A method, device, equipment, medium and product for assembling male and female heads of a flexible flat cable

Abstract

The application discloses a kind of universally used for soft cable male and female head assembly method, device, equipment, medium and product, it is related to industrial automation assembly field.The method comprises: based on Gaussian process regression model, real-time acquisition tactile information data is decomposed and interpolation processing each electrode;Gaussian process regression model is using each electrode independent modeling strategy, based on historical tactile information data is carried out statistical characteristic proportion analysis, to adaptively configure kernel function to adapt the nonlinear response of each electrode and the interpolation obtained after;By principal component analysis method, construct feature subspace, and using subspace alignment method to the structure consistency constraint of multi-electrode tactile feature, and carry out linear alignment processing;Based on Mahalanobis distance measurement model, determine the Mahalanobis distance between reference benchmark vector and aligned feature vector, to measure contact offset amount.The application can realize the position deviation perception of cross condition, high precision, and have universality.

Description

Technical Field

[0001] This application relates to the field of industrial automation assembly, and in particular to an assembly method, apparatus, equipment, medium, and product applicable to male and female connectors of flexible flat cable. Background Technology

[0002] In the field of industrial automation assembly, especially in the precision assembly of male and female connectors for 3C electronic products, current technology faces the following two key technical challenges: 1. In the working environment, high-sensitivity tactile sensors may encounter various abnormal data during data acquisition due to the physical structure of the flexible flat cable and noise interference in the device environment, which can easily obscure force characteristic information.

[0003] 2. Highly integrated sensor actuators are integrated into the end effector of industrial robotic arms to perform latching and data acquisition tasks. In the actual industrial production environment of precision assembly of male and female connectors for 3C electronic products, sensors have a theoretical lifespan, making sensor replacement a necessity.

[0004] To address the aforementioned issues, current technologies typically integrate high-resolution tactile sensors at the end effector of a robotic arm to collect force signals from the male and female connectors of a flexible flat cable during pressing and fastening processes in real time. Based on the collected data, peak force values, force slopes, or temporal characteristics are extracted to determine the assembly status or estimate assembly offset. However, in actual industrial assembly environments, the assembly process is inevitably affected by factors such as mechanical vibration, electromagnetic interference, structural gaps, and environmental disturbances. This causes a large amount of random noise and non-stationary interference components to be superimposed on the tactile sensor output signal, weakening or even submerging the effective force characteristics, thereby reducing the accuracy and stability of offset recognition based on force signals.

[0005] On the other hand, to improve the accuracy of offset prediction, current technologies generally rely on data-driven modeling methods, which involve collecting a large amount of calibration data under specific assembly scenarios and sensor conditions to train the prediction model. When sensor performance drifts, ages, or needs to be replaced due to failure, the new sensor differs from the original sensor in terms of individual electrode responses, zero drift, and noise characteristics, making it difficult for the original model to be directly adapted. Often, data acquisition and model training need to be repeated. This not only increases system deployment and maintenance costs but also causes production interruptions, reducing the overall flexibility and engineering practicality of the assembly system. Summary of the Invention

[0006] The purpose of this application is to provide an assembly method, apparatus, equipment, medium, and product applicable to male and female connectors of flexible flat cable, which can achieve high-precision position deviation sensing across working conditions and has versatility.

