Intelligent super-flexible wire harness sensing system

Abstract

The invention relates to the technical field of body sensing, in particular to a body intelligent super-flexible wire harness sensing system, which comprises a fiber response acquisition module, a state monitoring and correction module, a feature fusion module, a topological field reconstruction module and a control instruction generation module. According to the method, accurate extraction of the fiber response points is achieved through comparison of differential operation and the reference displacement, the bending position and angle change can be recognized in time and corrected in the threshold exceeding state, noise in signal transmission keeps stable after being weakened through a numerical method, and the stability of the fiber response points is improved. Further performing weighted integration on the deformation characteristics in multi-scale fusion to enable touch and deformation information to form continuous mapping under the same perception distribution, and finally positioning an action center through a probability model and outputting a topological mapping image, so that accurate capture and spatial position identification of an external force peak value can be realized in a complex environment, and the accuracy of the external force peak value is improved. And the sensitivity of the flexible wire harness to fine mechanical changes and the integrity of global perception are improved.

Description

Technical Field

[0001] This invention relates to the field of embodied sensing technology, and in particular to an ultra-flexible wire harness sensing system suitable for embodied intelligence. Background Technology

[0002] The field of embodied sensing technology involves combining sensor devices, materials science, and embodied intelligent systems to acquire and analyze environmental information and one's own state. Its core aspects include acquiring external physical signals through flexible sensing units, tactile sensing elements, and deformable conductive media, and realizing the capture of mechanical, morphological, and interactive information by embodied intelligent agents through the coupling of mechanical structures and sensing systems. This field as a whole covers the design of flexible sensing materials, signal acquisition and transmission methods, and the embedded integration of sensing systems and intelligent agent structures.

[0003] Among them, the traditional ultra-flexible wire harness sensing system refers to the sensing of external force by embedding conductive fibers or resistive sensing wires in the flexible wire harness. It uses resistance strain measurement to obtain bending or stretching information, and also uses the charge response of piezoelectric fibers to realize dynamic pressure sensing. Some designs also use the light intensity change caused by fiber bending to monitor deformation. It is a common approach to meet the deformation and tactile sensing needs of flexible wire harnesses in embodied intelligence applications.

[0004] Existing technologies rely on resistance strain gauges or piezoelectric fibers to acquire data. Signal acquisition often suffers from superimposed interference under multi-point conditions, making it difficult to distinguish between actual deformation and noise, leading to accumulated biases in monitoring results. When the deformation range is large or dynamic loads are frequent, the sensing system lacks sufficient correspondence between local displacement and the overall state, failing to accurately represent mechanical distribution characteristics, thus resulting in deficiencies in topology recognition and spatial positioning. For example, in continuous motion environments, errors gradually amplify, affecting the accuracy and reliability of tactile response. Summary of the Invention

[0005] To address the shortcomings of existing technologies that rely on resistance strain gauges or piezoelectric fibers for data acquisition, which often suffer from superimposed interference under multi-point conditions, making it difficult to distinguish between actual deformation and noise and leading to accumulated biases in monitoring results, this invention provides an ultra-flexible wire harness sensing system suitable for embodied intelligence. The technical solution is as follows: On the one hand, an ultra-flexible wire harness sensing system suitable for embodied intelligence is provided, the system comprising: The fiber response acquisition module acquires fiber array deformation data through strain sensors, performs differential calculations on transverse displacement and longitudinal extension, compares it with the reference displacement, marks the fiber coordinates as sensitive response points when the displacement exceeds the reference displacement, obtains the spatial coordinates of the sensitive response points, and transmits them to the condition monitoring and correction module. The status monitoring and correction module calculates the correlation degree of bending position based on the spatial coordinates of the sensitive response point, triggers correction when the real-time bending angle exceeds the safety threshold, and outputs the stable transmission status of the harness based on the least squares method combined with time series smoothing processing for noise reduction, and transmits it to the feature fusion module. The feature fusion module performs multi-scale sensing integration and weighted fusion of deformation features based on the stable transmission state of the harness. It sets the deformation weight coefficient and calculates the tactile response data to generate harness global sensing distribution data, which is then transmitted to the topology field reconstruction module. The topology field reconstruction module constructs a three-dimensional sensing topology mapping of the ultra-flexible wire harness using the full-domain sensing distribution data of the wire harness, identifies the response peak point, uses a Gaussian mixture model to locate the tactile action center, extracts the coordinates and response intensity to form a wire harness sensing topology image, and transmits it to the control command generation module.

[0006] As a further aspect of the present invention, the spatial coordinates of the sensitive response point include fiber array deformation, lateral displacement and longitudinal extension differential values; the stable transmission state of the harness includes bending angle, bending position correlation degree and denoised signal; the harness global sensing distribution data includes multi-scale sensing features, deformation feature weighted fusion and tactile response data; and the harness sensing topology image includes response peak position, Gaussian mixture model localization result and tactile action intensity.

[0007] As a further aspect of the present invention, the fiber response acquisition module includes: The deformation monitoring submodule acquires fiber array deformation data through strain sensors, detects the lateral and longitudinal displacement values ​​of the measuring points, calculates the displacement value of each sensing point in the fiber array based on the acquired displacement data, acquires and records the lateral and longitudinal displacement of a single fiber, and generates a displacement dataset. The differential calculation submodule performs differential calculations on the lateral and longitudinal displacements based on the displacement dataset. It then compares the differential results with the set reference displacement, filters out the intervals where the differential values ​​exceed the reference, and obtains the displacement differential intervals. The reference displacement refers to a displacement threshold set based on material properties, experimental calibration, or experience, which is used as a reference standard to determine whether the displacement difference result of the measuring point exceeds the normal range. The sensitive point marking submodule calls the displacement difference interval to mark the coordinate points whose difference value exceeds the reference displacement, summarizes the coordinate set of the points exceeding the reference displacement, and obtains the spatial coordinates of the sensitive response points.

