Calibration Errors Research Articles

Improving large angle flow measurement accuracy of five-hole probe using novel calibration coefficient and accuracy progressive neural network

Novel calibration coefficients and data-fitting techniques based on a two-stage accuracy progressive neural network were developed to improve the accuracy of a five-hole probe in measuring large-angle flows. By modifying the denominator and numerator of traditional calibration coefficients, the novel coefficients can solve the singularity and multi-value problems of large-angle flow measurement. By training the neural network using both global calibration data and the calibration data of large measurement error points, the two-stage accuracy progressive neural network can effectively improve the measurement accuracy of a five-hole probe when flow separation occurs around the probe head at large flow angles. The experimental results demonstrate that applying the novel calibration coefficients and accuracy progressive neural network ensures that the calibration error of the flow angle is less than 0.8°, and the flow pressure error is less than 0.1 % when the flow angle reaches 50° at low subsonic speeds.

Calibration and validation of a multi-beam LiDAR onboard China Chang’e lunar probe for landing obstacle avoidance

For automatic exploration of specific areas with high scientific value on the moon and other planets, high-precision obstacle detection and hazard avoidance capabilities are required. Three-dimensional (3D) LiDAR can obtain a high-resolution 3D point cloud, offering significant advantages in landing obstacle avoidance. During the landing of the Chang’e lander, a laser 3D imaging sensor (L3DIS) is used to measure and generate accurate topographic maps of the candidate landing area in real-time to help find a safe landing zone. The L3DIS employs 16 beamlets split from a Gaussian laser beam and 16 channels of linear-array avalanche photodiodes within the same optical path and scans the object with a two-axis galvanometer in a field of view of 29°×33°. The calibration of the systematic errors and the performance of obstacle detection on this sensor are conducted in this paper. First, the instrument components and main error sources of the sensor are introduced, and the systematic errors are calibrated based on planar targets, indicating that the sensor’s accuracy is greater than 4 cm, and the corresponding obstacle detection accuracy is better than 12 cm. Second, the ground validation of obstacle detection is demonstrated, and the obtained point cloud is compared with that obtained by a terrestrial laser scanner, indicating that the sensor has good performance. Finally, the performance of onboard measurement and obstacle detection is analyzed, showing that the sensor can identify obvious craters and rocks, i.e., nine major craters with a maximum and minimum depth of 1.36 m and 0.16 m, respectively, and major rocks with a maximum and minimum height of 0.1 m and 0.05 m, respectively, and main landing area with slop less than 10° except for the edges of craters and rocks.

High-Precision Calibration Method and Error Analysis of Infrared Binocular Target Ranging Systems

Infrared binocular cameras, leveraging their distinct thermal imaging capabilities, are well-suited for visual measurement and 3D reconstruction in challenging environments. The precision of camera calibration is essential for leveraging the full potential of these infrared cameras. To overcome the limitations of traditional calibration techniques, a novel method for calibrating infrared binocular cameras is introduced. By creating a virtual target plane that closely mimics the geometry of the real target plane, the method refines the feature point coordinates, leading to enhanced precision in infrared camera calibration. The virtual target plane is obtained by inverse projecting the centers of the imaging ellipses, which are estimated at sub-pixel edge, into three-dimensional space, and then optimized using the RANSAC least squares method. Subsequently, the imaging ellipses are inversely projected onto the virtual target plane, where its centers are identified. The corresponding world coordinates of the feature points are then refined through a linear optimization process. These coordinates are reprojected onto the imaging plane, yielding optimized pixel feature points. The calibration procedure is iteratively performed to determine the ultimate set of calibration parameters. The method has been validated through experiments, demonstrating an average reprojection error of less than 0.02 pixels and a significant 24.5% improvement in calibration accuracy over traditional methods. Furthermore, a comprehensive analysis has been conducted to identify the primary sources of calibration error. Ultimately, this achieves an error rate of less than 5% in infrared stereo ranging within a 55-m range.

