the leakage rate and leakage coefficient are estimated. To enhance the estimation accuracy caused by collecting, storing and transmitting measurements, compressed sensing is used to preprocess the test data. Experimental results show that the proposed method can locate the leak effectively when the reconstructed signal with a compression rate of 65%75% is used as the input of particle filter. The relative positioning error is about 1%.

Abstract:Due to the heavy axle load, high traffic volume and high traffic density, as well as the continuous high traffic volume, the number of rail detection increases and the types become more complicated. These factors bring some hidden troubles for the safe operation of the heavy haul railway. The detection rate of the rail head transverse defect by the large rail inspection car, which used in China for the heavy haul railway, can only reach 30%40%. The detection mainly relies on hand push defect detector. Its detection speed is slow and has high requirements on the operator, which bring high safety risk for operator and line. A new type of high speed rail inspection car based on the fast phased array ultrasonic technique is studied in this paper. The new car has some advantages, including the maximum speed of up to 80km/h and functions of intelligent detect classification, realtime alarm and threedimensional visual display. It is compatible with different heavy haul railway profiles. The detection rate, work efficiency and safety protection level are improved. More importantly, it is of great significance to improve the technical level of rail inspection for China′s heavy haul railway industry.

Abstract:The metal surface crack under paint coating has the problem of low signalnoise ratio using conventional detection methods. To address this issue, an eddy current thermography detection technique based on the directional modulation principle is proposed. Firstly, the principle of direction modulation is introduced. By deducing the mathematical model, the feasibility of the method is proved. The modulation frequency is selected according to the theoretical formula. According to the directional modulation, the experimental platform is established. Then, the reason that the eddy current cannot rotate in full angle is analyzed. One kind of improved method based on PWM drive signals flipping is proposed. To solve the problem of resonance frequency drift caused by heating, a twochannel frequency tracking strategy is proposed. Finally, the artificial cracks of 45# steel under different paint coating thicknesses and the cracks of 20# steel weld under 0517 mm paint coating thickness are detected and compared with the results of magnetic powder detection. Detection results show that the direction modulation method is competent to detect the artificial cracks under 0902 mm paint coating and the simulated natural cracks under 0517 mm paint coating. Compared with the magnetic particle detection results of the artificial cracks under 0486 mm paint coating and the simulated natural crack under 0451 mm paint coating, the directional modulation method effectively improves the detection ability of under paint coating cracks.

Abstract:For the electromagnetic acoustic testing, the transmitter coil utilizes the strong pulse current that excites the ultrasonic wave. It will cause the electromagnetic interference on the receiver coil. During the period of electromagnetic interference, the receiver coil is out of order and needs time to be stable. Thus, there is the electromagnetic interference blind zone. To solve this problem, this study proposes a new electromagnetic acoustic testing technique to restrain the electromagnetic interference blind zone. In this technique, a special receiver coil is utilized. On one hand, the distance between two adjacent lines in the coil structure satisfies the acoustic inphase matching relationship. On the other hand, the mutual inductance between the receiver coil and the transmitter coil is 0. In this way, it can be decoupled from the pulse excitation system. This study analyzes the principle of restraining the electromagnetic interference by the new technique. The theoretical design method of the special receiver coil is illustrated. Finally, by using this technique, the electromagnetic interference is suppressed by 948%, which proves its effectiveness.

Abstract:The inner detection of magnetic flux leakage (MFL) is an important way of nondestructive testing (NDT) in pipeline. For the nondestructive evaluation, anomaly edge detection is an important part because its accuracy directly affects the subsequent inversion process. Due to the data noise, the accuracy of edge detection is greatly reduced, especially for complex anomaly. Meanwhile, general machine learning algorithms take more time to process huge MFL data. To address these problems, an edge detection algorithm based on data fusion and wavelet transform is proposed. This algorithm is based on wavelet multiscale transformation and decomposition. The data layer fusion, feature layer fusion and decision layer fusion are combined. First, the original data are transformed by multicolor space, and the transformation results are fused together. The fused data are transformed by wavelet multiscale. Secondly, wavelet multilevel decomposition is executed for each scale data, and the wavelet modulus maximum edge detection is performed for each level. In addition, the edge detection results are combined with detail decomposition coefficient for fusion and reconstruction. Finally, the edge detection results with multiple scales are fused to obtain the final detection edge. Experiments are conducted on the simulated data and the real pipeline dataset, respectively. Results are compared with other edge detection algorithms, including Sobel, Canny, Roberts, Prewitt, and Log. Experimental results demonstrate that the proposed anomaly edge detection method has better performance than traditional edge detection methods. OA evaluation for our method exceeds 70%, which satisfies requirement of practical application.