[0008] As a further aspect of the present invention, the state monitoring and correction module includes: The bending correlation submodule obtains the spatial coordinates of the sensitive response points, establishes a spatial distance matrix between coordinate points based on the coordinates, calculates the angular offset between sensitive response points, classifies and quantifies the relationship between the angular offset and the bending position, and generates the bending position correlation degree. The bending position correlation degree represents the degree of influence of the bending position on the stability and performance of the wire harness signal transmission. The higher the value, the greater the influence. The threshold triggering submodule, based on the bending position correlation, calls real-time bending angle data, compares it with the set safety threshold, filters the bending angle range that exceeds the safety threshold, and then associates and marks the over-limit angle with the corresponding spatial coordinates to obtain the over-limit bending angle interval. The safety threshold represents the critical standard for the maximum permissible bending angle or offset that the harness can withstand at that location. The signal denoising submodule calls the over-limit bending angle range, uses the least squares method to fit the time series and calculate the residual based on the corresponding transmission signal data, smooths the residual series and corrects the signal to obtain the stable transmission state of the harness.

[0009] As a further aspect of the present invention, the feature fusion module includes: The deformation acquisition submodule monitors and acquires sensing data at different scales based on the stable transmission state of the wire harness, extracts deformation feature values ​​at each scale, and performs matching and comparison on the differentiated tactile response data to calculate the scale difference degree. The weight calculation submodule evaluates the weight of deformation features at different scales based on the scale difference degree, calculates the fusion coefficient at each scale through weighted operation by combining tactile response data, and combines them to obtain the fusion coefficient value. The global distribution generation submodule calculates and integrates multi-scale sensing features and tactile response data based on the fusion coefficient values, constructs a sensing distribution matrix of the wire harness across the entire spatial range, merges and corrects the node values ​​in the distribution matrix, and generates global sensing distribution data of the wire harness.

[0010] As a further aspect of the present invention, the scale difference degree refers to a quantitative value of the degree of difference between the deformation feature values ​​extracted under the differentiated scale and the tactile response data. The fusion coefficient value is a weighted fusion value calculated based on the scale difference degree and the corresponding weight.

[0011] As a further aspect of the present invention, the topological field reconstruction module includes: The harness sensing submodule analyzes the signal strength and location coordinates in the harness global sensing distribution data, calculates the signal strength and location coordinates, establishes a three-dimensional mapping and performs topological connections to obtain a topological distribution map. The tactile positioning submodule calls the topology distribution map, uses a Gaussian mixture model to cluster coordinate points and response signals, compares the signal amplitude of the region based on the clustering results and filters peak points, locates the tactile action center and records its spatial coordinates and intensity, and obtains the coordinates of the tactile action center. The perception image construction submodule calls the corresponding intensity value according to the coordinates of the tactile action center, arranges the coordinate distribution and intensity value into a matrix, analyzes and integrates the matrix points and numerical image units, and generates a wire harness perception topology image.

[0012] As a further embodiment of the present invention, the control command generation module creates control commands based on the harness sensing topology image, dynamically adjusts motion trajectory parameters, extracts the three-dimensional coordinates of the target contact position, calculates the bending angle and extension length to be adjusted, and generates a motion control command sequence. The motion control command sequence includes target contact position, bending angle adjustment, and extension / retraction length adjustment.

[0013] As a further aspect of the present invention, the control command generation module includes: The topology parsing submodule obtains node positions and connection paths based on the harness-sensing topology image, classifies nodes according to spatial coordinate relationships, calculates the distance difference between nodes, performs three-dimensional parameter mapping based on coordinate information, and generates node distribution parameter values. The trajectory adjustment submodule calls the node distribution parameter value, compares the target trajectory with the node distribution, calculates and corrects the trajectory offset, adjusts the dynamic position according to the spatial distance, integrates the corrected trajectory sequence, and obtains the trajectory correction amount. The angle calculation submodule extracts the three-dimensional coordinates of the target contact position based on the trajectory correction amount, calculates the bending angle based on the trajectory vector difference, obtains the path point extension length, integrates and adjusts the values, and generates a motion control command sequence.

[0014] As a further aspect of the present invention, the node distribution parameter value is a three-dimensional mapping parameter data generated based on the node positions, connection paths, and spatial coordinate relationships in the topological image; The trajectory correction amount is the trajectory position adjustment data calculated based on the comparison between the target trajectory and the node distribution parameter values.

[0015] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: By comparing differential operations with the reference displacement, the fiber response point is accurately extracted, enabling timely identification of bending position and angle changes and correction under over-threshold conditions. Noise in signal transmission is weakened by numerical methods to maintain stability. Furthermore, deformation features are weighted and integrated in multi-scale fusion, so that tactile and deformation information form a continuous mapping under the same perceptual distribution. Finally, the action center is located through a probabilistic model and a topological mapping image is output. This enables accurate capture and spatial identification of external force peaks in complex environments, improving the sensitivity of flexible harnesses to subtle mechanical changes and the integrity of global perception. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a system schematic diagram of the present invention; Figure 2 This is a schematic diagram of the system framework of the present invention; Figure 3 This is a flowchart of the fiber response acquisition module in this invention; Figure 4 This is a flowchart of the state monitoring and correction module in this invention; Figure 5 This is a flowchart of the feature fusion module in this invention; Figure 6 This is a flowchart of the topological field reconstruction module in this invention; Figure 7 This is a flowchart of the control instruction generation module in this invention. Detailed Implementation

[0018] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0019] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0020] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0021] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0022] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0023] This invention provides an ultra-flexible wire harness sensing system suitable for embodied intelligence, such as... Figures 1-2 The diagram shown illustrates an ultra-flexible wire harness sensing system suitable for embodied intelligence, which includes: The fiber response acquisition module acquires fiber array deformation data through strain sensors, performs differential calculations on transverse displacement and longitudinal extension, compares it with the reference displacement, marks the fiber coordinates as sensitive response points when the displacement exceeds the reference displacement, obtains the spatial coordinates of the sensitive response points, and transmits them to the condition monitoring and correction module. The status monitoring and correction module calculates the correlation of bending positions based on the spatial coordinates of sensitive response points, triggers correction when the real-time bending angle exceeds the safety threshold, and outputs the stable transmission status of the harness based on the least squares method combined with time series smoothing processing for noise reduction, which is then transmitted to the feature fusion module. The feature fusion module performs multi-scale sensing integration and weighted fusion of deformation features based on the stable transmission state of the wire harness. It sets the deformation weight coefficient and calculates the tactile response data to generate wire harness global sensing distribution data, which is then transmitted to the topology field reconstruction module. The topology field reconstruction module constructs a three-dimensional sensing topology mapping of the ultra-flexible wire harness using the full-domain sensing distribution data of the wire harness, identifies the response peak point, uses a Gaussian mixture model to locate the tactile action center, extracts the coordinates and response intensity to form a wire harness sensing topology image, and transmits it to the control command generation module. The control command generation module creates control commands based on the harness sensing topology image, dynamically adjusts motion trajectory parameters, extracts the three-dimensional coordinates of the target contact position, calculates the bending angle and extension length to be adjusted, and generates a motion control command sequence. The spatial coordinates of the sensitive response point include the fiber array deformation, lateral displacement and longitudinal extension differential values; the stable transmission state of the harness includes the bending angle, bending position correlation and denoised signal; the harness global sensing distribution data includes multi-scale sensing features, deformation feature weighted fusion and tactile response data; the harness sensing topology image includes the response peak position, Gaussian mixture model localization result and tactile action intensity; and the motion control command sequence includes the target contact position, bending angle adjustment and extension length adjustment.