[0007] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a general assembly method for male and female connectors of flexible flat cable, which is applied in the assembly scenario of male and female connectors of flexible flat cable; the general assembly method for male and female connectors of flexible flat cable includes: Real-time acquisition of tactile information data; the tactile information data includes multi-electrode time-series tactile signals acquired based on a multi-electrode tactile sensor installed on the end effector; Based on a Gaussian process regression model, the tactile information data is decomposed and interpolated electrode by electrode to obtain multi-electrode tactile features. The Gaussian process regression model adopts an electrode-by-electrode independent modeling strategy, performs statistical feature ratio analysis based on historical tactile information data, and adaptively configures the kernel function to adapt to the nonlinear response and interpolation of each electrode. The historical tactile information data includes tactile information data with known offsets. The multi-electrode tactile features include the offsets of each electrode after continuous interpolation. The feature subspace is constructed by principal component analysis, and the multi-electrode tactile features are subjected to structural consistency constraints and linear alignment processing by a subspace alignment method to obtain the aligned feature vector. The Mahalanobis distance between the reference vector and the aligned feature vector is determined based on the Mahalanobis distance metric model, and the contact offset is measured based on the Mahalanobis distance. The contact offset represents the position perception deviation of the assembly and is used for subsequent assembly position adjustment. The Mahalanobis distance metric model is determined based on the aligned feature vector corresponding to historical tactile information data. The Mahalanobis distance metric model is used to measure the degree of offset information.

[0008] Secondly, this application provides an assembly device universally applicable to male and female connectors of flexible flat cables, comprising: The data acquisition module is used to collect tactile information data in real time; the tactile information data includes multi-electrode time-series tactile signals acquired based on a multi-electrode tactile sensor installed on the end effector; The data augmentation module is used to perform electrode-by-electrode decomposition and interpolation processing on the tactile information data based on a Gaussian process regression model to obtain multi-electrode tactile features. The Gaussian process regression model adopts an electrode-by-electrode independent modeling strategy, performs statistical feature ratio analysis based on historical tactile information data, and adaptively configures the kernel function to adapt to the nonlinear response and interpolation of each electrode. The historical tactile information data includes tactile information data with known offsets. The multi-electrode tactile features include the offsets of each electrode after continuous interpolation. The principal component analysis and alignment module is used to construct a feature subspace using principal component analysis and to perform structural consistency constraints on the multi-electrode tactile features using a subspace alignment method, and to perform linear alignment processing to obtain the aligned feature vector. The measurement module is used to determine the Mahalanobis distance between the reference vector and the aligned feature vector based on the Mahalanobis distance measurement model, and to measure the contact offset based on the Mahalanobis distance; the contact offset represents the position perception deviation of the assembly, which is used for subsequent assembly position adjustment; wherein, the Mahalanobis distance measurement model is determined based on the aligned feature vector corresponding to historical tactile information data; the Mahalanobis distance measurement model is used to measure the degree of offset information.

[0009] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the assembly method for male and female connectors of flexible flat cables described above.

[0010] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the assembly method for male and female connectors of flexible flat cable described above.

[0011] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the assembly method for male and female connectors of flexible flat cable described above.

[0012] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a method, apparatus, equipment, medium, and product applicable to male and female connectors of flexible flat cables. Based on a Gaussian process regression model, it performs electrode-by-electrode decomposition and interpolation processing on real-time acquired tactile information data. The Gaussian process regression model adopts an electrode-by-electrode independent modeling strategy, performs statistical feature ratio analysis based on historical tactile information data, and adaptively configures the kernel function to adapt to the nonlinear response of each electrode and the interpolation. A feature subspace is constructed by principal component analysis, and a subspace alignment method is used to constrain the structural consistency of multi-electrode tactile features to extract the feature subspace, and linear alignment processing is performed to obtain aligned feature vectors. The Mahalanobis distance between the reference vector and the aligned feature vector is determined based on a Mahalanobis distance metric model, thereby measuring the contact offset. This application decomposes the multi-electrode tactile nonlinear response problem into an electrode-by-electrode Gaussian process regression model, and combines subspace alignment to achieve structural consistency constraints on feature distribution across batches and assembly states. While effectively suppressing tactile noise interference, it significantly reduces the dependence on high-density data acquisition and repeated sensor calibration, thereby improving the accuracy, stability, and universal interchangeability of offset sensing during the assembly of male and female connectors of flexible flat cable. As a result, it can achieve high-precision position deviation sensing across working conditions and has universality. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a structural diagram illustrating a common assembly method for male and female connectors of flexible flat cable. Figure 2 This is a schematic diagram of the overall design for an assembly method applicable to both male and female connectors of flexible flat cable. Figure 3 This is a schematic diagram of the framework of an adaptive high-precision interpolation method based on Gaussian process regression. Figure 4 This is a schematic diagram of the prediction process for a consistency processing method based on subspace alignment. Figure 5 This is a structural diagram of an assembly device for general-purpose male and female connectors of flat cable; Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0015] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] This application addresses the significant nonlinear changes in multi-electrode tactile signals caused by the offset of the male and female connectors during the assembly of flexible printed circuits (FPCs), as well as the issues related to different experimental batches, assembly states, or smart end conditions. It constructs a multi-electrode modeling and processing mechanism for sensing the assembly offset of male and female connectors, in order to ensure high prediction accuracy while also meeting the long-term stable operation requirements in industrial settings.

