Assembling system of electromagnetic compatibility shielding shell of piezoelectric buzzer

Abstract

The invention relates to the technical field of electromagnetic compatibility assembly of electronic devices, and discloses an assembly system of an electromagnetic compatibility shielding shell of a piezoelectric buzzer. The system comprises a simulation assembly generation module, an acquisition monitoring module, an anomaly detection and association module, a path correction and optimization module and an execution and report module. According to the system, the actual pose and electromagnetic parameters of a tool head are synchronously collected in the assembling process, and electromagnetic correlation abnormal points having key influences on electromagnetic performance are recognized from mechanical abnormal points through time alignment and condition screening. And generating a correction path based on the points, and dynamically and finely adjusting path coordinates according to real-time deformation feedback during execution to form a final assembly path. According to the system, online correlation diagnosis and closed-loop control of the mechanical assembling process and the electromagnetic performance of a product are achieved, and the system adapts to workpiece deformation through dynamic path compensation, so that the consistency of the assembling quality and the electromagnetic shielding effectiveness is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electromagnetic compatibility assembly of electronic devices, in particular to an assembly system of an electromagnetic compatibility shielding shell of a piezoelectric buzzer. BACKGROUND

[0002] In the production process of precision electronic components such as piezoelectric buzzers, installing an electromagnetic compatibility shielding shell is a key process to ensure stable operation of the device and avoid signal interference. Current mainstream automated assembly solutions generally rely on high-precision machine vision positioning and pre-set motion path programs. By monitoring the spatial pose of the tool head or pressure feedback during assembly, it can be determined whether the assembly action is accurately completed. The core goal of this method is to achieve precise fitting in geometric position, ensuring that the shielding shell is physically installed in place.

[0003] This conventional technical solution has limitations. They can only detect and respond to deviations in mechanical position or force, and cannot perceive the actual impact of the assembly process on the core functional indicator of electromagnetic shielding effectiveness. A seemingly minor assembly flaw in geometric size, such as uneven local contact pressure or slight deformation, can cause changes in the electrical contact characteristics of the shielding shell, leading to electromagnetic leakage or resonance frequency shift, but these key quality information is completely missing during assembly. In addition, the shielding shell or workpiece itself may have weak flexibility, which will produce unpredictable elastic deformation when the tool head is in contact and pressurized. Existing static or pre-programmed path correction methods cannot adapt to this dynamic deformation in real time, resulting in errors between the theoretical contact point and the actual effective contact point, affecting assembly consistency and stability of shielding performance. SUMMARY

[0004] The purpose of the present application is to provide an assembly system for an electromagnetic compatibility shielding shell of a piezoelectric buzzer to solve the problems raised in the background art.

[0005] To achieve the above purpose, the present application provides an assembly system for an electromagnetic compatibility shielding shell of a piezoelectric buzzer, which comprises: An analog assembly generation module for generating an analog assembly operation and a pre-planned motion path of the tool head; the pre-planned motion path includes a sequence of expected contact poses of the tool head; A collection and monitoring module for synchronously collecting a sequence of actual contact poses of the tool head and electromagnetic parameters during assembly; An anomaly detection and correlation module for comparing the sequence of actual contact poses with the sequence of expected contact poses point by point, extracting abnormal contact points whose pose differences exceed the allowed range, time-aligning each abnormal contact point with the synchronously collected electromagnetic parameters, and screening abnormal contact points whose electromagnetic parameter values meet the pre-set sensitive conditions and marking them as electromagnetic correlation abnormal points; a path correction and optimization module configured to reconstruct a corrected contact path of the tool head based on all the electromagnetic correlation abnormal points, the corrected contact path being used to replace a corresponding section in the pre-planned motion path, and to dynamically monitor deformation feedback data of the shield shell assembly during execution of the corrected contact path and to fine-tune and compensate spatial coordinates of the corrected contact path according to the deformation feedback data to generate a final assembly path; an execution and reporting module configured to control the tool head to complete an assembly action along the final assembly path and to generate a path execution report according to an action whole-process record.

[0006] Preferably, the generating the pre-planned motion path of the tool head for the simulated assembly operation comprises: Before the tool executes the assembly operation of the piezoelectric buzzer shield shell, a set of simulated assembly operations is generated, the simulated assembly operations including a series of assembly contact points arranged in a preset logical order, and a pre-planned motion path of an assembly tool is generated based on coordinate data of the assembly contact points, The step of generating the set of simulated assembly operations comprises: setting an initial installation posture of the shield shell as an original point of a reference coordinate system; labeling a plurality of theoretical fitting boundary lines in the reference coordinate system according to geometric characteristics of edges of the shield shell; sampling a plurality of discrete fitting feature points according to a preset density along each of the theoretical fitting boundary lines; based on spatial coordinates of the discrete fitting feature points, reversely deducing a tool head posture required for the assembly tool to contact the discrete fitting feature points; arranging and combining the tool head postures corresponding to all the discrete fitting feature points according to a spatial order of the theoretical fitting boundary lines to form an assembly contact point set; configuring a standard pressure value and a standard contact duration for each point in the assembly contact point set according to an operation attribute of each point; defining the assembly contact point set after the configuration as the simulated assembly operation.

[0007] Preferably, the step of the abnormality detection and correlation module extracting an abnormal contact point with a posture difference beyond an allowable range comprises: extracting a current actual contact posture in a time sequence from the actual contact posture sequence; finding an expected contact posture closest to the current actual contact posture in a time label from the expected contact posture sequence; calculating a spatial Euclidean distance between a position vector of the current actual contact posture and a position vector of the found expected contact posture; calculating a rotation angle deviation between an attitude rotation matrix of the current actual contact posture and an attitude rotation matrix of the found expected contact posture. the spatial Euclidean distance is compared with a preset distance tolerance threshold, and the rotation angle deviation is compared with a preset angle tolerance threshold; when the spatial Euclidean distance exceeds the distance tolerance threshold or the rotation angle deviation exceeds the angle tolerance threshold, the current actual contact pose is recorded as a difference pose point; the difference pose point is mapped back to the corresponding assembly contact point in the simulation assembly operation, and the assembly contact point is marked as an abnormal contact point; the comparison and marking steps are repeated for each pose in the actual contact pose sequence until all poses in the sequence are processed.

[0008] Preferably, the step of screening out abnormal contact points with electromagnetic parameter values meeting the preset sensitivity condition by the anomaly detection and association module comprises: acquiring electromagnetic parameter values time-aligned with the abnormal contact points, the electromagnetic parameter values containing field strength readings of multiple frequency bands; comparing the field strength readings of multiple frequency bands with the respective corresponding reference field strength thresholds; identifying the frequency bands whose field strength readings exceed the corresponding reference field strength thresholds, and marking the frequency bands as over-standard frequency bands; calculating the over-standard proportions of the field strength readings of all over-standard frequency bands relative to the reference field strength thresholds; comparing the over-standard proportions with a preset lower limit of sensitivity proportion; if the over-standard proportion is greater than or equal to the lower limit of sensitivity proportion, it is determined that the electromagnetic parameter values associated with the abnormal contact point meet the preset sensitivity condition; assigning an electromagnetic sensitivity mark to the abnormal contact point meeting the preset sensitivity condition, the electromagnetic sensitivity mark containing the number information of over-standard frequency bands and the comprehensive over-standard proportion information; classifying all abnormal contact points carrying electromagnetic sensitivity marks as electromagnetic association abnormal points.