[0024] Specifically, such as Figure 2 , 3 As shown, the fiber response acquisition module includes: The deformation monitoring submodule acquires fiber array deformation data through strain sensors, detects the lateral and longitudinal displacement values ​​of the measuring points, calculates the displacement value of each sensing point in the fiber array based on the acquired displacement data, acquires and records the lateral and longitudinal displacement of a single fiber, and generates a displacement dataset. The deformation monitoring submodule deploys multiple strain sensors across a fiber array. By installing several acquisition nodes at fixed intervals (e.g., 2 mm) along the fiber length, it simultaneously collects lateral and longitudinal displacement data. During execution, it sequentially reads the resistance changes output by the sensors and converts each resistance change into a corresponding displacement value (in millimeters) using a preset calibration curve. For example, the initial resistance of monitoring point A is... After being subjected to force, it rises to Combined with calibration coefficients Convertible The displacement is then calculated, and the displacement values ​​in both the horizontal and vertical directions are stored as two-dimensional arrays, with the horizontal data array being... The corresponding real-time displacement value of the measuring point in the lateral direction, and the longitudinal data array. For the corresponding longitudinal displacement value, and for the measuring point, its position coordinate index and the corresponding horizontal and vertical data are combined to form a data record. For example, measuring point No. 3 is located at coordinate... The lateral displacement value is The longitudinal displacement value is The complete record is input into the dataset to form a displacement data matrix containing the measurement points. During the execution process, data reading, unit conversion, and parameter storage need to be completed sequentially for each sensor node, and the acquisition points need to be acquired within the same time window to match the same motion state dataset. For example, when the system is used for mechanical finger joint bending monitoring, if a total of 5 sensor nodes are deployed, the displacement dataset is shown in Table 1.

[0025] Table 1: Fiber Array Displacement Monitoring Data Table As shown in Table 1, the measuring points in the fiber array are transformed and stored to form a structured numerical matrix. The specific values ​​of lateral and longitudinal displacements are derived from the direct calculation of the sensor's resistance change and the preset calibration coefficients. The entire process is executed in the order of "reading the original sensor output → converting it into displacement value → associating the measuring point coordinates → adding it to the dataset". In the acquisition of lateral displacement, when the value is in the range of... The time can be divided into low displacement ranges. Divided into the intermediate displacement range, Divided into high displacement ranges, and similarly in the longitudinal direction, in this embodiment, the lateral displacements of measuring points 2 and 3 are both in the middle range, and the longitudinal displacements are also in the middle range, thus completely recording the deformation data matrix of the robot arm in one grasping action.

[0026] The differential calculation submodule performs differential calculations on the lateral and longitudinal displacements based on the displacement dataset. It then compares the differential results with the set reference displacement, filters out the intervals where the differential values ​​exceed the reference, and obtains the displacement differential intervals. During execution, the differential calculation submodule first calls the lateral displacement data array. With longitudinal displacement data array For each measuring point The difference calculation between the lateral and longitudinal displacements is performed sequentially, that is, the lateral value is subtracted from the longitudinal value and the absolute value is taken as the difference result. The calculation is performed point by point throughout the process. For example, when the lateral displacement of measuring point 1 is... The longitudinal displacement is Then the difference value The lateral displacement of measuring point 2 is The longitudinal displacement is Difference value The difference results form a new array. Then, each difference value is compared with a preset reference displacement. During the comparison process, the reference displacement is first determined. The setting principle can be derived from the elastic limit difference component measured by material property experiments, the average difference value of the original calibration data, or a manually set standard. This embodiment takes experimental calibration as an example, setting the standard by taking the average value of multiple test results of the same type of fiber array under standard load. Here, it is assumed that under standard load... Below, the difference values ​​of the 5 measuring points are as follows: , , , , Then its average value is Therefore, the reference displacement is set to During the comparison, if the current difference value is greater than Then it is determined that the difference at that point exceeds the benchmark, for example when ,result If a value exceeds the benchmark, it is considered to exceed the benchmark; otherwise, it is considered not to exceed the benchmark. The judgment process is strictly based on numerical comparison rules. or Then, this logic is executed sequentially in the entire measurement point array to filter out the measurement point intervals with difference values ​​greater than the benchmark. The measurement point numbers of the start and end points are recorded in the form of intervals. For example, if the difference values ​​of measurement points 2 to 4 are all greater than 0.58, they are recorded as interval [2, 4]. In this process, the continuity of the intervals must be ensured, that is, the difference values ​​between the start and end points must meet the condition of being greater than the benchmark. If a point in the middle does not meet the condition, it must be truncated to form multiple intervals. The final displacement difference interval is directly connected to the subsequent output of this submodule.