[0017] Current technologies typically use tactile data collected under fixed discrete offset conditions to directly train the offset prediction model. This approach is limited by the number of discrete sampling points. In actual working conditions, a small number of data sampling points makes it difficult to accurately describe the nonlinear evolution of the tactile signal during continuous offset changes, and various noise interferences can easily occur during the data sampling process.

[0018] On the other hand, different electrodes exhibit significant differences in the mechanical paths, sensitivity, and noise characteristics experienced during elastic contact. Traditional modeling methods are prone to local overfitting or overall oversmoothing, making it difficult to accurately describe the force characteristic relationships of multiple electrodes under adjacent discrete offset labels, thereby weakening the accuracy of offset sensing. Furthermore, when sensors are replaced or assembly conditions change, the overall distribution of tactile data drifts, making it difficult to directly reuse the original model and severely restricting the versatility and interchangeability of smart terminals in industrial scenarios.

[0019] To address the aforementioned issues, the core idea of ​​this application is to decompose the complex multi-electrode nonlinear tactile response problem into an electrode-by-electrode Gaussian process regression modeling problem, and on this basis, to achieve cross-batch and cross-state feature distribution consistency constraints through subspace alignment.

[0020] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] In one exemplary embodiment, a method for assembling male and female connectors of flexible flat cables is provided, which is applied in the context of assembling male and female connectors of flexible flat cables.

[0022] like Figure 1As shown, the assembly method for the general-purpose male and female connectors of flexible flat cables includes: Step 100: Real-time acquisition of tactile information data. The tactile information data includes multi-electrode time-series tactile signals acquired based on a multi-electrode tactile sensor installed on the end effector.

[0023] Step 200: Based on the Gaussian process regression model, the tactile information data is decomposed and interpolated electrode by electrode to obtain multi-electrode tactile features. The Gaussian process regression model adopts an electrode-by-electrode independent modeling strategy, performs statistical feature ratio analysis based on historical tactile information data, and adaptively configures the kernel function to adapt to the nonlinear response of each electrode and the interpolation. The historical tactile information data includes tactile information data with known offsets. The multi-electrode tactile features include the offsets of each electrode after continuous interpolation.

[0024] The configuration strategy of the adaptive configuration kernel function is determined based on the electrode characteristic indicators; the electrode characteristic indicators include: dynamic range of response value, standard deviation, first-order difference statistical characteristics, and noise ratio.

[0025] Specifically, if the noise ratio is not greater than a preset threshold, a radial basis function kernel is used; otherwise, a Matérn kernel function is used.

[0026] Step 300: Construct a feature subspace using principal component analysis, and use a subspace alignment method to constrain the structural consistency of the multi-electrode tactile features, and perform linear alignment processing to obtain the aligned feature vectors.

[0027] As an optional implementation, a feature subspace is constructed using principal component analysis, and a subspace alignment method is employed to constrain the structural consistency of the multi-electrode tactile features, followed by linear alignment processing to obtain aligned feature vectors. Specifically, this includes: A feature subspace is constructed using principal component analysis. The feature subspace includes a source domain subspace and a target domain subspace. Principal components are selected from the dataset based on a preset number of principal components to obtain the source domain subspace and the target domain subspace. The dataset consists of any two sets of data from different sources in the multi-electrode tactile features. Different sources represent different collection batches or assembly conditions.