[0009] Preferably, the step of reconstructing the modified contact path of the tool head by the path correction and optimization module comprises: taking the pre-planned motion path as a basic reference path; locating the spatial position points corresponding to all electromagnetic association abnormal points on the basic reference path; taking each spatial position point as the center, performing path node resampling in a preset spherical neighborhood space; smoothly connecting the new path nodes obtained by resampling with the original path nodes in the basic reference path that are not marked as electromagnetic association abnormal points; processing all connection points through a path smoothing algorithm to ensure that the generated path curve is geometrically continuous and has a gentle change in curvature; The new path sequence obtained after the smooth connection and algorithm processing is defined as a modified contact path; A path modification identification code is assigned to each segment of the modified contact path, which is used to indicate the modification state of the path segment relative to the base reference path.

[0010] Preferably, the path modification and optimization module fine-tunes and compensates the spatial coordinates of the modified contact path according to the deformation feedback data, which includes: The real-time pressure distribution of the shield shell assembly during the contact process is obtained through the force sensor installed on the tool head; The real-time position offset of the shield shell assembly edge during the contact process is obtained through the visual measurement unit; The real-time pressure distribution is compared with the preset ideal pressure distribution model to generate a pressure difference vector field; The real-time position offset is compared with the preset allowable offset boundary to generate a position offset vector; The pressure difference vector field and the position offset vector are input into a path compensation controller; The path compensation controller converts the vector input into adjustment instructions for the spatial coordinates of the modified contact path based on the built-in compensation rules; The adjustment instructions drive the driving mechanism of the tool head, so that the motion trajectory of the tool head dynamically deviates according to the adjustment instructions during the contact process; The dynamic deviation process continues until the deformation feedback data indicates that the pressure distribution and the position offset are restored to within the allowable range; All spatial coordinate adjustment amounts finally adopted to achieve the allowable range are recorded, and the modified contact path is updated based on all the spatial coordinate adjustment amounts to generate a final assembly path.

[0011] Preferably, the path modification and optimization module reconstructs the modified contact path of the tool head based on all the electromagnetic correlation abnormal points, which includes: All electromagnetic correlation abnormal points are sorted according to their occurrence time in the assembly process to generate an ordered abnormal point list; Each electromagnetic correlation abnormal point in the ordered abnormal point list is processed in turn, and the corresponding local path replacement scheme is obtained by querying the preset path avoidance strategy library according to its over-standard frequency band information; The path endpoint after processing the previous electromagnetic correlation abnormal point is used as the path starting point for processing the next electromagnetic correlation abnormal point, and the local path replacement scheme is applied in sequence; When the local path replacement schemes of all electromagnetic correlation abnormal points are applied, the continuity of the overall path is checked; If there is a path breakpoint, a transition path curve that meets the tool kinematic constraints is inserted between the breakpoints; The length of the completed whole path is compared with the original pre-planned motion path, and if the length change exceeds a preset range, the whole path is scaled to return the total length to the allowable range; The adjusted whole path is output as the corrected contact path.

[0012] Preferably, the step of establishing a corresponding association between the abnormal contact point and the specific electromagnetic parameter value by the abnormality detection and association module comprises: A timestamp is created for each marked abnormal contact point; In the continuous time sequence data of the assembly environment electromagnetic parameters, the same time as the abnormal contact point timestamp is located; Extract all electromagnetic parameter data segments within a certain time window before and after the time; The extracted electromagnetic parameter data segments are filtered and denoised to eliminate transient interference pulses; The time domain statistical feature values of the filtered and denoised electromagnetic parameter data segments are calculated; The time domain statistical feature values are matched with the pre-learned electromagnetic interference mode feature library; The matching successful interference mode feature label is attached to the corresponding abnormal contact point, and the corresponding association relationship is established.

[0013] Preferably, the step of comparing the actual contact pose sequence with the expected contact pose sequence point by point by the abnormality detection and association module comprises: A pose comparison matrix is constructed, with the rows of the matrix corresponding to the expected contact pose sequence and the columns of the matrix corresponding to the actual contact pose sequence; The value of each element in the matrix is calculated, which is the composite difference degree between the expected pose of the corresponding row and the actual pose of the corresponding column; An optimal path from the top left corner to the bottom right corner in the matrix is found by using the dynamic time warping algorithm, so that the sum of the element values on the path is minimized; The optimal path indicates the best point-to-point matching relationship between the actual sequence and the expected sequence; According to the matching relationship of the optimal path, the isolated points in the actual sequence that cannot be well matched with any point in the expected sequence are found, which are the difference points to be extracted.

[0014] Preferably, the step of processing all connection points by the path smoothing algorithm to ensure that the generated path curve is geometrically continuous and has a smooth curvature change comprises: The spatial coordinate sequence of all connection points is obtained, including the new path nodes obtained by resampling and the original path nodes that are not marked as electromagnetic association abnormal points; The cubic spline interpolation algorithm is applied to the spatial coordinate sequence to generate a segmented cubic polynomial curve, so that the curve is continuous in position and consistent in tangent direction at the connecting points; The curvature function of each polynomial curve is calculated, and it is checked whether the curvature value exceeds the preset maximum allowable curvature threshold; If the curvature value exceeds the maximum allowable curvature threshold, the control weight of the adjacent connecting point is adjusted, and the curvature change rate is reduced to the smooth range through iterative optimization; The geometric continuity of the entire path curve is verified to ensure that there is no sharp point or breakpoint, and the smoothed path sequence is output.

[0015] Compared with the prior art, the beneficial effects of the present application are: By embedding the electromagnetic parameter monitoring depth into the assembly process and accurately timestamping the actual contact pose of the tool head, the system establishes a real-time mechanical-electrical performance correlation diagnosis mechanism. When the tool head motion deviates, this mechanism can intelligently filter out the key points that simultaneously cause electromagnetic parameter degradation from all mechanical abnormal points. This shifts the focus of quality control from simple geometric alignment to process defects that directly threaten the electromagnetic shielding function, enabling online prediction and proactive intervention of the final product performance, and avoiding ineffective assembly due to physical assembly but unqualified electromagnetic performance.