[0027] The sensitive point marking submodule calls the displacement difference interval to mark the coordinates of points whose difference values ​​exceed the reference displacement, and summarizes the set of coordinates of the points exceeding the reference displacement to obtain the spatial coordinates of the sensitive response points. During execution, the sensitive point marking submodule first calls the displacement difference interval data. For each difference interval, it sequentially reads the coordinate information of the measurement points within it. This coordinate information comes from the measurement point coordinate index of the displacement dataset and is represented in two-dimensional coordinates. When performing the marking action, for measurement points whose difference value exceeds the reference displacement, their coordinates are extracted and stored in the sensitive response point coordinate set. The specific process is as follows: traverse the measurement point number in each interval. Call the corresponding lateral displacement With longitudinal displacement The numerical value, through matching the index, returns the physical coordinates of the measurement point within the fiber array. Add the coordinates to the set In the process, for example, when the interval [2, 4] is determined to be an interval where the difference exceeds the baseline, the measurement points 2, 3, and 4 are processed in sequence, assuming that their coordinates are respectively... , , Then the set Updated to In the specific implementation of the marker, sequential traversal and direct insertion are used. During the traversal operation, it is first determined whether the current coordinate already exists in the set. If it does not exist, insert it; if it already exists, skip it to avoid duplicate marking. The judgment logic is to compare whether the two components of the coordinates are simultaneously equal. Insertion is performed only if the values ​​are not equal. After the interval processing is completed, the set... The internal storage contains the coordinates of the sensitive response points for differential over-datum displacement during this acquisition and calculation process. The determination of "over-datum" and "not over-datum" strictly follows the numerical comparison standard. Furthermore, the number and distribution of the output sensitive response points are entirely determined by this judgment. The final output result is the set of coordinates of the sensitive response points in space, which can be used for spatial positioning or further analysis and processing in subsequent processes.

[0028] Specifically, such as Figure 2 , 4 As shown, the condition monitoring and correction module includes: The bending correlation submodule obtains the spatial coordinates of the sensitive response points, establishes a spatial distance matrix between the coordinate points, calculates the angular offset between the sensitive response points, classifies and quantifies the relationship between the angular offset and the bending position, and generates the bending position correlation degree. The bending position correlation indicates the degree to which the bending position affects the stability and performance of the signal transmission of the wire harness; the higher the value, the greater the impact. The bending correlation submodule obtains the set of spatial coordinates of sensitive response points. First, it sorts the coordinate points in the set in ascending order according to the measurement point number. Then, it calculates the Euclidean distance between any two coordinate points. The method for calculating this distance is to first find the difference in the x-coordinates of the two points. Difference from the ordinate Then square the two differences separately, add them together, and take the square root to get the final distance value. For example, when At time, point and The difference in the x-coordinates is The difference in the vertical coordinates is The sum of the squares of the two differences is The square root is ,point and The difference in the x-coordinates is The difference in the vertical coordinates is The sum of the squares of the two differences is The square root is This process is repeated to obtain the complete distance matrix. Next, for three adjacent sensitive response points, the angular offset is calculated. The first segment vector formed by two adjacent points is first represented as... Then, the sub-segment vector formed by the last two points is expressed as... Calculate the lengths of the two vector segments separately (by squared the differences in their x and y coordinates, summing the squares and taking the square root). Simultaneously, calculate the inner product (by multiplying the x and y coordinates, then adding the sum of the x and y coordinates). Divide this inner product by the product of the lengths of the two vector segments to obtain a cosine ratio. Then, use the inverse cosine to obtain the angle value. For example, when... , , At that time, the first segment vector is The second segment vector is The length of the first segment is The second segment is approximately [length missing] The inner product is The ratio of the inner product to the length product is approximately The inverse cosine is obtained Then, the measured angular offset is classified according to the preset bending position angle range, for example... to less than This is a slightly curved area. to less than This is the middle bend zone, greater than or equal to For large bends, each angular offset is categorized into its corresponding bend location according to its interval. The frequency of occurrence of each type of location in the sensitive point combination is counted, and this frequency is divided by the total number of sensitive point combinations to obtain the correlation degree of the bend location. For example, when the frequency of occurrence in the large bend area is... The total number of combinations is At that time, correlation for Finally, the corresponding bending position correlation results are generated.

[0029] The threshold triggering submodule, based on the correlation of bending position, calls real-time bending angle data, compares it with the set safety threshold, filters the bending angle range that exceeds the safety threshold, and then associates and marks the over-limit angle with the corresponding spatial coordinates to obtain the over-limit bending angle range. The safety threshold represents the critical standard for the maximum permissible bending angle or offset that the harness can withstand at that location; During execution, the threshold triggering submodule first calls the bending position correlation data, and then calls the set of bending angle data collected in real time. The angle data is obtained by measuring and quantifying the real-time deformation of the wire harness at different locations using sensors, and then setting a safety threshold. Its value needs to be determined comprehensively based on multiple data, including the maximum allowable bending angle of the wire harness material, long-term fatigue test results, and original usage records. For example, after conducting 5000 bending cycles on the same wire harness in an experiment, the minimum angle at which damage occurs due to bending is statistically determined to be... If we take 10% of that as a safety margin, then... During the comparison step, each real-time bending angle is recorded. and When performing direct numerical comparisons, If the angle exceeds the threshold, it is determined that the angle is within the threshold range; otherwise, it is determined that the angle is within the threshold range. The comparison logic strictly follows the binary judgment standard of "if greater than, it exceeds the limit; if less than or equal to, it does not exceed the limit". The angle data of the measurement points are traversed, all bending angle records that exceed the limit are filtered, and the spatial coordinates of the corresponding angles are extracted. During the screening process, each out-of-limit angle is associated with and stored along with its corresponding spatial coordinates. For example, if And the corresponding position is Then it will be recorded as in the result data. After data processing is complete, the continuously distributed out-of-limit angle measurement points will be merged into intervals to form out-of-limit bending angle intervals. An interval is defined as a range where the real-time angles of all measurement points between the start and end coordinates are in an out-of-limit state. For example, when... to The measuring point angles are as follows: And the threshold is At that time, the continuous spatial range formed by these three measuring points is classified into the same over-limit bending angle interval, and the final output interval data is used by the next signal denoising submodule.