[0028] The linear alignment matrix is ​​optimized using a subspace alignment method to obtain the optimized linear alignment matrix; the linear alignment matrix is ​​determined based on the feature subspace.

[0029] The source domain subspace is mapped to the target domain subspace based on the optimized linear alignment matrix to achieve linear alignment and obtain the aligned feature vector.

[0030] In one embodiment, a subspace alignment method is used to optimize the linear alignment matrix to obtain an optimized linear alignment matrix, specifically including: By minimizing the eigenspace mapping error in the Frobenius norm sense The linear alignment matrix is ​​optimized and solved to obtain a closed-form solution, which is used as the optimized linear alignment matrix.

[0031] The expression for the optimized linear alignment matrix is: .

[0032] in, This is the optimized linear alignment matrix; For the source domain subspace; For the target domain subspace; It is the Frobenius norm; This is a transpose.

[0033] when When the problem is irreversible, the Moore–Penrose pseudo-inverse is used for the solution.

[0034] Step 400: Determine the Mahalanobis distance between the reference vector and the aligned feature vector based on the Mahalanobis distance metric model, and measure the contact offset based on the Mahalanobis distance. The contact offset characterizes the position perception deviation of the assembly and is used for subsequent assembly position adjustment; wherein, the Mahalanobis distance metric model is determined based on the aligned feature vector corresponding to historical tactile information data; the Mahalanobis distance metric model is used to measure the degree of offset information.

[0035] In one embodiment, the electrode-by-electrode independent modeling strategy is a nonparametric regression method based on Bayesian theory, which models the nonlinear relationship of each electrode to model the nonlinear response of each electrode as a function related to the offset; the expression of the function is: .

[0036] in, For the first The tactile information data of each electrode represents the corresponding nonlinear response value; This is the offset; It is a mean function; For kernel functions; For Gaussian process regression.

[0037] In practical applications, the specific operational steps of the method mentioned in this application are as follows: Step 1: During the assembly of flexible printed circuits (FPCs), a multi-electrode tactile sensor mounted on an intelligent end effector is used to collect data on the assembly process of the FPC and connector under different preset contact offset conditions. Before the assembly experiment, random offset position labels are determined in the positive and negative directions of the assembly through pose control, with the ideal alignment position as the zero point. Under each preset offset condition, the intelligent end effector is controlled to complete one full pressing contact process, and the timing tactile signals from the multi-electrode tactile sensor are collected simultaneously. To reduce the impact of transient shocks and noise interference in the early stages of assembly, only the stable contact phase of the contact process is selected as the effective data range.

[0038] Step 2: Constructing the Training Dataset. Based on the raw tactile data (historical tactile information data) collected in Step 1, a training dataset for regression modeling is constructed. The multi-electrode steady-state response tactile data (multi-electrode time-series tactile signals) obtained under different offset conditions are categorized and organized according to offset labels. The offset values ​​are used as supervision labels, and the corresponding multi-electrode steady-state response tactile data are used as feature inputs. For each electrode, its response sequence under different offset conditions is extracted, thus decomposing the original multi-electrode modeling problem into multiple electrode-by-electrode one-dimensional regression modeling problems. Through the above processing, a standardized training dataset for electrode-by-electrode regression modeling is obtained, providing a foundation for the subsequent construction of the Gaussian process regression model.

[0039] Step 3: Construct independent Gaussian process regression models and perform continuous offset interpolation based on these models. To address the heterogeneity of the response characteristics of each electrode, an electrode-by-electrode independent modeling strategy is adopted, using the offset as input to construct a separate Gaussian process regression model for each electrode. By analyzing the statistical characteristics of the tactile information data, such as the noise ratio, the kernel function is adaptively configured to accommodate the nonlinear responses of different electrodes. The trained model is used to transform discrete experimental data into continuous, high-density interpolation curves and obtain corresponding uncertainty estimates, thereby solving the problem of sparse original data.