Enhancing the reliability of deep learning-based head and neck tumour segmentation using uncertainty estimation with multi-modal images

Objective. Deep learning shows promise in autosegmentation of head and neck cancer (HNC) primary tumours (GTV-T) and nodal metastases (GTV-N). However, errors such as including non-tumour regions or missing nodal metastases still occur. Conventional methods often make overconfident predictions, compromising reliability. Incorporating uncertainty estimation, which provides calibrated confidence intervals can address this issue. Our aim was to investigate the efficacy of various uncertainty estimation methods in improving segmentation reliability. We evaluated their confidence levels in voxel predictions and ability to reveal potential segmentation errors. Approach. We retrospectively collected data from 567 HNC patients with diverse cancer sites and multi-modality images (CT, PET, T1-, and T2-weighted MRI) along with their clinical GTV-T/N delineations. Using the nnUNet 3D segmentation pipeline, we compared seven uncertainty estimation methods, evaluating them based on segmentation accuracy (Dice similarity coefficient, DSC), confidence calibration (Expected Calibration Error, ECE), and their ability to reveal segmentation errors (Uncertainty-Error overlap using DSC, UE-DSC). Main results. Evaluated on the hold-out test dataset (n = 97), the median DSC scores for GTV-T and GTV-N segmentation across all uncertainty estimation methods had a narrow range, from 0.73 to 0.76 and 0.78 to 0.80, respectively. In contrast, the median ECE exhibited a wider range, from 0.30 to 0.12 for GTV-T and 0.25 to 0.09 for GTV-N. Similarly, the median UE-DSC also ranged broadly, from 0.21 to 0.38 for GTV-T and 0.22 to 0.36 for GTV-N. A probabilistic network—PhiSeg method consistently demonstrated the best performance in terms of ECE and UE-DSC. Significance. Our study highlights the importance of uncertainty estimation in enhancing the reliability of deep learning for autosegmentation of HNC GTV. The results show that while segmentation accuracy can be similar across methods, their reliability, measured by calibration error and uncertainty-error overlap, varies significantly. Used with visualisation maps, these methods may effectively pinpoint uncertainties and potential errors at the voxel level.

Measurement System for the Calibration of Accelerometer Arrays

This paper addresses accelerometer array calibration, focusing on determining the errors between multiple sensors. Micro-electromechanical system (MEMS) based triaxial accelerometers, key components of Inertial Measurement Units (IMUs), are used in localization, robotics, and navigation systems. The requirements of these applications necessitate low-cost sensors, which makes MEMS IMUs a reasonable choice. However, these low-cost IMUs are significantly affected by systematic (i.e., bias, misalignment, scale-factor) and random errors. Achieving reliable sensor output depends on the precision of the executed calibration method. While traditional laboratory-based sensor calibration using specialized equipment (i.e., three-axis turntable) is accurate, it is time-consuming and costly. In contrast, in-field calibration techniques, which can be performed using a mechatronic actuator or a robotic arm, have gained popularity. These techniques involve comparing sensor measurements to established reference values. The MEMS sensors are increasingly being used in multi-sensor applications, which demands not only individual sensor error calibration but also important to determine the axis misalignment between the used sensors. During calibration process, various optimization algorithms (e.g., GA, PSO) can also be used to find the error parameters. The proposed measurement system allows for individual calibration of misalignment, bias, and scale factor of the sensor array, and eliminates between-sensor misalignment errors.

Discrimination and Quantification of Cotton and Polyester Textile Samples Using Near-Infrared and Mid-Infrared Spectroscopies.

In the textile industry, cotton and polyester (PES) are among the most used fibres to produce clothes. The correct identification and accurate composition estimate of fibres are mandatory, and environmentally friendly and precise techniques are welcome. In this context, the use of near-infrared (NIR) and mid-infrared (MIR) spectroscopies to distinguish between cotton and PES samples and further estimate the cotton content of blended samples were evaluated. Infrared spectra were acquired and modelled through diverse chemometric models: principal component analysis; partial least squares discriminant analysis; and partial least squares (PLS) regression. Both techniques (NIR and MIR) presented good potential for cotton and PES sample discrimination, although the results obtained with NIR spectroscopy were slightly better. Regarding cotton content estimates, the calibration errors of the PLS models were 3.3% and 6.5% for NIR and MIR spectroscopy, respectively. The PLS models were validated with two different sets of samples: prediction set 1, containing blended cotton + PES samples (like those used in the calibration step), and prediction set 2, containing cotton + PES + distinct fibre samples. Prediction set 2 was included to address one of the biggest known drawbacks of such chemometric models, which is the prediction of sample types that are not used in the calibration. Despite the poorer results obtained for prediction set 2, all the errors were lower than 8%, proving the suitability of the techniques for cotton content estimation. It should be stressed that the textile samples used in this work came from different geographic origins (cotton) and were of distinct presentations (raw, yarn, knitted/woven fabric), which strengthens our findings.