Abstract:The low signaltonoise ratio of leakage vibration signal in watersupply pipelines (WSP) may result in the large leak location error. To address this problem, a leak location method based on the improved empirical wavelet transform (EWT) and crosspower phase difference spectrum is proposed. Firstly, the local minimum of wavelet packet energy spectrum is used to segment the signal spectrum. The signal energy bands of different scales are achieved through wavelet packet decomposition. The distribution of energy value is used to determine spectrum segmentation interval, which solves the spectrum division problem of EWT. Then, the leakage vibroacoustic signal is decomposed into multiple components by EWT according to the spectrum segmentation interval. Effective components can be selected according to the correlation coefficients, and the effective components are bandpass filtered using the frequency band with horizontal variation of the crosspower phase difference spectrum of the effective components. Finally, the leakage can be determined by crosscorrelation timedelay estimation of the filtered effective component signal. Compared with crosscorrelation, variational mode decomposition and crossspectrum analysis, simulation and experimental results show that the proposed method can effectively locate the leak point, and the average relative location error is reduced by 67 times and 15 times respectively.

Abstract:As a key technology in intelligent driving field, lane detection plays an important role in advanced driver assistant system (ADAS), which includes lane departure warning (LDW) and lane keeping (LK), lane changing (LC) and forward collision warning (FCW), adaptive cruise control (ACC). The visionbased method is dominant in the research on lane detection technology, which is also the future development direction. This paper reviews the research progress in lane detection methods based on vision in recent twenty years. Firstly, the classification and characteristics of lane are briefly described. The general process of lane detection and its faced challenges are clarified. On this basis, the lane detection principle of the lane detection methods, including the featurebased method, modelbased method, learningbased method and etc. are emphatically expounded. Their advantages and disadvantages are reviewed, analyzed and compared. Then, the commonly used datasets and the performance evaluation indexes for lane detection are introduced. Finally, aiming at the current existing problems of lane detection, the further research direction is prospected.

Abstract:At present, the palpation examination of oral cancer relies on doctor′s experience; the detection method of intraoperative frozen section is invasive detection and the accuracy is low. In order to meet the actual needs of clinical surgical oral cancer detection, the oral cancer detection method based on highwavenumber (2 400～4 000 cm-1) fiber laser Raman spectroscopy is studied. The principle using fiber Raman spectroscopy to realize oral cancer detection is studied. The feasibility of the Principal Component AnalysisLinear Discriminant Analysis (PCALDA) statistical algorithm is analyzed. A portable oral cancer tissue detection fiber Raman system was built, and the histopathological sections of the oral cancer tissues and normal tissues were prepared. The highwavenumber Raman spectra of the oral cancer tissues and normal tissues were collected, and the spectra were processed and analyzed with PCALDA algorithm. The spectral analysis results show that the average spectra of the cancer tissues and normal tissues are significantly different at 2 780, 2 890, 2 936, 3 180, 3 285, 3 300 and 3 650 cm-1. The Raman spectroscopy of the normal tissues and oral cancer tissues were dimensionreduced and classified with the PCALDA analysis, and the detection accuracy is 966%. Compared with the result of confocal microscopy Raman spectroscopy, the detection efficiency is improved. Compared with the result of fingerprint Raman spectrum detection, the detection accuracy is improved by about 06%. The study result indicates that combining statistical method, the highwavenumber Raman spectroscopy method can nicely distinguish between normal tissues and cancerous tissues, which is expected to be an oral cancer detection method with high sensitivity, high timeefficiency and high accuracy, and has important significance for intraoperative oral cancer detection.

Abstract:To address the issue of sparse lidar point cloud and camera distortion caused by ambient light, an automatic registration method of lidar and camera based on point cloud center is proposed. The manual selection of feature points and continuous acquisition of multiple frames in traditional joint calibration can be avoided. After preprocessing the data of laser point clouds and camera image, the multichecker board point clouds are segmented automatically by the consistency of the plane normal vector. The point clouds of each checkerboard are extracted in the laser coordinate system and camera coordinate system, respectively. Then, the center points are solved iteratively by point cloud aggregation. The rough registration of the corresponding relationship between the center points of the two sensor checkerboards is realized. Finally, the iterative closest point algorithm is used for precision registration, and the calibration parameter matrix is obtained to complete the joint calibration. Measurement results show that the correct projection ratio of point cloud can reach 9793% within the range of lidar error ±3cm. This method can effectively obtain highprecision joint calibration parameters and meet the requirements of data fusion between lidar and camera in the space environment.