[0040] Step 4: Apply subspace alignment to constrain the structural consistency of the interpolated high-dimensional multi-electrode tactile features. To address the data distribution differences between the original and spare sensors, a subspace alignment method is introduced to process the interpolated high-dimensional features. This step uses principal component analysis to extract the feature subspaces of the source domain data (original) and the target domain data (spare), respectively, and calculates the optimal linear alignment matrix. This maps the source domain features to a common subspace semantically consistent with the target domain, thereby eliminating the distribution shift caused by differences in experimental batches or hardware without introducing additional supervisory information.

[0041] Step 5: Complete the preparation of prediction parameters. Construct a statistical space to measure the degree of offset based on the aligned feature data. Specifically, calculate the covariance matrix using the aligned feature matrix and establish a Mahalanobis distance metric model accordingly. At the same time, select a standard sample with zero offset as a reference benchmark, and obtain a reference vector after the same subspace projection transformation. This reference vector will serve as the unified physical zero point for subsequent calculation of offset distance and determination of offset direction.

[0042] Step 6: Collect real-time tactile information data. Using the steps above, align the spare part data features to the original part data, and accurately predict the offset based on the Mahalanobis distance metric space. In actual assembly or application scenarios, the system collects real-time tactile response data (tactile information data) from the current sensor and extracts steady-state features. The system projects this real-time feature onto the target subspace, uses a pre-calculated alignment matrix to align the features with the original part data, and then calculates the Mahalanobis distance between its real-time feature vector and the reference vector. Finally, based on this distance metric, the system directly regresses and predicts the current contact offset, achieving high-precision position deviation sensing across operating conditions.

[0043] Through the synergistic effect of the above-mentioned electrode-by-electrode Gaussian process regression interpolation modeling and subspace alignment, this application has achieved accurate modeling of the continuous change law of multi-electrode tactile signals in principle, while significantly reducing the impact of sensor replacement or assembly condition changes on the offset prediction model, thus forming an intelligent end-point sensing and modeling solution with universality, interchangeability and engineering practical value.

[0044] Figure 2 This is a schematic diagram of the overall design of the method mentioned in this application. Addressing the nonlinear response and general interchangeability issues of multi-electrode tactile signals caused by contact position offsets in flexible printed circuit (FPC) assembly, the method first extracts tactile information data during the contact stabilization phase, constructs a tactile dataset with offset labels, and employs an electrode-by-electrode independent modeling strategy to overcome heterogeneity interference between electrodes. Using Gaussian Process Regression (GPR) technology, high-precision nonlinear interpolation and uncertainty estimation of electrode responses are achieved through an adaptive kernel function (Radial Basis Function (RBF) or Matérn) configuration based on the signal-to-noise ratio. After interpolating the original and spare parts data, a subspace alignment (SA) method is introduced. Principal component analysis is used to construct a feature subspace and optimize the alignment matrix, eliminating data distribution differences between different experimental batches. This allows for robust offset prediction across operating conditions based on the aligned Mahalanobis distance metric.

[0045] Figure 3This is a schematic diagram of the adaptive high-precision interpolation method framework based on Gaussian process regression. Under different preset offset conditions, a multi-electrode tactile sensor is used to collect tactile signals during the elastic contact process between the flexible cable and the connector. Considering the presence of transient impacts and noise interference during assembly, this application only extracts sensor data and offset labels during the contact stabilization phase as modeling input. Specifically, the sampled data corresponding to each offset is statistically averaged over time to obtain the steady-state response value of each electrode at that offset. Through the above processing, each offset sample is represented as a 12-dimensional electrode response vector, forming the basic dataset for subsequent modeling. Considering the significant differences in the position, force path, and sensitivity of different electrodes during the elastic contact process, their responses often exhibit non-uniform, nonlinear, and electrode-dependent characteristics as they change with offset. Therefore, this application does not adopt a multi-electrode independent modeling approach, but instead introduces an electrode-by-electrode independent modeling strategy, i.e.: 1. Use the offset as a one-dimensional input variable; 2. Build an independent regression model for the response value of each electrode; 3. Avoid interference from the heterogeneity between electrodes on the model fitting; 4. This strategy improves the ability to characterize local nonlinear features while maintaining the simplicity of the model structure.