[0030] The signal denoising submodule calls the over-limit bending angle range, uses the least squares method to fit the time series and calculate the residual based on the corresponding transmission signal data, smooths the residual series and corrects the signal to obtain the stable transmission state of the harness. During execution, the signal denoising submodule calls the data for the excessive bending angle range and extracts the corresponding transmission signal time series data based on the range. Its corresponding signal amplitude data The data comes from sampling records of the sensor channels embedded in the harness, with amplitude units of [missing information]. In processing, the first step is to perform fitting calculations on the data within each interval. The steps are as follows: first, the time series... With amplitude Data pairs are paired to form a set, and then the fitting parameters of the fitted curve are calculated to make the curve as close as possible to the sampling points, for example, by taking an interval. to The time data is as follows: ; The signal amplitude data is as follows: ; The fitted amplitude sequence is obtained by solving the parameters of the fitted curve. Next, the residual sequence is calculated and a new array is formed. The calculation method is to subtract the corresponding fitted amplitude data from the amplitude data. A residual value > 0 indicates that the measured signal is higher than the fitted curve, and < 0 indicates that the measured signal is lower than the fitted curve. For example, when and At that time, the residual was The specific steps for smoothing the residual sequence are: taking the arithmetic mean of each residual value with its immediate and adjacent residual values ​​and replacing the original value with the result; for example, the first segment of the residual sequence... In the middle, the second term, after smoothing, is... This process iterates through the entire residual array to generate a smoothed residual sequence. Finally, the smoothed residual values ​​are added back to the corresponding positions of the fitted amplitude sequence to obtain the corrected signal amplitude array. This set of corrected values ​​is defined as the stable transmission state data of the harness under the over-limit bending range. An example of a correction result in this embodiment is shown in Table 2.

[0031] Table 2: Corrected Harness Signal Amplitude Table As shown in Table 2, the corrected signal amplitude has been slightly adjusted numerically compared to the original value, reflecting the stable state characteristics after removing signal noise. This output result will be saved for subsequent system analysis.

[0032] Specifically, such as Figure 2 , 5 As shown, the feature fusion module includes: The deformation acquisition submodule monitors and acquires sensing data at different scales based on the stable transmission state of the wire harness, extracts the deformation feature values ​​at each scale, and performs matching and comparison on the differentiated tactile response data to calculate the scale difference degree. The deformation acquisition submodule, based on the stable transmission state data of the wire harness, first extracts a set of sensing data at different spatial or temporal scales for each sensing node under this state, and sorts them in ascending order of scale. For example, the time scale can be sorted by... , , The three intervals can be divided into three spatial scales according to... , , The interval is subdivided, and the signal amplitude and deformation displacement of the sensing node within each scale interval are read. The original values ​​are standardized to ensure dimensional consistency across different scales. Then, the deformation characteristic value is calculated at each scale. The specific calculation process is as follows: based on the mean, maximum, and fluctuation range of the displacement of the sampling points within that scale, a single parameter describing the deformation characteristics of that scale is extracted. For example, when... Displacement data within the scale are When, its mean is The maximum value is The fluctuation range is The three feature values ​​are stored sequentially in the feature dataset corresponding to that scale, and then the haptic response dataset is called. This dataset consists of standardized pressure or force-based values ​​recorded by multiple tactile sensor nodes, expressed in units of... The record shows that during the point-to-point matching process, deformation characteristic values ​​and tactile response values ​​are aligned according to the same node number and timestamp to ensure comparison at the same measurement moment. The matching action involves taking the difference point by point, and then using the absolute value as the matching error value. For example, if a node is at... Deformation characteristic value at scale tactile response value (Converted to equivalent displacement values), then the error is After performing the traversal, obtain the average error of the nodes at that scale. This is called the scale difference degree. For example, the error value of the three nodes is , , hour, For different scales, repeat the above process to finally obtain the scale difference result for each scale, so that it can be called in the subsequent weight calculation submodule.

[0033] The weight calculation submodule evaluates the weight of deformation features at different scales based on the scale difference degree, combines tactile response data, calculates the fusion coefficient at each scale through weighted operation, and combines them to obtain the fusion coefficient value. The weight calculation submodule calls the scale difference result set. Each degree of difference is processed sequentially. To avoid imbalance when differences in different scales are directly superimposed, a reference baseline degree of difference needs to be determined first. This value is obtained by averaging the scale differences. For example, when the scale differences are respectively... , and At that time, the average of the three was This mean is set as the baseline difference. When calculating the weight value for each scale, the ratio of the baseline difference to the difference at that scale is used. The operation is as follows: the baseline difference value is divided by the current scale difference value, and the resulting ratio is the weight result. For example, if the current scale difference is... At that time, the benchmark difference The ratio is approximately This ratio is the weight value for that scale; for a difference of 1... The scale is used to calculate the weights. For differences of ; The scale is used to calculate the weights. This differentiated scale assigns different numerical weights: scales with smaller differences correspond to larger numerical weights, and scales with larger differences correspond to smaller numerical weights. Subsequently, each haptic response data point from the multi-scale dataset is extracted and multiplied by its corresponding weight to generate a new weighted haptic dataset, for example, at the spatial scale. Below, the tactile response data is The weight is Then, by calculating the product one by one, the weighted result is: The set is named the Weighted Haptic Dataset. Then it is compared with the set of deformation eigenvalues ​​corresponding to the same scale. Perform point-by-point addition, for example when At that time, the corresponding cumulative calculation result is To form a fused dataset Finally, the fused datasets obtained at the differentiated scales are aligned one by one according to node number and time series position, and the average value is taken in the vertical direction to obtain the overall fusion coefficient. This value represents the unified result after weighted integration of differentiated scale features and haptic response. For example, the fusion values ​​of a certain node at the three scales are as follows: The average result is This value is defined as the final fusion coefficient and input into the subsequent global distribution generation submodule.

[0034] The global distribution generation submodule calculates and integrates multi-scale sensing features and tactile response data based on the fusion coefficient value, constructs the sensing distribution matrix of the wire harness in the entire space, merges and corrects the node values ​​in the distribution matrix, and generates global sensing distribution data of the wire harness. During execution, the global distribution generation submodule first calls the fusion coefficient values ​​and their corresponding node numbers and spatial coordinates to align the data one-to-one with the multi-scale perceptual feature set and tactile response data set. The alignment process involves sorting the nodes by number from smallest to largest, ensuring that the three datasets for each node are perfectly matched in location and timestamp. Then, a combination operation is performed on each node, that is, the fusion coefficient values ​​are... The sensory feature values ​​and tactile response values ​​of this node at multiple scales are directly superimposed or differentially processed. The superposition is selected according to design requirements to form a comprehensive sensory value. For example, the fusion coefficient of a certain node is... The mean of multi-scale perceptual features is The mean tactile response is The overall perceived value is This value directly reflects the overall sensing intensity of the node in the global distribution. By traversing all nodes in this way, a preliminary sensing distribution matrix is ​​generated with node numbers as rows and spatial locations as columns. The difference between adjacent nodes in the matrix is ​​set to a correction threshold. The elements are merged. The merging action involves replacing the original values ​​of two nodes with the arithmetic average of the perception values ​​of adjacent nodes. For example, when node... With nodes The perceived values ​​are respectively and The difference between the two is and Set as Then the perceived values ​​of both nodes are updated to Then update it in the matrix. After the traversal of nodes completes the correction process, the final wire harness global sensing distribution data matrix is ​​output. Some results of this embodiment are shown in Table 3.