[0016] After planning a new path based on the filtered key abnormal points, the system does not execute it as a fixed instruction, but introduces real-time dynamic monitoring of the deformation of the shielding shell assembly. Deformation feedback data obtained through strain sensing or macro vision is used to continuously fine-tune and compensate the tool head path coordinates in real time. This dynamic compensation process enables the tool head to "adapt" to the deformation of the workpiece caused by stress, maintaining the desired optimal contact state at all times. It effectively overcomes assembly uncertainties caused by workpiece flexibility, fixture errors, or material batch differences, improving the physical contact quality and consistency of assembly to a new level, providing process assurance for obtaining stable and excellent electromagnetic shielding effect. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 The timing diagram of the piezoelectric buzzer electromagnetic compatibility shielding shell assembly system described in the present application; Figure 2 The flowchart generated for simulating the assembly operation; Figure 3 The flowchart for reconstructing the contact path after correction; Figure 4 The two-axis comparison diagram for the piezoelectric buzzer shielding shell assembly path avoidance strategy; Figure 5 The timing comparison diagram of the piezoelectric buzzer shielding shell assembly path. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 This invention provides an assembly system for an electromagnetic compatibility (EMC) shielding shell of a piezoelectric buzzer. The system includes: a simulation assembly generation module, a data acquisition and monitoring module, an anomaly detection and correlation module, a path correction and optimization module, and an execution and reporting module. The simulation assembly generation module generates a simulated assembly operation and a pre-planned motion path for the tool head, which includes the expected contact pose sequence of the tool head. The data acquisition and monitoring module simultaneously acquires the actual contact pose sequence and electromagnetic parameters of the tool head during the assembly process. The anomaly detection and correlation module compares the actual contact pose sequence with the expected contact pose sequence point by point, extracts abnormal contact points where the pose difference exceeds the allowable range, aligns each abnormal contact point with the synchronously acquired electromagnetic parameters in time, and filters out abnormal contact points whose electromagnetic parameter values ​​meet preset sensitivity conditions, marking them as electromagnetically correlated anomaly points. The path correction and optimization module reconstructs the corrected contact path of the toolhead based on all electromagnetic correlation anomalies. This corrected contact path replaces the corresponding segments in the pre-planned motion path. During the execution of the corrected contact path, the deformation feedback data of the shielding shell assembly is dynamically monitored, and the spatial coordinates of the corrected contact path are fine-tuned and compensated based on the deformation feedback data to generate the final assembly path. The execution and reporting module controls the toolhead to complete the assembly action along the final assembly path and generates a path execution report based on the entire action process record.

[0020] Example 1: See Figure 2Before the tool performs the assembly operation of the piezoelectric buzzer shielding shell, a set of simulation assembly operations is generated, the simulation assembly operation includes a series of assembly contact points arranged in a preset logical order, and a pre-planned motion path of the assembly tool is generated based on coordinate data of the assembly contact points. The initial installation posture of the shielding shell is set as the origin of the reference coordinate system, and a plurality of theoretical fitting boundary lines are marked in the reference coordinate system according to the geometric characteristics of the edges of the shielding shell. A plurality of discrete fitting feature points are sampled according to a preset density along each theoretical fitting boundary line, and the tool head posture required by the assembly tool to contact the discrete fitting feature points is reversely deduced based on the spatial coordinates of the discrete fitting feature points. The tool head postures corresponding to all the discrete fitting feature points are arranged and combined according to the spatial order of the theoretical fitting boundary lines to form a set of assembly contact points, and a standard pressure value and a standard contact time are configured for each point in the set of assembly contact points according to the operation attribute of each point. The set of assembly contact points configured is defined as the simulation assembly operation.

[0021] In specific implementation, the initial step of generating the simulation assembly operation is to set the initial installation posture of the shielding shell as the origin of the reference coordinate system, and the reference coordinate system adopts a three-dimensional rectangular coordinate system, and the origin position corresponds to the center positioning point of the piezoelectric buzzer base in the assembly fixture. In specific implementation, marking the theoretical fitting boundary line needs to be based on the three-dimensional model data of the shielding shell, and the three-dimensional model data of the shielding shell is usually imported from a computer-aided design file, and the theoretical fitting boundary line is a mathematical expression of the geometric edge in the reference coordinate system, which is physically contacted between the shielding shell and the piezoelectric buzzer base in the ideal assembly state. In some embodiments, the equal arc length sampling method is adopted to sample the discrete fitting feature points along the theoretical fitting boundary line, and the preset density parameter determines the number of points sampled per unit length of the theoretical fitting boundary line, and the spatial coordinates of the discrete fitting feature points are directly obtained through analytic geometry calculation. In specific implementation, the tool head posture reversely deduced based on the spatial coordinates of the discrete fitting feature points involves coordinate transformation, and the tool head posture is defined by a position vector and an attitude rotation matrix, the position vector is obtained by translating the coordinate of the discrete fitting feature point by a preset bias vector of the tool head, and the attitude rotation matrix is calculated and generated according to the direction of the normal vector of the theoretical fitting boundary line at the discrete fitting feature point.

[0022] In some embodiments, the permutation and combination of the tool head poses corresponding to all the discrete conforming feature points follow a spatial connection order of the theoretical conforming boundary lines, which is determined according to the closed contour structure of the shielding shell. The formation of the assembly contact point set is an ordered data structure, and each element in the data structure stores a unique tool head pose. In a specific implementation, each point in the assembly contact point set is configured with a standard pressure value and a standard contact duration according to the operating properties of the point, including the type of the theoretical conforming boundary line where the point is located and the relative position of the point on the boundary line. The standard pressure value is a scalar parameter, and the standard contact duration is a time parameter. In a specific implementation, defining a simulation assembly operation is to encapsulate the assembly contact point set configured with the standard pressure value and the standard contact duration into an executable instruction sequence, which can be read and parsed by the control unit of the assembly system. It can be understood that generating the pre-planned motion path of the assembly tool is based on the sequence of tool head poses in the assembly contact point set for trajectory interpolation, and the trajectory interpolation algorithm ensures that the tool head smoothly passes through each specified tool head pose point in the continuous motion process.

[0023] It can be understood that, in order to optimize the uniformity of the distribution of the discrete conforming feature points, adaptive density adjustment is introduced in some embodiments, which dynamically modifies the sampling interval according to the local curvature variation of the theoretical conforming boundary line. In a specific implementation, the local curvature value of the theoretical conforming boundary line is calculated using the differential geometry method, and the system automatically increases the sampling number of the discrete conforming feature points for the boundary line segment with larger curvature. The curvature calculation formula is as follows:

[0024] wherein the character represents the curvature value of the i-th sampling point on the theoretical conforming boundary line, and the characters , , respectively represent the spatial coordinate vectors of the three adjacent discrete conforming feature points on the boundary line, the symbol represents the vector cross multiplication operation, and the symbol represents the calculation of the length of the vector.

[0025] In a specific implementation, a current actual contact pose is extracted from the actual contact pose sequence in chronological order, which contains a three-dimensional position coordinate measured by a six-degree-of-freedom sensor and a spatial orientation represented by a quaternion. In a specific implementation, finding the expected contact pose closest to the current actual contact pose in time label from the expected contact pose sequence is accomplished by a timestamp matching algorithm, which calculates the absolute time difference between the acquisition time of the current actual contact pose and the planning time of each pose in the expected contact pose sequence, and selects the expected contact pose with the smallest absolute time difference as the matching object.

[0026] In a specific implementation, a current actual contact pose is extracted from the actual contact pose sequence in chronological order, which contains a three-dimensional position coordinate measured by a six-degree-of-freedom sensor and a spatial orientation represented by a quaternion. In a specific implementation, finding the expected contact pose closest to the current actual contact pose in time label from the expected contact pose sequence is accomplished by a timestamp matching algorithm, which calculates the absolute time difference between the acquisition time of the current actual contact pose and the planning time of each pose in the expected contact pose sequence, and selects the expected contact pose with the smallest absolute time difference as the matching object.

[0027] In some embodiments, calculating the spatial Euclidean distance involves finding the magnitude of the difference between the position vector of the current actual contact pose and the position vector of the matched expected contact pose, where the position vector is a three-dimensional spatial coordinate. In a specific implementation, calculating the rotation angle deviation involves performing a matrix multiplication operation between the attitude rotation matrix of the current actual contact pose and the attitude rotation matrix of the expected contact pose. The result of the multiplication operation is a new rotation matrix. By calculating the trace of this new rotation matrix and applying the inverse cosine function, the minimum rotation angle between the two poses can be solved. This minimum rotation angle is the rotation angle deviation.