[0046] Gaussian process regression is a nonparametric regression method based on Bayesian theory. Its core advantage lies in its ability to model complex nonlinear relationships without requiring a predefined functional form, and it naturally provides an estimate of uncertainty in the prediction results. For the i-th electrode, its response value is modeled as a function of the offset: .

[0047] in, It is a mean function. It is the covariance function (kernel function).

[0048] To avoid overfitting or excessive smoothing caused by uniform kernel function parameters, this application introduces an adaptive kernel function configuration strategy based on the statistical characteristics of electrode response data. Specifically, the following characteristic indices are calculated for each electrode: 1. Dynamic range of response value; 2. Standard deviation; 3. Statistical properties of the first-order difference; 4. Noise ratio (used to measure the severity of local changes). Based on the noise ratio, electrodes are divided into two modeling cases: Case 1: The response changes relatively smoothly (low noise ratio): For electrodes whose response changes relatively smoothly with offset, a radial basis function (RBF) kernel is used to obtain smoother interpolation results and limit the upper bound of the noise term to avoid the model being overly sensitive to local perturbations.

[0049] Scenario 2: Response changes drastically (high noise ratio): For electrodes exhibiting significant non-smooth changes or local fluctuations, the Matérn kernel function is employed to enhance the model's ability to express non-smooth functions and to appropriately relax the boundary of the noise term, thereby improving the model's robustness.

[0050] Figure 4 This is a schematic diagram of the prediction process for the subspace alignment-based consistency processing method. After completing the training of the electrode-by-electrode GPR model, predictions are made for each electrode for any set of target offsets, yielding: the mean of the interpolated electrode response; and the corresponding prediction standard deviation as an uncertainty measure. Prediction uncertainty reflects the confidence level of the model across different offset intervals, providing important reference information for subsequent offset regression, anomaly detection, and subspace alignment. After completing continuous offset interpolation, to reduce the impact of differences in tactile data distribution across different experimental batches or assembly states on subsequent offset predictions, this application further introduces a subspace alignment (SA) method to constrain the distribution consistency of the interpolated high-dimensional tactile features.

[0051] First, for two datasets from different sources but with consistent semantics (e.g., data from different acquisition batches or under different assembly conditions), their original tactile feature matrices are constructed. Each sample consists of a 12-dimensional electrode steady-state response. To reduce noise interference and extract key structural information, principal component analysis (PCA) is performed on both datasets, retaining the first three principal components to obtain the source domain subspace. With the target domain subspace .in, For the sample.

[0052] Subsequently, the optimal linear alignment matrix is ​​solved by minimizing the subspace mapping error in the Frobenius norm sense. .

[0053] .

[0054] This optimization problem has a closed-ended solution: .

[0055] when When irreversibility is not possible, the Moore-Penrose pseudo-inverse is used for solving. Through this alignment matrix, source domain samples are mapped to a feature space with semantic consistency with the target domain, thereby achieving cross-condition structural alignment without introducing additional supervision information.

[0056] After completing subspace alignment, this application does not directly use the original alignment features for modeling, but further constructs a structural description based on the aligned source domain data. Specifically, the aligned feature matrix is ​​used to calculate the covariance matrix, and a Mahalanobis distance metric space is constructed accordingly for subsequent offset sensing and regression modeling. Furthermore, a sample with zero spare part offset is selected as a reference state, and the representation of the reference vector in the aligned subspace is obtained through the same PCA transformation and subspace alignment mapping. This reference vector serves as the benchmark for subsequent Mahalanobis distance calculation and offset direction determination, enabling tactile responses under different offset states to be compared under a unified statistical meaning.