[0035] Table 3: Distribution Matrix of the Whole-Domain Sensing Harness (Partial Nodes) As shown in Table 3, the comprehensive sensing value of each unit in the matrix corresponds to the multi-parameter fusion result of the line bundle at that location. The values ​​after merging and correction have continuity and can be directly used for spatial distribution visualization and data analysis input.

[0036] Specifically, such as Figure 2 , 6 As shown, the topology field reconstruction module includes: The harness sensing submodule analyzes the signal strength and location coordinates in the harness global sensing distribution data, calculates the signal strength and location coordinates, establishes a three-dimensional mapping and performs topological connections to obtain the topological distribution map; During execution, the harness sensing submodule first calls the harness global sensing distribution data matrix. Each row of this matrix records the node number and spatial coordinates. and comprehensive perception value After transmission, the electrical signal amplitude and displacement response values ​​of the target wire bundle at differentiated positions are synchronously acquired by a three-dimensional sensor array deployed within the monitoring area. The acquired results are then compared with... A one-to-one matching process is performed between node numbers and locations to ensure that the original integrated sensing value at the same coordinate point corresponds to the real-time acquired signal. After matching, numerical calculations are performed on the signal strength and spatial coordinates of each matched node. The calculation process involves normalizing the signal strength to... to Between, and , , The three coordinate values ​​maintain the physical quantity units (units) It directly participates in the mapping operation, using signal strength as the third-dimensional color or height value for three-dimensional spatial mapping. For example, when a node... The real-time electrical signal is The maximum value of the signal set is And the minimum value is When, its normalization strength is Then With strength Positioned in a three-dimensional coordinate system All node data are processed sequentially; the mapped coordinate point set is connected according to the node topology, and the connection principle is to sort the nodes in ascending order of spatial distance, only connecting nodes with a distance less than a set connection threshold. Node pairs, for example, let When node to The horizontal distance is If a connection is established, a topological link is created; otherwise, no connection is established. After the connections are completed, the resulting 3D topological distribution map, consisting of coordinate points and edges, is recorded as follows: Its structure includes node positions, intensity values, and connection relationships between adjacent nodes, which are used as inputs for subsequent tactile positioning submodules.

[0037] The tactile localization submodule calls the topology distribution map, uses a Gaussian mixture model to cluster coordinate points and response signals, compares the signal amplitude of the region based on the clustering results and filters the peak points, locates the tactile action center and records its spatial coordinates and intensity, and obtains the coordinates of the tactile action center. During execution, the haptic positioning submodule first calls the topology map. Read the spatial coordinates of each node. and the corresponding signal strength The node data is then arranged into a node feature set according to coordinate order. Then, clustering operations are performed using the spatial distance between nodes and signal strength as dual input conditions. Before clustering, the signal strength of the nodes is first... Perform normalization processing, and set the normalization range to... to For example, the signal strength of a certain node is The maximum and minimum values ​​of the entire set are respectively and Its normalized value is After normalization, the data are grouped according to coordinate distance and intensity difference, with clusters whose three-dimensional Euclidean distance is less than the set cluster radius. The nodes are grouped into one category, for example When, if node and The spatial distance is And the absolute value of the strength difference is less than If the values ​​are the same, they belong to the same group. After grouping the nodes, the signal amplitudes of the nodes within each group are compared, and the node with the largest value is directly identified as the peak point of that group; the comparison process involves iterating through each node in the group. If the strength of the current node is found to be greater than the recorded maximum value, then update the maximum value and the node number. For example, if the strength value of a node in a certain group is... Then the peak node is the intensity The corresponding spatial location. The peak node in each category is recorded as a tactile action center point, and its output information is... For example, the center point obtained is By repeating the process for all clusters, the set of coordinates of the tactile action center can be obtained. This set will serve as the input data for the next perceptual image construction submodule.

[0038] The perception image construction submodule calls the corresponding intensity value according to the coordinates of the tactile action center, arranges the coordinate distribution and intensity value matrix, analyzes and integrates the matrix points and numerical values ​​of the image units, and generates a line harness perception topology image. During execution, the perceptual image construction submodule first calls the set of coordinates of the tactile action center. Each record in the set is represented by the spatial coordinates of the center point. and the corresponding signal strength After the composition and invocation are completed, the center point will be arranged according to... and The signals are arranged in ascending order to form the index order of the two-dimensional coordinate grid. Then, the corresponding signal strength values ​​are directly filled into the matrix cells to generate the initial intensity matrix. The row and column indices of the matrix correspond to and Spatial coordinates of direction. For example, when the coordinate set of the center of tactile sensation is: ; Sorted Direction is , Direction is The generated two-dimensional intensity matrix For: Line 1 , line 2 Before outputting the matrix, numerical analysis is performed on the matrix cells. The analysis involves checking whether the difference between adjacent cells exceeds a set threshold. If the value is greater than the threshold, it is marked as a gradient change region; if it is less than or equal to the threshold, it is considered a continuous region. For example, when the threshold is... hour, and The difference is , i.e., continuous region; and The difference is This refers to the gradient change region, where the labeled information is added to the matrix's additional attribute table. Finally, the two-dimensional coordinates, signal strength values, and gradient labels are integrated together to form the final wire harness sensing topology image data. Some matrix data in this embodiment is shown in Table 4.

[0039] Table 4: Matrix of topological images for harness sensing (partial data) As shown in Table 4, this matrix data fully records the two-dimensional spatial location and intensity distribution information of the tactile action center, and marks the distribution characteristics of the continuous area and the gradient change area. It can be directly used as input data for visually drawing perceptual topology images.