[0028] In some embodiments, comparing the spatial Euclidean distance with a distance tolerance threshold is a numerical judgment process, where the distance tolerance threshold is a preset scalar value. Similarly, comparing the rotation angle deviation with an angle tolerance threshold is also a numerical judgment process, where the angle tolerance threshold is a preset scalar value. When the spatial Euclidean distance exceeds the distance tolerance threshold or the rotation angle deviation exceeds the angle tolerance threshold, the system generates a recording event, packaging and storing the current actual contact pose's three-dimensional position coordinates, quaternion orientation, and corresponding timestamp as a difference pose point.

[0029] It is understandable that mapping differing pose points back to corresponding assembly contact points in the simulated assembly operation requires the use of a time stamp and spatial location association index. Each assembly contact point in the simulated assembly operation has a unique sequence number and a planning timestamp. In specific implementation, marking assembly contact points as abnormal contact points is achieved by modifying the data label attribute of the assembly contact point. The data label attribute is changed from "normal" to "abnormal," and the corresponding unique identifier of the differing pose point is associated and stored. Repeatedly executing the comparison and marking steps for each pose in the actual contact pose sequence is a loop processing flow, which continues until the last pose element of the actual contact pose sequence is processed.

[0030] In practice, the electromagnetic parameter values, time-aligned with the abnormal contact point, are read from a circular data buffer. This buffer continuously stores multi-band field strength readings synchronously acquired by the electromagnetic field probe, achieving millisecond-level time alignment accuracy. The electromagnetic parameter values ​​consist of field strength readings from multiple discrete frequency bands, covering the specific spectral range of electromagnetic interference generated by the piezoelectric buzzer. In practice, the comparison of the field strength readings from multiple frequency bands with their respective corresponding reference field strength thresholds is performed in parallel. Each channel has an independently stored reference field strength threshold, and the comparison operation outputs a Boolean sequence indicating whether each frequency band has exceeded the limit.

[0031] In some embodiments, the identification and marking of the over-standard frequency bands is achieved by traversing the sequence of Boolean values, for the frequency channels with a Boolean value of true, the corresponding frequency band number is added to a temporary list, which is the over-standard frequency band list. In a specific implementation, the over-standard proportion of the field strength reading of all over-standard frequency bands relative to the corresponding reference field strength threshold is calculated using a weighted comprehensive calculation method, and the calculation formula is as follows:

[0032] wherein the character represents the comprehensive over-standard proportion, the character represents the total number of over-standard frequency bands, the character is a frequency band index variable, the character represents the measured field strength reading of the th over-standard frequency band, the character represents the reference field strength threshold corresponding to the th over-standard frequency band, and the character represents the weight coefficient pre-assigned to the th frequency band. After the calculation is completed, the comprehensive over-standard proportion is compared with a globally set sensitivity proportion lower limit.

[0033] Optionally, if the comprehensive over-standard proportion is greater than or equal to the sensitivity proportion lower limit, the system determines that the electromagnetic parameter value associated with the abnormal contact point meets the pre-set sensitivity condition. In a specific implementation, the allocation of the electromagnetic sensitivity mark is to add a field to the data structure of the abnormal contact point, which records the number information of the over-standard frequency bands and the specific value of the calculated comprehensive over-standard proportion . The classification of all electromagnetic-associated abnormal points carrying the electromagnetic sensitivity mark is achieved by a data screening operation, which traverses all the assembly contact points marked as “abnormal” and checks whether they contain the electromagnetic sensitivity mark field. The set of contact points containing the field is output as the electromagnetic-associated abnormal point list.

[0034] Embodiment 3: refer to Figure 3With the pre-planned motion path as the base reference path, the spatial position points corresponding to all the electromagnetic correlation abnormal points are located on the base reference path. Path node resampling is performed in a preset spherical neighborhood space centered on each spatial position point. The newly resampled path nodes are smoothly connected with the original path nodes in the base reference path that are not marked as electromagnetic correlation abnormal points. All the connection points, including the newly resampled path nodes and the original path nodes that are not marked as electromagnetic correlation abnormal points, are processed by a path smoothing algorithm to ensure that the generated path curve is geometrically continuous and has a gentle curvature change. A cubic spline interpolation algorithm is applied to the spatial coordinate sequence to generate a piecewise cubic polynomial curve, ensuring that the curve is continuous in position and consistent in tangent direction at the connection points. The curvature function of each polynomial curve is calculated, and it is checked whether the curvature value exceeds the preset maximum allowable curvature threshold. If the curvature value exceeds the maximum allowable curvature threshold, the control weights of adjacent connection points are adjusted, and the curvature change rate is reduced to a smooth range through iterative optimization. The geometric continuity of the entire path curve is verified to ensure that there are no sharp points or breakpoints, and the smoothed path sequence is output. The new path sequence obtained after smoothing connection and algorithm processing is defined as the modified contact path. Each segment of the modified contact path is assigned a path modification identification code to indicate the modification state of the path segment relative to the base reference path.

[0035] In a specific implementation, the pre-planned motion path is used as the base reference path, and the base reference path is a data sequence composed of a series of ordered three-dimensional spatial coordinate points and corresponding tool head poses. In a specific implementation, the spatial position points corresponding to all electromagnetic correlation abnormal points are located on the base reference path by querying the index. Each electromagnetic correlation abnormal point has an associated assembly contact point sequence number in the simulation assembly operation. The assembly contact point sequence number can be used to retrieve the path point with the same sequence number on the base reference path, and the three-dimensional coordinate of the path point is the spatial position point. In some embodiments, path node resampling is performed in a preset spherical neighborhood space centered on each spatial position point. The radius of the spherical neighborhood space is a configurable parameter. The specific method of path node resampling is to randomly generate or generate a series of new three-dimensional coordinate candidate points in the spherical neighborhood space according to a regular grid. In a specific implementation, the newly resampled path nodes are smoothly connected with the original path nodes in the base reference path that are not marked as electromagnetic correlation abnormal points. The smooth connection operation needs to ensure that the connected path is geometrically continuous. The smooth connection operation needs to insert a transition segment between the new path node and the adjacent original path nodes before and after it.

[0036] It is appreciated that the path smoothing algorithm receives a sequence of spatial coordinates of the connection points as input to ensure that the generated path curve is geometrically continuous and the curvature variation is smooth. In a specific implementation, a sequence of spatial coordinates of all the connection points is obtained, which includes the newly generated path nodes from the resampling operation and the original path nodes that are not marked as electromagnetic anomaly points. All the nodes are arranged in the order of execution of the assembly operation. A cubic spline interpolation algorithm is applied to the sequence of spatial coordinates to generate a piecewise cubic polynomial curve, which ensures that the position of the curve is continuous and the first derivative is continuous at each connection point, thereby achieving consistent tangent direction. A curvature function of each polynomial curve is calculated based on the first derivative and the second derivative of the curve, and it is checked whether the curvature value exceeds a preset maximum allowed curvature threshold. In some embodiments, if the curvature value exceeds the maximum allowed curvature threshold, the control weight of the adjacent connection points is adjusted, and the adjustment of the control weight is realized by an iterative optimization algorithm, the goal of which is to reduce the curvature variation rate of the path curve to within a smooth range.