Application of machine learning in on-line calibration of flow measurement errors

With the vigorous construction of an intelligent pipe network, an automatic flowmeter was the key to realize online flow measurement of oil and gas. Ensuring the accuracy of flow measurement was very important to achieve real-time feedback on oil–gas production status for adjusting the oil recovery technology. However, due to the changing flow state of the oil–gas in the pipe and the error of the flow-meter itself, there were some errors during flow measurement. Therefore, based on the machine learning method, this paper corrected the errors caused by the pulsating flow and the flowmeter itself in data drift. The results showed that the toy–LSTM model can measure the pulsating flow accurately. When the number of toy–LSTM model neurons was ranged between 10 and 50, both RMSE and R2 showed good performance. The average of multiple predicted results presented higher accuracy than that of single predicted results, and furthermore, the drop rate can significantly increase the robustness of the toy–LSTM model. It should be noted that the optimal case proposed in this paper was based on a specific condition in the field. In the actual application, it was necessary to comprehensively consider the number of branch pipes, the change in oil well production, the characteristics of the oil–gas–water mixture, and other factors to determine the specific values of the number of neurons, the drop rate, and other parameters. Aiming at the distortion caused by the data drift of the flowmeter itself, the program was designed and data were collected from the software platform, information collector, and base station. The data acquisition of temperature–pressure integrated sensor and flowmeter was carried out in the field. In 20 groups of experiments, the online corrected program based on the Mexican Hat wavelet transform can realize accurate identification and automatic corrected responses, and the corrected time was within 5 min.

A 0.037 pJ K2$\text{K}^2$ 338 pW temperature sensor based on dynamic leakage‐suppression logic

AbstractThis letter introduces an ultra‐low‐power temperature sensor utilizing dynamic leakage‐suppression (DLS) logic and thoroughly analyses its working principle. The sensor effectively tackles the weak pull‐up challenge inherent in DLS logic ensuring its compatibility with standard digital logic. By capitalizing on the super cut‐off attribute of DLS logic, the frontend of the sensor achieves ultra‐low power consumption, without compromising on measurement precision or the breadth of the temperature range. The digital part of the proposed utilizes the output frequency of the sensor's frontend as the clock source, in conjunction with an external 50 Hz reference clock, achieving a low overall power consumption. The frontend of the temperature sensor was fabricated using a 180 nm process, occupying a minimal area of 374 . The digital part of the circuit is implemented using FPGA. Following a two‐point calibration and system error removal, the sensor, operating at a supply voltage of 0.8 V, demonstrated a error of across the temperature range of −20 to 125 . At 25 , the resolution figure of merit of the sensor was 0.037 pJ , with a maximum voltage sensitivity of 4.2 .

Development and evaluation of predictive models for pregnancy risk in UK dairy cows

One suggested approach to improve the reproductive performance of dairy herds is through the targeted management of subgroups of biologically similar animals, such as those with similar probabilities of becoming pregnant, termed pregnancy risk. We aimed to use readily available farm data to develop predictive models of pregnancy risk in dairy cows. Data from a convenience sample of 108 dairy herds in the UK were collated and each herd was randomly allocated, at a ratio of 80:20, to either training or testing data sets. Following data cleaning, there were a total of 78 herds in the training data set and 20 herds in the testing data set. Data were further split by parity into nulliparous, primiparous, and multiparous subsets. An XGBoost model was trained to predict the insemination outcome in each parity subset, with predictors from farm records of breeding, calving and milk recording. Training data comprised 74,511 inseminations in 45,909 nulliparous animals, 86,420 inseminations in 39,439 primiparous animals, and 158,294 inseminations in 32,520 multiparous animals. The final models were evaluated by predicting with the testing data, comprising 31,740 inseminations in 19,647 nulliparous animals, 38,588 inseminations in 16,215 primiparous animals, and 65,049 inseminations in 12,439 multiparous animals. Model discrimination was assessed by calculating the area under receiver operating characteristic curves (AUC); model calibration was assessed by plotting calibration curves and compared across test herds by calculating the expected calibration error (ECE) in each test herd. The models were unable to discriminate between insemination outcomes with high accuracy, with an AUC of 0.63, 0.59 and 0.62 in the nulliparous, primiparous and multiparous subsets, respectively. The models were generally well-calibrated, meaning the model-predicted pregnancy risks were similar to the observed pregnancy risks. The mean (SD) ECE in the test herds was 0.038 (0.023), 0.028 (0.012) and 0.020 (0.008) in the nulliparous, primiparous and multiparous subsets respectively. The predictive models reported here could theoretically be used to identify subgroups of animals with similar pregnancy risk to facilitate targeted reproductive management; or provide information about cows' relative pregnancy risk compared with the herd average, which may support on-farm decision-making. Further research is needed to evaluate the generalizability of these predictive models and understand the source of variation in ECE between herds; however, this study demonstrates that it is possible to accurately predict pregnancy risk in dairy cows using readily available farm data.