[0040] Specifically, such as Figure 2 , 7 As shown, the control command generation module includes: The topology parsing submodule obtains node positions and connection paths based on the harness-sensing topology image, classifies nodes according to spatial coordinate relationships, calculates the distance difference between nodes, performs three-dimensional parameter mapping based on coordinate information, and generates node distribution parameter values. During execution, the topology parsing submodule first calls the harness-sensing topology image data. Read the spatial coordinates of each node sequentially from this data. And its connection path information, categorizing the nodes according to The coordinate values ​​are sorted in ascending order. When coordinates are the same, Sort the coordinates in ascending order to obtain the sorted node index sequence. Used for subsequent path and distance calculations. Then, when classifying nodes based on spatial coordinate relationships, the classification is first performed on nodes within the connecting path. Coordinate difference as horizontal spacing Coordinate difference as vertical spacing The coordinate difference is used as the height spacing for sequential calculation. For example, for the coordinates of node A... Coordinates of node B The horizontal spacing is Vertical spacing is The height spacing is After obtaining three sets of spacing data, they are added sequentially and compared with the reference distance. Subtract to obtain the distance difference for each connected segment, for example Set as If the sum of the distances in the three directions is The difference is If the sum of the distances in the three directions is The difference is A positive difference indicates that the distance between nodes is greater than the reference distance, while a negative value indicates that it is smaller. Next, the distance difference is combined with the spatial coordinates of the nodes. As a coordinate axis position parameter, the distance difference is bound as an additional parameter label. For example, the mapping parameter value corresponding to node B in the above case is... The coordinates of the other node C The corresponding mapping parameter value is In this way, the mapping parameters of the nodes are sequentially written into the node distribution parameter matrix to form a complete dataset of node distribution parameter values. This dataset will serve as the input data for the previous step of the trajectory adjustment submodule.

[0041] The trajectory adjustment submodule calls the node distribution parameter values, compares the target trajectory with the node distribution, calculates and corrects the trajectory offset, adjusts the dynamic position according to the spatial distance, integrates the corrected trajectory sequence, and obtains the trajectory correction amount. During execution, the trajectory adjustment submodule first calls the node distribution parameter value dataset. Combined with a pre-set sequence of target trajectory points The two are then aligned point-by-point. The matching process involves: for each point on the target trajectory, finding the node with the closest spatial distance to that point in the node distribution parameter value dataset, and recording the distance between them. , , Coordinate difference in direction, such as the target trajectory point With node parameters In comparison, the differences are as follows: , , This indicates that the trajectory point has a connection with the corresponding node in the height direction. The offset. After calculating the difference between the target trajectory point and its nearest neighbor node, the absolute value of the three-direction difference of each trajectory point is judged. The judgment criteria are: if the absolute value is greater than or equal to the position adjustment threshold... If the direction is not corrected, the threshold can be set according to the actual accuracy requirements of the harness movement, for example... . On point With trajectory points When comparing, the difference in the X direction , Y direction All need correction, Z direction It also needs to be corrected. The direction difference to be corrected is directly added to the original target trajectory point coordinates according to its sign to obtain the corrected trajectory point coordinates. For example, the above trajectory point is corrected to... The corrected trajectory points are then rearranged according to the original target trajectory order to form the corrected trajectory sequence. Then take the original trajectory sequence With the corrected trajectory sequence Point-by-point comparison to obtain the set of trajectory correction values. Each record represents a correction value for the direction of a trajectory point, such as point... The correction amount is The final set of trajectory correction values ​​will serve as the input data for the angle calculation submodule.

[0042] The angle calculation submodule extracts the three-dimensional coordinates of the target contact position based on the trajectory correction amount, calculates the bending angle based on the trajectory vector difference, obtains the path point extension length, integrates and adjusts the values, and generates a motion control command sequence. During execution, the angle calculation submodule first calls the trajectory correction set. and the corrected trajectory sequence The three-dimensional coordinate difference between each pair of adjacent trajectory points is extracted as the trajectory vector. For example, trajectory points and The vector difference is The angle between this trajectory vector and its subsequent adjacent vectors is calculated. For example, another vector is... arrive Then the vector difference is To compare the angle between two vectors, the specific steps are to first obtain the lengths of each vector, and then calculate the bending angle corresponding to the angle based on the numerical relationship between the vector components. For example, if the calculated angle is approximately... While calculating the angle, the sum of the displacements in three directions for each trajectory segment is extracted and used as the scaling length of the path points. For example... The sum of the absolute values ​​of the components is Then, the stretching length is adjusted by combining the trajectory correction amount, for example, the correction amount. The corresponding total change is The corrected telescopic length is The bending angle, path point extension / retraction length, and adjustment value of each vector segment are registered in the motion control command table for sequence generation. Some calculation results from this embodiment are shown in Table 5.

[0043] Table 5: Calculation Results of Trajectory Segment Angle and Extension Length As shown in Table 5, this data table records the bending angles and stretching changes of some adjacent path segments in the trajectory, which can be directly used as the basis for generating motion control command sequences.

[0044] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A flexible wire harness sensing system suitable for embodied intelligence, characterized in that, The system includes: The fiber response acquisition module acquires fiber array deformation data through strain sensors, performs differential calculations on transverse displacement and longitudinal extension, compares it with the reference displacement, marks the fiber coordinates as sensitive response points when the displacement exceeds the reference displacement, obtains the spatial coordinates of the sensitive response points, and transmits them to the condition monitoring and correction module. The status monitoring and correction module calculates the correlation degree of bending position based on the spatial coordinates of the sensitive response point, triggers correction when the real-time bending angle exceeds the safety threshold, and outputs the stable transmission status of the harness based on the least squares method combined with time series smoothing processing for noise reduction, and transmits it to the feature fusion module. The feature fusion module performs multi-scale sensing integration and weighted fusion of deformation features based on the stable transmission state of the harness. It sets the deformation weight coefficient and calculates the tactile response data to generate harness global sensing distribution data, which is then transmitted to the topology field reconstruction module. The topology field reconstruction module constructs a three-dimensional sensing topology mapping of the ultra-flexible wire harness using the full-domain sensing distribution data of the wire harness, identifies the response peak point, uses a Gaussian mixture model to locate the tactile action center, extracts the coordinates and response intensity to form a wire harness sensing topology image, and transmits it to the control command generation module.