[0037] In a specific implementation, verifying the geometric continuity of the entire path curve requires checking whether there are points of position discontinuity or points of derivative discontinuity in the curve, and the existence of cusps or breakpoints means that the geometric continuity is destroyed. The smoothed path sequence is output as the input of subsequent processing, and the new path sequence obtained after the smoothing connection and algorithm processing is defined as the modified contact path. It is appreciated that each segment of the modified contact path is assigned a path modification identification code, which is a numerical or character code indicating the modification state of the path segment relative to the base reference path, such as "new", "replace" or "keep". In a specific implementation, whether a path segment is modified is determined by comparing the coordinate difference between its start point and end point and the coordinates of the corresponding segment in the base reference path, and if the coordinate difference exceeds a certain threshold, it is marked as a modification state. The path smoothing algorithm minimizes a comprehensive objective function during optimization, which considers both the total length of the path and the smoothness of the curvature variation, and its expression is:

[0038] wherein the character represents the comprehensive objective value of the path smoothing optimization, the character and the character are preset positive weight coefficients, the character represents the total arc length of the current path curve, the character is a characteristic length for dimensionless normalization, for example, it can be taken as one tenth of the total length of the pre-planned motion path, the character represents the curvature value of the path curve at the arc length parameter , and the symbol represents the curvature the first derivative of the arc length , i.e. the rate of change of curvature, is denoted by . The spatial coordinates of the path nodes are adjusted by numerical optimization methods to minimize the integrated objective value , thus obtaining a final path with moderate length and gentle rate of change of curvature under the constraint of geometric continuity.

[0039] In embodiment 4, the real-time pressure distribution of the shield shell assembly during the contact process is obtained by the force sensor installed on the tool head, and the real-time position offset of the edge of the shield shell assembly during the contact process is obtained by the visual measurement unit. The real-time pressure distribution is compared with the preset ideal pressure distribution model to generate a pressure difference vector field, and the real-time position offset is compared with the preset allowable offset boundary to generate a position offset vector. The pressure difference vector field and the position offset vector are input into a path compensation controller, and the path compensation controller converts the vector input into adjustment instructions for the spatial coordinates of the modified contact path based on the built-in compensation rules. The adjustment instructions drive the driving mechanism of the tool head, so that the motion trajectory of the tool head dynamically deviates according to the adjustment instructions during the contact process. This dynamic deviation process continues until the deformation feedback data indicates that the pressure distribution and the position offset are restored to the allowable range. Record all the spatial coordinate adjustment amounts finally adopted to achieve the allowable range, and update the modified contact path based on all the spatial coordinate adjustment amounts to generate a final assembly path. All electromagnetic correlation abnormal points are sorted according to their appearance time sequence in the assembly process to generate an ordered abnormal point list. Each electromagnetic correlation abnormal point in the ordered abnormal point list is processed in turn, and the corresponding local path replacement scheme is obtained from the preset path avoidance strategy library according to the over-standard frequency band information. The path endpoint after processing the previous electromagnetic correlation abnormal point is taken as the path starting point for processing the next electromagnetic correlation abnormal point. The local path replacement scheme is applied in turn, and the continuity of the overall path is checked after the application of the local path replacement scheme for all electromagnetic correlation abnormal points is completed. If there is a path breakpoint, a transition path curve that meets the tool kinematics constraint is inserted between the breakpoints. The length of the connected overall path is compared with the original pre-planned motion path, and if the length change exceeds the preset range, the overall path is scaled by the same proportion to make its total length return to the allowable range. The adjusted overall path is output as the modified contact path.

[0040] In specific implementation, the real-time pressure distribution of the shield shell assembly during the contacting process is acquired by a force sensor installed on the tool head, which is a multi-array tactile sensor, and the real-time pressure distribution data is transmitted in the form of a two-dimensional matrix, each element of which corresponds to the pressure reading of a sensor unit on the sensor array. In specific implementation, the real-time position offset of the edge of the shield shell assembly during the contacting process is acquired by a visual measurement unit, which includes a high-resolution industrial camera and a set of structured light projectors, and the real-time position offset is a physical displacement value converted from the pixel deviation calculated by comparing the currently captured shield shell edge image with the pre-stored reference edge image and combining the camera calibration parameters. The real-time pressure distribution is compared with a preset ideal pressure distribution model to generate a pressure difference vector field, the ideal pressure distribution model defines the magnitude and direction of the normal pressure that each expected contact point on the surface of the shield shell should bear in the ideal assembly state, and each vector in the pressure difference vector field represents the difference in magnitude and direction between the real-time pressure and the ideal pressure. The real-time position offset is compared with a preset allowable offset boundary to generate a position offset vector, the allowable offset boundary defines the maximum and minimum position tolerance allowed for the edge of the shield shell in X, Y and Z dimensions, and the position offset vector indicates the direction and distance of the current real-time position from the center of the allowable offset boundary.

[0041] In some embodiments, the pressure difference vector field and the position offset vector are input into a path compensation controller, and the compensation rules built in the path compensation controller are a set of pre-defined mapping functions. In specific implementation, the adjustment instructions drive the driving mechanism of the tool head, and the driving mechanism of the tool head includes multi-axis servo motors and precision ball screws, so that the motion trajectory of the tool head is dynamically offset according to the adjustment instructions during the contacting process. The dynamic offset process continues until the deformation feedback data indicates that the pressure distribution and the position offset are both restored to the allowable range, and this judgment is completed by continuously monitoring whether the modulus average of the pressure difference vector field and the modulus of the position offset vector are simultaneously lower than their respective thresholds. All the final adopted spatial coordinate adjustment amounts for achieving the allowable range are recorded, the spatial coordinate adjustment amounts are the tool head pose correction values recorded in time sequence, and the final assembly path is generated by updating the corrected contacting path based on all the spatial coordinate adjustment amounts.

[0042] In a specific implementation, all electromagnetic correlation abnormal points are sorted according to their occurrence time in the assembly process, and the occurrence time is determined according to the timestamp of the electromagnetic correlation abnormal point in the data stream of the monitoring module. Each electromagnetic correlation abnormal point in the ordered abnormal point list is processed in turn, and the corresponding local path replacement scheme is obtained according to the exceeding frequency band information of the electromagnetic correlation abnormal point by querying the preset path avoidance strategy library. The path avoidance strategy library stores the mapping relationship between different exceeding frequency band modes and recommended avoidance actions. The path endpoint after the previous electromagnetic correlation abnormal point is processed is taken as the path starting point for processing the next electromagnetic correlation abnormal point, and the local path replacement scheme is applied in turn. When the local path replacement scheme of all electromagnetic correlation abnormal points is applied, the continuity of the overall path is checked, and the checking method includes verifying whether the endpoint coordinates of adjacent path segments are consistent and whether the path tangent is smoothly transitioned.