Research on calibration method for measurement error of multiple rate carrier energy meter

Abstract With the rapid development of smart grid technology, the multi-rate carrier watt-hour meter is an important measurement tool in power systems; its accuracy directly affects economic benefits and energy management efficiency. This paper proposes a calibration method of measurement error for a multi-rate carrier watt-hour meter to improve the measurement accuracy and reliability of the watt-hour meter under changing load conditions. The multi-rate carrier energy data is collected in real-time by high-precision data acquisition equipment to ensure the comprehensiveness and accuracy of the data. Using advanced data analysis technology to extract the amplitude characteristics of abnormal metering data, on this basis, the output deviation of equipment without input signal is eliminated, and the metering error calibration of the electric energy meter is completed. The experimental results show that this method can significantly improve the metering accuracy of multi-rate carrier watt-hour meters under different power consumption conditions and has good adaptability and practical value.

A Newton method with characteristic value correction for geometric error calibration of parallel mechanism

A Newton method with characteristic value correction for geometric error calibration of parallel mechanism

HiFAST: An H i Data Calibration and Imaging Pipeline for FAST. II. Flux Density Calibration

Accurate flux density calibration is essential for precise analysis and interpretation of observations across different observation modes and instruments. In this research, we first introduce the flux calibration model that incorporated in HiFAST pipeline, and designed for processing H i 21 cm spectra. Furthermore, we investigate different calibration techniques and assess the dependence of the gain parameter on the time and environmental factors. A comparison is carried out in various observation modes (e.g., tracking and scanning modes) to determine the flux density gain (G), revealing insignificant discrepancies in G among different methods. Long-term monitoring data shows a linear correlation between G and atmospheric temperature. After subtracting the G–Temperature dependence, the dispersion of G is reduced to <3% over a one-year timescale. The stability of the receiver response of Five-hundred-meter Aperture Spherical radio Telescope (FAST) is considered sufficient to facilitate H i observations that can accommodate a moderate error in flux calibration (e.g., > ∼ 5%) when utilizing a constant G for calibration purposes. Our study will serve as a useful addition to the results provided by Jiang et al. Detailed measurement of G for the 19 beams of FAST, covering the frequency range 1000–1500 MHz, can be found on the HiFAST homepage: https://hifast.readthedocs.io/fluxgain.

Measuring Tilt with an IMU Using the Taylor Algorithm

This article addresses the important problem of tilt measurement and stabilization. This is particularly important in the case of drone stabilization and navigation in underwater environments, multibeam sonar mapping, aerial photogrammetry in densely urbanized areas, etc. The tilt measurement process involves the fusion of information from at least two different sensors. Inertial sensors (IMUs) are unique in this context because they are both autonomous and passive at the same time and are therefore very attractive. Their calibration and systematic errors or bias are known problems, briefly discussed in the article due to their importance, and are relatively simple to solve. However, problems related to the accumulation of these errors over time and their autonomous and dynamic correction remain. This article proposes a solution to the problem of IMU tilt calibration, i.e., the pitch and roll and the accelerometer bias correction in dynamic conditions, and presents the process of calculating these parameters based on combined accelerometer and gyroscope records using a new approach based on measuring increments or differences in tilt measurement. Verification was performed by simulation under typical conditions and for many different inertial units, i.e., IMU devices, which brings the proposed method closer to the real application context. The article also addresses, to some extent, the issue of navigation, especially in the context of dead reckoning.

System error analysis and target localization of distributed pulse radars

AbstractTarget localization for distributed radars requires precise knowledge of the calibration errors in time, frequency, and space. Previous studies focus on the sensor location error and sensor directional error in space. This paper concentrates on the system errors in time and frequency which directly affect the targets estimation of range and velocity. Based on linear frequency modulation waveform, the impact of the system errors including sample interval, pulse repetition interval, and centre frequency are analysed. An alternating optimization algorithm is proposed to achieve the target localization, velocity estimation and system errors estimation simultaneously. Simulation shows the effectiveness of the algorithm and the influence on localization due to varying number of targets and radars.