[0057] This application decomposes the multi-electrode tactile nonlinear response problem into an electrode-by-electrode Gaussian process regression model, and combines subspace alignment to achieve consistent feature distribution constraints across batches and assembly states. This effectively suppresses tactile noise interference while significantly reducing the dependence on high-density data acquisition and repeated sensor calibration, thereby improving the accuracy and stability of offset sensing during the assembly of male and female connectors of flexible flat cable and the universal interchangeability of the intelligent terminal.

[0058] In one exemplary embodiment, such as Figure 5 As shown, an assembly device for general-purpose male and female connectors of flexible flat cables is provided, comprising: The data acquisition module is used to collect tactile information data in real time; the tactile information data includes multi-electrode time-series tactile signals acquired based on a multi-electrode tactile sensor installed on the end effector.

[0059] The data augmentation module is used to perform electrode-by-electrode decomposition and interpolation processing on the tactile information data based on a Gaussian process regression model to obtain multi-electrode tactile features. The Gaussian process regression model adopts an electrode-by-electrode independent modeling strategy, performs statistical feature ratio analysis based on historical tactile information data, and adaptively configures the kernel function to adapt to the nonlinear response and interpolation of each electrode. The historical tactile information data includes tactile information data with known offsets. The multi-electrode tactile features include the offsets of each electrode after continuous interpolation.

[0060] The principal component analysis and alignment module is used to construct a feature subspace using principal component analysis and to apply a subspace alignment method to constrain the structural consistency of the multi-electrode tactile features and perform linear alignment processing to obtain the aligned feature vector.

[0061] The measurement module is used to determine the Mahalanobis distance between the reference vector and the aligned feature vector based on the Mahalanobis distance measurement model, and to measure the contact offset based on the Mahalanobis distance; the contact offset represents the position perception deviation of the assembly, which is used for subsequent assembly position adjustment; wherein, the Mahalanobis distance measurement model is determined based on the aligned feature vector corresponding to historical tactile information data; the Mahalanobis distance measurement model is used to measure the degree of offset information.

[0062] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 6 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores assembly data applicable to male and female connectors of flexible flat cable. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the assembly method applicable to male and female connectors of flexible flat cable.

[0063] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0064] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0065] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0066] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0067] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0068] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0069] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logic devices, etc., and are not limited to these.

[0070] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0071] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for assembling a universal connector for a flexible flat cable, characterized in that, The assembly method for male and female connectors of flexible flat cables is applied in the assembly scenario of male and female connectors of flexible flat cables; The assembly method for general-purpose male and female connectors of flat cable includes: Real-time acquisition of tactile information data; the tactile information data includes multi-electrode time-series tactile signals acquired based on a multi-electrode tactile sensor installed on the end effector; Based on a Gaussian process regression model, the tactile information data is decomposed and interpolated electrode by electrode to obtain multi-electrode tactile features. The Gaussian process regression model adopts an electrode-by-electrode independent modeling strategy, performs statistical feature ratio analysis based on historical tactile information data, and adaptively configures the kernel function to adapt to the nonlinear response and interpolation of each electrode. The historical tactile information data includes tactile information data with known offsets. The multi-electrode tactile features include the offsets of each electrode after continuous interpolation. The feature subspace is constructed by principal component analysis, and the multi-electrode tactile features are subjected to structural consistency constraints and linear alignment processing by a subspace alignment method to obtain the aligned feature vector. The Mahalanobis distance between the reference vector and the aligned feature vector is determined based on the Mahalanobis distance metric model, and the contact offset is measured based on the Mahalanobis distance. The contact offset represents the position perception deviation of the assembly and is used for subsequent assembly position adjustment. The Mahalanobis distance metric model is determined based on the aligned feature vector corresponding to historical tactile information data. The Mahalanobis distance metric model is used to measure the degree of offset information.

2. The assembly method for general-purpose male and female connectors of flexible flat cable according to claim 1, characterized in that, The configuration strategy for adaptive configuration kernel functions is determined based on electrode characteristic indices; The electrode characteristic indicators include: dynamic range of response value, standard deviation, first-order difference statistical characteristics, and noise ratio; Wherein, if the noise ratio is not greater than a preset threshold, a radial basis function kernel is used; otherwise, a Matérn kernel function is used.