[0043] It can be understood that if there is a path breakpoint, a transition path curve conforming to the kinematic constraints of the tool is inserted between the breakpoints, and the transition path curve is usually generated in the form of a polynomial or a Bezier curve. In a specific implementation, the completed overall path is compared with the original pre-planned motion path in terms of length, and the total spatial arc length of the two paths is calculated. If the length change exceeds the preset range, the overall path is scaled to adjust the total length to return to the allowable range. The path compensation controller uses a model prediction-based optimization method when performing coordinate fine-tuning, which minimizes a comprehensive cost function including pressure deviation, position deviation, and energy consumption to calculate the optimal adjustment instruction. The specific implementation is as follows: in each control period, the controller uses the currently collected real-time pressure distribution and position offset as the initial state to predict the pressure deviation and position deviation evolution caused by the tool head movement in the future finite time domain; by using the built-in optimization algorithm, a series of future adjustment instruction sequences are solved, so that the comprehensive cost of the cumulative pressure deviation, position deviation, and energy consumption of the adjustment instruction itself after applying the sequence is minimized; only the first adjustment instruction in the sequence is applied each time, and the prediction and optimization are performed again according to the new deformation feedback data in the next control period, so as to realize rolling optimization and feedback correction, and ensure that the tool head path can dynamically adapt to the deformation of the shield shell assembly. The expression of the cost function is as follows:

[0044] wherein the character represents the total cost function value in the prediction time domain, the character represents the prediction step index, the character represents the total number of steps in the prediction time domain, the character represents the real-time pressure distribution vector predicted at the step, the character represents the ideal pressure distribution vector, and the character represents the position deviation vector predicted at the The predicted position offset vector, character Indicates the first The step requires calculating the path coordinates to adjust the command vector, character. , and These are the weighting coefficients for the pressure deviation term, position deviation term, and control quantity change term, respectively. , and These are the characteristic pressure value, characteristic displacement value, and characteristic control value used for dimensional normalization, with symbols... This represents calculating the squared L2 norm of a vector. Solving for this allows us to minimize the total cost. Minimum future adjustment instruction sequence This is to achieve dynamic fine-tuning compensation.

[0045] Optionally, the generation of local path alternatives considers not only frequency band information but also the spatial location of electromagnetic correlation anomalies. The system selects the best-matching template from a library containing various preset avoidance templates and instantiates it using parameterization. Refer to Table 1, which shows a simplified lookup table for the path avoidance strategy library, illustrating the correspondence between out-of-range frequency band patterns and local path alternative types.

[0046] Table 1: Mapping Table of Exceeding Frequency Band Modes and Local Path Alternatives Over-range frequency band mode description Local path replacement scheme type Main adjustment parameter Only low frequency band (e.g. below 100 kHz) over-range Contact point lifting Lifting height ΔZ, contact pressure reduction ratio High frequency band (e.g. above 1 MHz) over-range Contact point lateral translation Translation direction angle θ, translation distance ΔL Wide frequency band multi-point over-range Change contact posture and slow down Tool head deflection angle α, path segment speed coefficient In some embodiments, the path avoidance strategy library is queried based on the specific combination and intensity level of the out-of-range frequency bands. The system performs similarity matching between the out-of-range frequency band information carried by the electromagnetic correlation anomaly points and the pattern entries in the strategy library. It is understood that when inserting a transition path curve, it is necessary to ensure that the newly inserted curve segment and the original path segments before and after it achieve C1 continuity at the connection point, i.e., position and tangent continuity, to meet the requirements of smooth tool head movement. In specific implementations, when scaling the overall path proportionally, the scaling factor is calculated based on the ratio of the total arc length of the original pre-planned motion path to the total arc length of the current overall path. The scaling operation is applied uniformly to the three-dimensional coordinates of all points on the path.

[0047] See Figure 4This is a two-axis comparison chart of piezoelectric buzzer shielding shell assembly path avoidance strategy, which is used to show the correlation between the number of strategy use and success rate under different over-standard frequency band modes. The left vertical axis is "strategy use times", and the right vertical axis is "strategy success rate". The intuitive correlation between the application frequency and effect of the strategy under different frequency band modes is shown. The strategy use times (18) and success rate (0.92) of the high frequency band over-standard mode are both high, indicating that the path avoidance strategy in this scenario is mature and effective; the success rate of the wide frequency band multi-frequency point over-standard mode is only 0.10, reflecting the insufficient adaptability of the strategy in the multi-frequency band interference scenario, which needs to be optimized; the success rate is 1.00 when there is no frequency band over-standard, verifying the stability of the assembly path in the normal scenario. The chart corresponds to the "path avoidance stage", and its value lies in locating the short board of the strategy under different electromagnetic interference scenarios, providing a basis for strategy library iteration; it is a key evaluation tool for industrial assembly path optimization, helping technicians to match high-success-rate strategies and improve assembly efficiency and electromagnetic compatibility.

[0048] In embodiment 5, a timestamp is created for each labeled abnormal contact point, and the same time as the abnormal contact point timestamp is located in the continuous time sequence data of the electromagnetic parameters of the assembly environment. All electromagnetic parameter data segments within a certain time window before and after the time are extracted, and the extracted electromagnetic parameter data segments are filtered and denoised to eliminate transient interference pulses. The time domain statistical feature values of the filtered and denoised electromagnetic parameter data segments are calculated, and the time domain statistical feature values are matched with the pre-learned electromagnetic interference mode feature library. The matched interference mode feature label is attached to the corresponding abnormal contact point to complete the establishment of the corresponding association relationship. A pose comparison matrix is constructed, the rows of the matrix correspond to the expected contact pose sequence, and the columns of the matrix correspond to the actual contact pose sequence. The value of each element in the matrix is calculated, which is the composite difference degree between the corresponding row of expected pose and the corresponding column of actual pose. The dynamic time warping algorithm is used to find an optimal path from the top left corner to the bottom right corner in the matrix, so that the sum of the element values on the path is minimized. The optimal path indicates the best point-to-point matching relationship between the actual sequence and the expected sequence. According to the matching relationship of the optimal path, the isolated points in the actual sequence that fail to match well with any point in the expected sequence are found, which are the difference points that need to be extracted.

[0049] In a specific implementation, a timestamp is created for each labeled abnormal contact point, with a precision of microseconds and synchronized with the global clock of the data acquisition system. In a specific implementation, the same time point as the abnormal contact point timestamp is located in the continuous time sequence data of the assembly environment electromagnetic parameters, which is stored in a circular database with time as the primary key. The system retrieves the exact same time point in the database according to the timestamp of the abnormal contact point "1.356 seconds after the assembly cycle begins". Extract all electromagnetic parameter data segments within a certain time window before and after the time point, and the exemplary time window width can be set to 100 milliseconds, which means that all electromagnetic parameter sampling data from "1.306 seconds after the assembly cycle begins" to "1.406 seconds after the assembly cycle begins" is extracted.

[0050] In some embodiments, the extracted electromagnetic parameter data segments are filtered and denoised to eliminate transient interference pulses, using a digital band-stop filter combined with a median filter algorithm. The digital band-stop filter is used to suppress the known 50Hz power frequency and its harmonic interference, and the median filter algorithm is used to filter out narrow pulse noise caused by random discharge. In a specific implementation, the time domain statistical feature values of the electromagnetic parameter data segments after filtering and denoising are calculated, including the root mean square value, peak-to-peak value and waveform factor of the data segment within a 100 millisecond window. The time domain statistical feature values are matched with the pre-learned electromagnetic interference mode feature library, which stores the time domain statistical feature vector templates of multiple typical electromagnetic interference events.