Abstract:Abrasive belt grinding is widely applied in the industry field. Due to its flexible contact with the workpiece and the nonuniformity of abrasive distribution on the belt, the material removal rate is difficult to be predicted accurately in theory. It directly affects the efficiency and quality control of the abrasive belt grinding. This study proposes a method for identifying the material removal rate of abrasive belt grinding based on spark images. A segmentation algorithm for spark images is presented. Quantitative feature models of the color, brightness, area, and contour features of the spark image are formulated. Pearson coefficient is used to analyze the correlation between the feature of the spark image and the material removal rate of abrasive belt grinding. A linear regression model based on the single feature of the spark image and a multifeature regression model based on the support vector regression (SVR) are established, respectively. The maximum error, the mean square error, and the determination coefficient are used as the evaluation metrics. Experimental results show that the multifeature SVR model based on the radial basis kernel function can achieve high prediction accuracy with the determination coefficient of 0976. The proposed method in this paper provides a new way to effectively control the material removal rate of abrasive belt grinding.

Abstract:In the robot grasping task in unstructured environment, it is important to acquire stable and reliable grasp pose of the object. In this paper, a dynamic multitarget 3D grasp pose detection approach based on deep convolutional network is proposed. Firstly, the Faster RCNN is utilized to conduct dynamic multitarget detection, and a stabilization detection filter is proposed to reject the noise and jitter in real time detection. Then, based on proposing depth target adapter, the GGCNN model is used to estimate the 2D grasp pose. Furthermore, the target detection result, 2D grasp pose and object depth information are fused to reconstruct the point cloud of the object, and calculate the 3D grasp pose. Finally, a robot grasping platform was established. The experiment results show that the statistical grasping success rate reaches 956%, which not only verifies the feasibility and effectiveness of the proposed approach, but also overcomes the defect of fixed and single result for 2D grasp pose.

Abstract:With the rapid development of driverless cars, the environment perception solution relying on single sensor cannot meet the demands of vehicles target detection in complex traffic scenarios. Fusion of multiple sensors has become a mainstream perception solution for driverless vehicles. In this study, a vehicle detection method in traffic environment based on the fusion laser point cloud and image information is proposed. Firstly, the deep learning method is used to detect the object data collected by the lidar and the camera sensor. Secondly, the Hungarian algorithm is utilized to track the target detection results in realtime. Then, the characteristics of the detection and tracking results from the two sensors are optimally matched with each other. Finally, the matched and unmatched targets are picked and outputted as the final perception results. The proposed algorithm is evaluated in some traffic environment tracking sequence of the public dataset KITTI and real road testing. Experimental results show that the realtime vehicle detection accuracy of the proposed fusion method increases more than 12% and the number of false detection decreases more than 50%.

Abstract:In order to achieve the fine classification of different suspicious substances in onsite rapid security inspection, this paper discusses an idea to obtain the density of substances by means of Xray backscattering signal, and completes the design and functional verification of a linear scintillator detector based on silicon photomultiplier (SiPM), which is the core component to realize the proposed method. The detector has energy resolution ability and is able to analyze the scattering attenuation of different substances under different energies. The output pulse amplitude of the detector is stable under the ambient temperature of -10℃ to 50℃, and the energy resolution varies from 221% to 257% (the energy resolution is 236% @ 595 KeV at room temperature). The structure of the linear array lays a foundation for realizing spatial onetoone correspondence between scattering and perspective images, and then hopefully realizing the fusion of scattering and perspective information. The test results of scattering energy spectra of different substances show that the energy curve detected with the detector can reflect the density characteristics of the substances, which verifies the feasibility of using the detector to measure scattering energy spectrum and achieve substance classification. The detector designed in this paper has great application prospect in intelligent identification of spot investigation equipment.