2. The ultra-flexible wire harness sensing system suitable for embodied intelligence according to claim 1, characterized in that, The spatial coordinates of the sensitive response point include the fiber array deformation, lateral displacement and longitudinal extension differential values; the stable transmission state of the harness includes the bending angle, bending position correlation degree and denoised signal; the harness global sensing distribution data includes multi-scale sensing features, deformation feature weighted fusion and tactile response data; the harness sensing topology image includes the response peak position, Gaussian mixture model localization result and tactile action intensity.

3. The ultra-flexible wire harness sensing system suitable for embodied intelligence according to claim 1, characterized in that, The fiber response acquisition module includes: The deformation monitoring submodule acquires fiber array deformation data through strain sensors, detects the lateral and longitudinal displacement values ​​of the measuring points, calculates the displacement value of each sensing point in the fiber array based on the acquired displacement data, acquires and records the lateral and longitudinal displacement of a single fiber, and generates a displacement dataset. The differential calculation submodule performs differential calculations on the lateral and longitudinal displacements based on the displacement dataset. It then compares the differential results with the set reference displacement, filters out the intervals where the differential values ​​exceed the reference, and obtains the displacement differential intervals. The sensitive point marking submodule calls the displacement difference interval to mark the coordinate points whose difference value exceeds the reference displacement, summarizes the coordinate set of the points exceeding the reference displacement, and obtains the spatial coordinates of the sensitive response points.

4. The ultra-flexible wire harness sensing system suitable for embodied intelligence according to claim 1, characterized in that, The condition monitoring and correction module includes: The bending correlation submodule obtains the spatial coordinates of the sensitive response points, establishes a spatial distance matrix between coordinate points based on the coordinates, calculates the angular offset between sensitive response points, classifies and quantifies the relationship between the angular offset and the bending position, and generates the bending position correlation degree. The threshold triggering submodule, based on the bending position correlation, calls real-time bending angle data, compares it with the set safety threshold, filters the bending angle range that exceeds the safety threshold, and then associates and marks the over-limit angle with the corresponding spatial coordinates to obtain the over-limit bending angle interval. The signal denoising submodule calls the over-limit bending angle range, uses the least squares method to fit the time series and calculate the residuals based on the corresponding transmission signal data, smooths the residual series and corrects the signal to obtain the stable transmission state of the harness.

5. The ultra-flexible wire harness sensing system suitable for embodied intelligence according to claim 1, characterized in that, The feature fusion module includes: The deformation acquisition submodule monitors and acquires sensing data at different scales based on the stable transmission state of the wire harness, extracts deformation feature values ​​at each scale, and performs matching and comparison on the differentiated tactile response data to calculate the scale difference degree. The weight calculation submodule evaluates the weight of deformation features at different scales based on the scale difference degree, calculates the fusion coefficient at each scale through weighted operation by combining tactile response data, and combines them to obtain the fusion coefficient value. The global distribution generation submodule calculates and integrates multi-scale sensing features and tactile response data based on the fusion coefficient values, constructs a sensing distribution matrix of the wire harness across the entire spatial range, merges and corrects the node values ​​in the distribution matrix, and generates global sensing distribution data of the wire harness.

6. The ultra-flexible wire harness sensing system suitable for embodied intelligence according to claim 5, characterized in that, The scale difference refers to the quantitative value of the degree of difference between the deformation feature values ​​extracted under the differentiated scale and the tactile response data; The fusion coefficient value is a weighted fusion value calculated based on the scale difference degree and the corresponding weight.

7. The ultra-flexible wire harness sensing system suitable for embodied intelligence according to claim 1, characterized in that, The topology field reconstruction module includes: The harness sensing submodule analyzes the signal strength and location coordinates in the harness global sensing distribution data, calculates the signal strength and location coordinates, establishes a three-dimensional mapping and performs topological connections to obtain a topological distribution map. The tactile positioning submodule calls the topology distribution map, uses a Gaussian mixture model to cluster coordinate points and response signals, compares the signal amplitude of the region based on the clustering results and filters peak points, locates the tactile action center and records its spatial coordinates and intensity, and obtains the coordinates of the tactile action center. The perception image construction submodule calls the corresponding intensity value according to the coordinates of the tactile action center, arranges the coordinate distribution and intensity value into a matrix, analyzes and integrates the matrix points and numerical image units, and generates a wire harness perception topology image.

8. The ultra-flexible wire harness sensing system suitable for embodied intelligence according to claim 1, characterized in that, The control command generation module creates control commands based on the harness sensing topology image, dynamically adjusts motion trajectory parameters, extracts the three-dimensional coordinates of the target contact position, calculates the bending angle and extension length to be adjusted, and generates a motion control command sequence. The motion control command sequence includes target contact position, bending angle adjustment, and extension / retraction length adjustment.

9. The ultra-flexible wire harness sensing system suitable for embodied intelligence according to claim 8, characterized in that, The control command generation module includes: The topology parsing submodule acquires node positions and connection paths based on the harness-sensing topology image, classifies nodes according to spatial coordinate relationships, calculates the distance difference between nodes, performs three-dimensional parameter mapping based on coordinate information, and generates node distribution parameter values. The trajectory adjustment submodule calls the node distribution parameter value, compares the target trajectory with the node distribution, calculates and corrects the trajectory offset, adjusts the dynamic position according to the spatial distance, integrates the corrected trajectory sequence, and obtains the trajectory correction amount. The angle calculation submodule extracts the three-dimensional coordinates of the target contact position based on the trajectory correction amount, calculates the bending angle based on the trajectory vector difference, obtains the path point extension length, integrates and adjusts the values, and generates a motion control command sequence.

10. The ultra-flexible wire harness sensing system for embodied intelligence according to claim 9, characterized in that, The node distribution parameter values ​​are three-dimensional mapping parameter data generated based on the node positions, connection paths, and spatial coordinate relationships in the topological image; The trajectory correction amount is the trajectory position adjustment data calculated based on the comparison between the target trajectory and the node distribution parameter values.

Images