Machine-learning-based prediction of disability progression in multiple sclerosis: An observational, international, multi-center study.

Disability progression is a key milestone in the disease evolution of people with multiple sclerosis (PwMS). Prediction models of the probability of disability progression have not yet reached the level of trust needed to be adopted in the clinic. A common benchmark to assess model development in multiple sclerosis is also currently lacking. Data of adult PwMS with a follow-up of at least three years from 146 MS centers, spread over 40 countries and collected by the MSBase consortium was used. With basic inclusion criteria for quality requirements, it represents a total of 15, 240 PwMS. External validation was performed and repeated five times to assess the significance of the results. Transparent Reporting for Individual Prognosis Or Diagnosis (TRIPOD) guidelines were followed. Confirmed disability progression after two years was predicted, with a confirmation window of six months. Only routinely collected variables were used such as the expanded disability status scale, treatment, relapse information, and MS course. To learn the probability of disability progression, state-of-the-art machine learning models were investigated. The discrimination performance of the models is evaluated with the area under the receiver operator curve (ROC-AUC) and under the precision recall curve (AUC-PR), and their calibration via the Brier score and the expected calibration error. All our preprocessing and model code are available at https://gitlab.com/edebrouwer/ms_benchmark, making this task an ideal benchmark for predicting disability progression in MS. Machine learning models achieved a ROC-AUC of 0⋅71 ± 0⋅01, an AUC-PR of 0⋅26 ± 0⋅02, a Brier score of 0⋅1 ± 0⋅01 and an expected calibration error of 0⋅07 ± 0⋅04. The history of disability progression was identified as being more predictive for future disability progression than the treatment or relapses history. Good discrimination and calibration performance on an external validation set is achieved, using only routinely collected variables. This suggests machine-learning models can reliably inform clinicians about the future occurrence of progression and are mature for a clinical impact study.

Calibrating the global network of gravitational wave observatories via laser power calibration at NIST and PTB

Current gravitational wave (GW) observatories rely on photon calibrators that use laser radiation pressure to generate displacement fiducials used to calibrate detector output signals. Reducing calibration uncertainty enables optimal extraction of astrophysical information such as source distance and sky position from detected signals. For the ongoing O4 observing run that started on 24 May 2023, the global GW detector network is employing a new calibration scheme with transfer standards calibrated at both the National Institute of Standards and Technology (NIST) and the Physikalisch-Technische Bundesanstalt (PTB). These transfer standards will circulate between the observatories and the metrology institutes to provide laser power calibration traceable to the International System of Units (SI) and enable assessment and reduction of relative calibration errors for the observatory network. The Laser Interferometer Gravitational-Wave Observatory (LIGO) project and the Virgo project are currently participating in the new calibration scheme. The Large-scale Cryogenic Gravitational-wave Telescope project (KAGRA) is expected to join in 2024, with the LIGO Aundha Observatory in India joining later. Before implementing this new scheme, a NIST-PTB bilateral comparison was conducted. It validated the scale representation by both laboratories, with a degree of equivalence of −0.2% and an associated expanded uncertainty of 0.32% (k = 2) which is significantly lower than previous studies. We describe the transfer of power sensor calibration, including detailed uncertainty estimates, from the transfer standards calibrated by NIST and PTB to the sensors operating continuously at the interferometer end stations. Finally, we discuss the ongoing calibration of Pcal-induced displacement fiducials for the O4 observing run. Achieved combined standard uncertainty levels as low as 0.3% facilitate calibrating the interferometer output signals with sub-percent accuracy.

Asynchronous calibration of a CT scanner for bone mineral density estimation: sources of error and correction.

The estimation of BMD with CT scans requires a calibration method, usually based on a phantom. In asynchronous calibration, the phantom is scanned separately from the patient. A standardized acquisition protocol must be used to avoid variations between patient and phantom. However, variations can still be induced, for example, by temporal fluctuations or patient characteristics. Based on the further use of 739 forensic and 111 clinical CT scans, this study uses the proximal femur BMD value ("total hip") to assess asynchronous calibration accuracy, using in-scan calibration as ground truth. It identifies the parameters affecting the calibration accuracy and quantifies their impact. For time interval and table height, the impact was measured by calibrating the CT scan twice (once using the phantom scan with closest acquisition parameters and once using a phantom scan with standard values) and comparing the calibration accuracy. For other parameters such as body weight, the impact was measured by computing a linear regression between parameter values and calibration accuracy. Finally, this study proposes correction methods to reduce the effect of these parameters and improve the calibration accuracy. The BMD error of the asynchronous calibration, using the phantom scan with the closest acquisition parameters, was -1.2 ± 1.7% for the forensic and - 1.6 ± 3.5% for the clinical dataset. Among the parameters studied, time interval and body weight were identified as the main sources of error for asynchronous calibration, followed by table height and reconstruction kernel. Based on these results, a correction method was proposed to improve the calibration accuracy.