Abstract:The recovery of the complex oil and gas reservoir faces serious overflow, leakage and other drilling accidents. The real time information near bit annular flow information can help to identify the abnormality early, which can reduce the risk of well control effectively. In this paper, a method of electromagnetic measurement system for annular flow in the drilling process is proposed. Firstly, the theory of downhole annular flow electromagnetic measurement system based on Faraday′s law is studied. Then, the principle of downhole annular flow electromagnetic measurement system is analyzed. The virtual current density distribution and the mechanical model are deducted in detail. The optimum structure of the measurement system is obtained by the finite simulation software combined with excitation coil structure optimization and main structure mechanics analysis. Finally, the prototype of downhole annular flow electromagnetic measurement is constructed and the evaluation experiment is implemented. Experimental results show that the electromagnetic measurement prototype can reach the accuracy of around 1. Its safety factor is over 9, which satisfies the downhole annular flow measurement. The downhole annular electromagnetic measurement theory has great significance to the following research on downhole annular flow measurement and realization of early well control and secure drilling.

Abstract:To address the slow imaging speed of scanning electrochemical cell microscopy (SECCM), a new scanning method based on Archimedean spiral is proposed. The probe height of conventional hopping mode is set by experience without any prior knowledge. To avoid collision, the value of the hopping height is often large. Therefore, the long scanning trajectory may lead to long scanning time and low imaging efficiency. Archimedes spiralbased prescanning is used to obtain the highest point of the plane. This method can effectively reduce the hopping height of the scanning process. Then, the SECCM scanning speed is improved. Moreover, due to the continuous and smooth trajectory of the Archimedes spiral, no impact and no sample deviation occur in the scanning process. It has the advantage of high stability. Compared with the traditional hopping mode, experiments on imaging the metal surface show that the proposed method can improve the scanning speed by 110%. Meanwhile, the imaging quality can be guaranteed. The proposed method is of great significance to improve SECCM imaging speed and quality.

Abstract:Data feature selection of converter steelmaking process is the key step to realize the end point carbon content and temperature prediction. The highdimensional data of production process are not conducive to the rapid and accurate prediction of the end point carbon temperature. To address this problem, an improved genetic algorithm is proposed to select the data feature of converter steelmaking process. Firstly, Pearson correlation coefficient is used to measure the important contribution of different features. Then, the objective function is formulated to reflect the correlation between process data feature and terminal carbon temperature. The maximum, minimum, average fitness and random individual fitness of the population are defined by the objective function. In this way, an adaptive crossover mutation probability mechanism is established. This method not only makes the population distribution more reasonable during the iteration optimization, but also improves the late convergence speed to prevent the algorithm from falling into local optimization. Through verification and comparison experiments of data feature selection in actual steel mills, results show that the average time of feature selection is 025 s, the accuracy of temperature error within ±5℃ in terminal prediction is 8567%, and the accuracy of carbon content prediction error within ±001% is 8067%.

Abstract:To reduce the random error of the microelectromechanical system (MEMS) gyroscope, a new method combining the improved empirical mode decomposition (EMD) with the traditional modeling and filtering method is proposed. Firstly, the traditional EMD algorithm is used to decompose the signal into a finite number of intrinsic mode functions (IMF). Based on Pearson correlation coefficient criterion and statistics of noise, a screening mechanism is proposed to divide IMFs into three categories, including noise dominated IMFs, mixed IMFs, and signal IMFs. Then, the timeseries model of the mixed IMFs is formulated. Kalman filtering algorithm is fitted after the modeling is finished. Finally, the mixed IMFs after modeling and filtering and signal IMFs are reconstructed to obtain the denoised signal. Experimental analysis results show that the proposed method has obvious advantages in suppressing the effect of random errors, which can significantly improve the signal quality and the accuracy of the inertial navigation system.

Abstract:In the actual operation of machinery, the normal data are abundant and the fault data are rare. The recognition rate of the minority class is low when the convolutional neural network is used to process these imbalanced data. To solve this problem, an imbalanced fault diagnosis method for machinery based on the cost sensitive convolutional neural network is proposed. Firstly, the intrinsic performance state knowledge is achieved in raw data of machinery through multilevel convolution and pooling operations. Then, the intrinsic performance state knowledge is mapped to mechanical health by fully connected layer. Finally, the cost sensitive loss function is used to set a large cost to the misclassification of the minority class. The effective classification of mechanical imbalanced data is realized. The proposed method is evaluated by tool monitoring data and bearing monitoring data with different imbalanced ratio. Compared with the traditional convolutional neural networks, experimental results show that the recognition rate of minority samples in imbalanced datasets has been improved by more than 22%.