[0051] It can be understood that the matching successful interference mode feature label is attached to the corresponding abnormal contact point to complete the establishment of the corresponding association relationship. If the features of the current data segment match template A successfully, the string label "high frequency switch noise" is written into the data record of the abnormal contact point. A pose comparison matrix is constructed, with rows corresponding to the expected contact pose sequence and columns corresponding to the actual contact pose sequence. For an exemplary assembly segment containing 50 poses, the pose comparison matrix is a 50x50 square matrix. In a specific implementation, the value of each element in the matrix is calculated, which is the composite difference degree between the expected pose of the corresponding row and the actual pose of the corresponding column. A dynamic time warping algorithm is used to find an optimal path from the top left corner to the bottom right corner in the matrix, so that the sum of the element values on the path is minimized. The dynamic time warping algorithm allows the actual sequence and the expected sequence to be nonlinearly stretched or compressed on the time axis to find the best matching relationship.

[0052] In some embodiments, the optimal path indicates the best point-to-point matching relationship between the actual sequence and the expected sequence, and the optimal path is composed of a series of coordinate index pairs of matrix cells, for example, the path points include (expected point 1, actual point 1), (expected point 2, actual point 2), (expected point 3, actual point 4), and the like, where (expected point 3, actual point 4) indicates that the fourth point of the actual sequence matches the third point of the expected sequence. In a specific implementation, according to the matching relationship of the optimal path, an isolated point in the actual sequence that fails to be well matched with any point of the expected sequence is found, and the judgment method is to check whether each index of the actual sequence appears in the second item of any coordinate pair of the optimal path, if a certain actual sequence index does not appear, the actual contact pose corresponding to the index is determined as an isolated point. Composite difference The calculation formula of the composite difference is designed as:

[0053] Wherein, the character represents the composite difference between the th pose in the expected contact pose sequence and the th pose in the actual contact pose sequence. The character represents the three-dimensional position vector of the th expected pose, and the character represents the three-dimensional position vector of the th actual pose, and the symbol represents the calculation of the spatial Euclidean distance between the two position vectors. The character is a characteristic length for distance normalization. The character represents the attitude rotation angle deviation between the th expected pose and the th actual pose, and the value range is between 0 and π radians. The character is a weight coefficient between 0 and 1, used to adjust the relative importance of the position difference and the attitude difference in the composite difference calculation.

[0054] Optionally, when the dynamic time warping algorithm is applied, additional constraint conditions can be set for path search, such as limiting the local slope of the path to be no more than 2:1, to avoid excessive distortion and time alignment relationship that does not conform to the physical actuality. It can be understood that the electromagnetic interference pattern feature library is pre-constructed by clustering analysis and feature extraction on a large amount of electromagnetic interference event data collected in the historical assembly process.

[0055] Referring to Figure 5This is a time sequence comparison chart of the shielding shell assembly path of the piezoelectric buzzer, which shows the difference between the pre-planned path and the corrected path in the abnormal area. Both charts show the correlation of "pre-planned path → abnormal area → corrected path": in the abnormal area (0.75-1.25 seconds), the offset of the corrected path fluctuates greatly (up to 0.8 mm), reflecting the response of the path correction strategy to the anomaly. The fluctuation amplitude of the corrected path in the second chart is slightly smaller, indicating that the path smoothing optimization effect is more obvious, and the stability of the tool head movement has been improved. The chart corresponds to the "path correction and optimization stage", and its value lies in intuitively presenting the deviation between the pre-planned path and the actual corrected path, verifying the effectiveness of the path adjustment in the abnormal area; it is a key visualization tool for industrial assembly path dynamic optimization, helping technicians evaluate the smoothness and stability of path correction, and improving assembly accuracy and electromagnetic compatibility.

[0056] It should be noted that the relational terms herein, such as first and second, are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any such actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0057] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and changes can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. An assembly system for an electromagnetic compatibility shielding shell for a piezoelectric buzzer, characterized in that, The system includes: The simulation assembly generation module is used to generate simulated assembly operations and pre-planned motion paths for the tool head; the pre-planned motion paths include the expected contact pose sequence of the tool head; The acquisition and monitoring module is used to synchronously acquire the actual contact pose sequence and electromagnetic parameters of the tool head during the assembly process; The anomaly detection and correlation module is used to compare the actual contact pose sequence with the expected contact pose sequence point by point, extract the abnormal contact points whose pose differences exceed the allowable range, align each abnormal contact point with the synchronously acquired electromagnetic parameters in time, filter out the abnormal contact points whose electromagnetic parameter values ​​meet the preset sensitivity conditions, and mark them as electromagnetic correlation anomaly points. The path correction and optimization module is used to reconstruct the corrected contact path of the tool head based on all electromagnetic correlation anomalies. The corrected contact path is used to replace the corresponding segment in the pre-planned motion path. During the execution of the corrected contact path, the deformation feedback data of the shield shell assembly is dynamically monitored, and the spatial coordinates of the corrected contact path are finely adjusted and compensated according to the deformation feedback data to generate the final assembly path. The execution and reporting module is used to control the tool head to complete the assembly action along the final assembly path and generate a path execution report based on the entire action process record.

2. The assembly system for the piezoelectric buzzer electromagnetic compatibility shielding shell according to claim 1, characterized in that, The generation of the simulated assembly operation and the pre-planned motion path of the tool head includes: Before the tool performs the assembly operation of the piezoelectric buzzer shield, a set of simulated assembly operations is generated. The simulated assembly operations include a series of assembly contact points arranged in a preset logical order, and based on the coordinate data of the assembly contact points, a pre-planned motion path for the assembly tool is generated. The steps for generating a set of simulated assembly operations include: The initial installation orientation of the shielding shell is set as the origin of the reference coordinate system; Based on the geometric features of each edge of the shielding shell, multiple theoretical fitting boundary lines are marked in the reference coordinate system; Along each theoretical fitting boundary line, multiple discrete fitting feature points are obtained by sampling according to a preset density. Based on the spatial coordinates of discrete fitting feature points, the tool head pose required for the assembly tool to contact the discrete fitting feature points is deduced in reverse. Arrange and combine the tool head poses corresponding to all discrete fitting feature points according to the spatial order of the theoretical fitting boundary lines to form a set of assembly contact points; Based on the operational attributes of each point in the assembly contact point set, configure a standard pressure value and a standard contact duration for each point; Define the configured set of assembly contact points as the simulated assembly operation.

3. The assembly system for the piezoelectric buzzer electromagnetic compatibility shielding shell according to claim 1, characterized in that, The step of the anomaly detection and correlation module to extract abnormal contact points whose pose differences exceed the allowable range includes: Extract a current actual contact pose from the actual contact pose sequence in chronological order; Find the expected contact pose that is closest to the current actual contact pose time label from the expected contact pose sequence; Calculate the spatial Euclidean distance between the position vector of the current actual contact pose and the position vector of the found expected contact pose; Calculate the rotation angle deviation between the current actual contact pose attitude rotation matrix and the found expected contact pose attitude rotation matrix; The spatial Euclidean distance is compared with a preset distance tolerance threshold, and the rotation angle deviation is compared with a preset angle tolerance threshold. When the spatial Euclidean distance exceeds the distance tolerance threshold or the rotation angle deviation exceeds the angle tolerance threshold, the current actual contact pose is recorded as a difference pose point. Map the different pose points back to the corresponding assembly contact points in the simulated assembly operation, and mark the assembly contact points as abnormal contact points; Repeat the comparison and labeling steps for each pose in the actual contact pose sequence until all poses in the sequence have been processed.