Enhancement of Foley Catheter Assembly Section Quality using Six Sigma (A Case Study of an Indonesian Medical Device Company)

This study aims to enhance the quality of the Foley catheter assembly section at PT XYZ, an Indonesian medical device manufacturer, using Six Sigma methodology. PT XYZ currently faces an 8% defect rate in its Foley catheter assembly section, surpassing the acceptable 3% threshold. The current Sigma level of the Shrink process within the assembly section is 2.36, while the Balloon Vulcanizing and Assembly process stands at 2.68. The Six Sigma DMAIC (Define, Measure, Analyze, Improve, Control) framework was used to identify and address root causes of defects. Data collection methods included interviews, questionnaires, and field observations, revealing key issues such as incorrect tray loading, machine calibration errors, operator skill variability, and inadequate handling techniques. Proposed improvements include updating Standard Operating Procedures (SOPs), implementing SCAMPER for calibration, comprehensive training programs, and incentive schemes. Additional measures involve template overlays for precise cutting, vendor maintenance, and control charts for consistency. The Control phase ensures sustainability through documentation, training, and monitoring using control charts and Failure Modes and Effects Analysis (FMEA). Implementing these solutions is expected to reduce the defect rate and elevate the Sigma level of both processes to 3, enhancing process efficiency and product quality. This will improve operational efficiency and customer satisfaction, providing a robust framework for PT XYZ to elevate its quality control and contribute to better healthcare device manufacturing in Indonesia.

Multidimensional Quality Characteristics of Sichuan South-Road Dark Tea and Its Chemical Prediction

The distinctive quality of Sichuan south-road dark tea (SSDT) is gradually disappearing with processing innovation. Here, near-infrared (NIR) spectroscopy (NIRS) and spectrofluorometric techniques were utilized to determine the spectral characteristics of dried SSDT and its brew, respectively. Combined with chemical analysis, the multidimensional quality characteristics of SSDT will be presented. Finally, the NIR spectral fingerprint of dried SSDT was observed, with Kangzhuan (KZ) and Jinjian (JJ) showing a very similar NIR spectrum. The SiPLS models effectively predicted the levels of theabrownin, caffeine, and epigallocatechin gallate, based on the NIR spectrum, with root-mean-square errors of calibration of 0.15, 0.12, and 0.02 for each chemical compound, root-mean-square errors of prediction of 0.20, 0.09, and 0.03, and both corrected and predicted correlation coefficients greater than 0.90. Meanwhile, the fluorescence characteristics of the SSDT brew were identified based on the parallel factor analysis for the fluorescence excitation–emission matrix (EEM). The KZ and JJ brews could be classified with 100% accuracy using extreme-gradient-boosting discriminant analysis. The integration of NIRS and fluorometric EEM seems to be a powerful technique for characterizing SSDTs, and the results can greatly benefit the production and quality control of SSDTs.

Design and Self-Calibration Method of a Rope-Driven Cleaning Robot for Complex Glass Curtain Walls

Rope-driven robots are increasingly used to safely and efficiently clean complex glass curtain walls. However, continuous global cleaning is difficult for most robots because of their poor performance in overcoming obstacles and adapting to curved surfaces, and an inconvenient winch calibration on complex surfaces further burdens such work. This paper presents a 3-DOF rope-driven robot and a winch self-calibration method for efficient cleaning. The robot, by integrating a 5-rope parallel configuration, a self-adaptive cleaning body, and a self-compensating driving winch, is designed to perform continuous, compliant, and accurate spatial motion on curved walls with obstacles. By deducing the kinematic model, the constraint relationship related to orderly arranged winch positions, maneuverable body positions, and accessible rope lengths is established during the robot tracking 3D trajectory. Combining the established relationship, a series of regulations are formulated for easily acquiring body positions and rope lengths, and then a self-calibration method is proposed by accurately calculating winch positions without using professional instruments. Experimental results show that the robot can perform global and precise movement on complex glass surfaces. Applying the proposed method, the maximum winch calibration error is 6.6 mm, and the maximum body tracking error is controlled within 9.6 mm.