3. The assembly method for general-purpose male and female connectors of flexible flat cable according to claim 1, characterized in that, The proposed electrode-by-electrode independent modeling strategy uses a nonparametric regression method based on Bayesian theory to model the nonlinear relationships of each electrode, thereby modeling the nonlinear response of each electrode as a function related to the offset; the expression of the function is: ； in, For the first The tactile information data of each electrode represents the corresponding nonlinear response value; This is the offset; It is a mean function; For kernel functions; For Gaussian process regression.

4. The assembly method for general-purpose male and female connectors of flexible flat cable according to claim 1, characterized in that, A feature subspace is constructed using principal component analysis, and a subspace alignment method is used to constrain the structural consistency of the multi-electrode tactile features. Linear alignment is then performed to obtain the aligned feature vectors, specifically including: A feature subspace is constructed using principal component analysis. The feature subspace includes a source domain subspace and a target domain subspace. The data set is filtered by principal components based on a preset number of principal components to obtain the source domain subspace and the target domain subspace. The data set is a set composed of any two sets of data from different sources in the multi-electrode tactile features. Different sources represent different collection batches or assembly conditions. The linear alignment matrix is ​​optimized using a subspace alignment method to obtain an optimized linear alignment matrix; the linear alignment matrix is ​​determined based on the feature subspace. The source domain subspace is mapped to the target domain subspace based on the optimized linear alignment matrix to achieve linear alignment and obtain the aligned feature vector.

5. The assembly method for general-purpose male and female connectors of flexible flat cable according to claim 4, characterized in that, The linear alignment matrix is ​​optimized using a subspace alignment method, resulting in an optimized linear alignment matrix, which includes: By minimizing the eigenspace mapping error in the Frobenius norm sense The linear alignment matrix is ​​optimized and solved to obtain a closed-form solution, which is used as the optimized linear alignment matrix. The expression for the optimized linear alignment matrix is: ； in, This is the optimized linear alignment matrix; For the source domain subspace; For the target domain subspace; It is the Frobenius norm; This is a transpose.

6. The assembly method for general-purpose male and female connectors of flexible flat cable according to claim 5, characterized in that, when When the problem is irreversible, the Moore–Penrose pseudo-inverse is used for the solution.

7. An assembly device universally applicable to male and female connectors of flexible flat cables, characterized in that, include: The data acquisition module is used to collect tactile information data in real time; The tactile information data includes multi-electrode time-series tactile signals acquired based on a multi-electrode tactile sensor installed in the end effector; The data augmentation module is used to perform electrode-by-electrode decomposition and interpolation processing on the tactile information data based on a Gaussian process regression model to obtain multi-electrode tactile features. The Gaussian process regression model adopts an electrode-by-electrode independent modeling strategy, performs statistical feature ratio analysis based on historical tactile information data, and adaptively configures the kernel function to adapt to the nonlinear response and interpolation of each electrode; the historical tactile information data includes tactile information data with known offsets; the multi-electrode tactile features include the offsets of each electrode after continuous interpolation. The principal component analysis and alignment module is used to construct a feature subspace using principal component analysis and to apply a subspace alignment method to constrain the structural consistency of the multi-electrode tactile features and perform linear alignment processing to obtain the aligned feature vector. The measurement module is used to determine the Mahalanobis distance between the reference vector and the aligned feature vector based on the Mahalanobis distance measurement model, and to measure the contact offset based on the Mahalanobis distance; the contact offset represents the position perception deviation of the assembly, which is used for subsequent assembly position adjustment; wherein, the Mahalanobis distance measurement model is determined based on the aligned feature vector corresponding to historical tactile information data; the Mahalanobis distance measurement model is used to measure the degree of offset information.

8. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the assembly method for male and female connectors of flexible flat cable as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the assembly method for male and female connectors of flexible flat cable as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the assembly method for male and female connectors of flexible flat cable as described in any one of claims 1-6.

Images