4. The assembly system for the piezoelectric buzzer electromagnetic compatibility shielding shell according to claim 1, characterized in that, The step of the anomaly detection and correlation module to filter out abnormal contact points whose electromagnetic parameter values ​​meet preset sensitivity conditions includes: Obtain electromagnetic parameter values ​​that are time-aligned with the abnormal contact point, the electromagnetic parameter values ​​including field strength readings for multiple frequency bands; The field strength readings of multiple frequency bands are compared with their respective reference field strength thresholds. Identify frequency bands whose field strength readings exceed their corresponding reference field strength thresholds, and mark the frequency bands as exceeding the standard frequency bands; Calculate the percentage of field strength readings exceeding the standard for all frequency bands relative to their reference field strength threshold; The percentage exceeding the standard is compared with the preset lower limit of the sensitivity ratio; If the excess ratio is greater than or equal to the lower limit of the sensitivity ratio, then the electromagnetic parameter value associated with the abnormal contact point is determined to meet the preset sensitivity condition. An electromagnetic sensitivity marker is assigned to an abnormal contact point that meets the preset sensitivity conditions. The electromagnetic sensitivity marker contains information on the number of frequency bands exceeding the standard and information on the overall exceeding ratio. All abnormal contact points carrying electromagnetic sensitivity markers are classified as electromagnetic correlation anomalies.

5. The assembly system for the electromagnetic compatibility shielding shell of the piezoelectric buzzer according to claim 1, characterized in that, The steps for reconstructing the corrected contact path in the path correction and optimization module's toolhead include: Use the pre-planned motion path as the basic reference path; Locate the spatial positions of all electromagnetic correlation anomalies along the basic reference path; For each spatial location point, path nodes are resampled within a pre-defined spherical neighborhood space; The new path nodes obtained by resampling are smoothly connected to the original path nodes in the basic reference path that were not marked as electromagnetic correlation anomalies. All connection points are processed by a path smoothing algorithm to ensure that the generated path curve is geometrically continuous and has a gentle curvature change. The new path sequence obtained after smooth connection and algorithm processing is defined as the corrected contact path; Each segment in the corrected contact path is assigned a path correction identifier code to indicate the modification status of the path segment relative to the base reference path.

6. The assembly system for the piezoelectric buzzer electromagnetic compatibility shielding shell according to claim 1, characterized in that, The path correction and optimization module performs fine-tuning compensation on the spatial coordinates of the corrected contact path based on deformation feedback data, including: The real-time pressure distribution of the shielding shell assembly during the contact process is obtained by a force sensor installed on the tool head; The real-time positional offset of the shielding shell assembly edge during the contact process is obtained through a vision measurement unit; The real-time pressure distribution is compared with a preset ideal pressure distribution model to generate a pressure difference vector field. The real-time position offset is compared with the preset allowable offset boundary to generate a position offset vector; Input the pressure difference vector field and the position offset vector into a path compensation controller; The path compensation controller, based on built-in compensation rules, converts vector input into adjustment instructions for the spatial coordinates of the corrected contact path; Adjust the drive mechanism of the tool head to make the movement trajectory of the tool head dynamically offset according to the adjustment command during the contact process; The dynamic offset process continues until the deformation feedback data shows that the pressure distribution and position offset have both returned to within the allowable range. Record all spatial coordinate adjustments ultimately adopted to achieve the allowable range, and update the corrected contact path based on all spatial coordinate adjustments to generate the final assembly path.

7. The assembly system for the electromagnetic compatibility shielding shell of the piezoelectric buzzer according to claim 1, characterized in that, The path correction and optimization module reconstructs the corrected contact path of the toolhead based on all electromagnetic correlation anomalies, including the following steps: All electromagnetic correlation anomalies are sorted according to their order of appearance during the assembly process to generate an ordered list of anomalies. Each electromagnetic correlation anomaly in the ordered list of anomalies is processed sequentially, and the corresponding local path avoidance strategy is obtained by querying the preset path avoidance strategy library based on its out-of-range frequency band information. The path endpoint after the previous electromagnetic correlation anomaly is processed is used as the path starting point for the next electromagnetic correlation anomaly, and local path replacement schemes are applied sequentially. After all local path alternatives for electromagnetic correlation anomalies have been applied, check the continuity of the overall path. If path breaks exist, a transition path curve that conforms to tool kinematic constraints is inserted between the breaks. The length of the completed path is compared with the original pre-planned motion path. If the length change exceeds the preset range, the overall path is scaled and adjusted proportionally to bring its total length back to the allowable range. The adjusted overall path is output as the corrected contact path.

8. The assembly system for the electromagnetic compatibility shielding shell of the piezoelectric buzzer according to claim 1, characterized in that, The step of establishing a correspondence between abnormal contact points and specific electromagnetic parameter values ​​in the anomaly detection and association module includes: Create a timestamp for each marked abnormal contact point; In the continuous time series data of the electromagnetic parameters of the assembly environment, the time stamp that is the same as the timestamp of the abnormal contact point is located. Extract all electromagnetic parameter data segments within a specific time window before and after the stated time. The extracted electromagnetic parameter data segments are filtered and denoised to eliminate transient interference pulses; Calculate the time-domain statistical characteristics of the electromagnetic parameter data segment after filtering and noise reduction; Match time-domain statistical feature values ​​with a pre-learned electromagnetic interference pattern feature library; The successfully matched interference pattern feature labels are attached to the corresponding abnormal contact points to complete the establishment of the corresponding association.

9. The assembly system for the electromagnetic compatibility shielding shell of the piezoelectric buzzer according to claim 1, characterized in that, The anomaly detection and correlation module compares the actual contact pose sequence with the expected contact pose sequence point by point, including: Construct a pose comparison matrix, where the rows of the matrix correspond to the expected contact pose sequence and the columns of the matrix correspond to the actual contact pose sequence. Calculate the value of each element in the matrix, where the value is the composite difference between the expected pose of the corresponding row and the actual pose of the corresponding column; The dynamic time warping algorithm is used to find an optimal path from the top left corner to the bottom right corner in the matrix, such that the sum of the values ​​of the elements along the path is minimized. The optimal path indicates the best point-to-point matching relationship between the actual sequence and the expected sequence; Based on the matching relationship of the optimal path, identify isolated points in the actual sequence that fail to match any point in the expected sequence well. These isolated points are the difference points that need to be extracted.

10. The assembly system for the electromagnetic compatibility shielding shell of the piezoelectric buzzer according to claim 5, characterized in that, The step of processing all connection points using a path smoothing algorithm to ensure that the generated path curve is geometrically continuous and has a gentle curvature change includes: Obtain the spatial coordinate sequence of all connection points, including newly sampled path nodes and original path nodes that were not marked as electromagnetic correlation anomalies; Apply cubic spline interpolation algorithm to spatial coordinate sequence to generate piecewise cubic polynomial curves, ensuring that the curves are continuous in position and have consistent tangent direction at connection points; Calculate the curvature function of each polynomial curve and check whether the curvature value exceeds the preset maximum allowable curvature threshold; If the curvature value exceeds the maximum allowable curvature threshold, the control weights of adjacent connection points are adjusted, and the curvature change rate is reduced to a smooth range through iterative optimization. Verify the geometric continuity of the entire path curve, ensuring there are no cusps or breaks, and output the smoothed path sequence.

Images