Abstract:In this paper, taking improving the flexibility of peak load regulation and frequency regulation of thermal power units and promoting the consumption of renewable energy sources as the target, the combustion stability and economy of a certain thermal power unit during operation is studied. The adaptive genetic algorithm is adopted to optimize the kernel function parameters and normalization parameters, and a least square support vector machine (LSSVM) boiler combustion process model is established. On the basis of the established LSSVM model, an offline optimized case base is established using the adaptive genetic algorithm. Then, from the perspective of facilitating engineering application, a casebased reasoning (CBR) optimization method is proposed. In consideration of subjective and objective factors, the genetic algorithm is used to optimize the feature weight of CBR, which improves the retrieval accuracy and adaptively retrieves the case matching with the target case from the huge case base. The application of the CBR adaptive optimization algorithm ensures the stable combustion of the unit, and at the same time, considers the boiler combustion efficiency and the concentration of NOx emission. This algorithm reasonably gives the opening instructions of the secondary and tertiary valve baffles and the fixed value of oxygen, and realizes the economic combustion of the boiler. The system was applied to a certain 350MW coalfired generation unit, which simplifies the process of optimization calculation, shortens the optimization time and has high stability. The system is suitable for online realtime optimization.

Abstract:Planning a highefficiency and lowcost threedimensional (3D) flight track has become an urgent problem to be solved for UAV extensive application. Aiming at the problems of track length and lack of smoothness of ant colony algorithm in the flight track planning, this paper proposes the ant colony particle swarm fusion algorithm, which improves the node movement rules in ant colony system, constructs multiple heuristic information and combines the global search ability of particle swarm optimization algorithm. Meanwhile, to solve the problems of dynamic obstacle avoidance and target point change in the flight track, an improved bioinspired neural dynamics model algorithm is proposed, which realizes local online flight track adjustment for the obstacles and target point change in the 3D static optimal flight track. Experiment simulation results show that the ant colony particle swarm fusion algorithm can plan an expected track in 3D static environment. At the same time, the improved bioinspired neural dynamics model algorithm can not only dynamically avoid sudden obstacles, but also track the changes of dynamic target points in real time.

Abstract:The compliant motion of robot helps to improve its safety and stability in interactive motion, which has attracted much more attention in recent years. In this paper, aiming at a wiredriven flexible robotic joint with active and passive compliance, a decoupling compliant control method for the stiffness and position of the wiredriven variablestiffness joint is proposed, which realizes the joint position control and also achieves the uniformity of joint compliance. The stiffness model of the joint is obtained using the Jacobian matrix and the static relationship between the models. The nonlinear equations composed of the mechanical model and stiffness model of the variable stiffness device are solved with optimization method to realize the nonlinear decoupling of the stiffness and position of the variablestiffness joint. Based on the decoupling control, a torque observation method is proposed to realize the torque compensation of the joint and enhance the joint position control ability. The prototype and control system of the wiredriven variablestiffness joint were built, and simulation and experiment analysis verify the feasibility and effectiveness of the above compliant control method.

Abstract:In the magnetic resonance imaging environment, it is difficult for breast biopsy robot to obtain highquality scanned images and precise biopsy needle interventions due to material and drive compatibility. Therefore, a double tendonsheath flexible system with lebus grooves used on the robots, that could meet the requirements of nuclear magnetic compatibility, longdistance and largestroke transmission is proposed in this paper. Firstly, the friction analysis of the single joint is implemented for the double tendonsheath remote transmission system with lebus grooves. A bidirectional coupling transmission model is formulated based on the analysis of nonlinear coupling characteristics of the singlejoint double tendonsheath transmission. The experimental platform of singlejoint bidirectional coupling transmission is established, and the analysis and compensation experiments of singlejoint bidirectional coupling transmission are carried out. Secondly, a multijoint breast biopsy robot experimental platform with double tendonsheath transmission is built. In addition, the analysis of bidirectional coupled motion error and compensation experiment are performed on the multijoint breast biopsy robot with double tendonsheath transmission. After compensation, the average error in the X, Y, and Z directions of the needle tip entering the tissue are 139, 189, and 160 mm, respectively. Experimental results verify that the multijoint breast biopsy robot with double tendon sheath transmission can meet the precision requirements of breast biopsy under nuclear magnetic environment.