Urinary surgery simulation training method and system based on tactile feedback

Abstract

The invention belongs to the field of medical teaching training, and discloses a urinary surgery simulation training method and system based on tactile feedback, and the method comprises the steps: collecting operation behavior data and force change data, and obtaining a preliminary perception sequence; obtaining a classified organization state cluster according to the preliminary sensing sequence; extracting a hand feeling perception feature from the classified organization state cluster to obtain an update cluster; aiming at the update cluster and the operation behavior data, constructing a support vector machine model to train a tactile feedback response to obtain a dynamic matching function; processing the tactile feedback response data through a dynamic matching function to obtain an optimized feedback sequence; acquiring tissue state update from the optimized feedback sequence, and performing real-time calibration on hand feeling perception by adopting an information processing link to obtain a calibrated perception model; and according to the calibrated sensing model and the force change data, generating a corresponding operation processing track. According to the method, high authenticity and low delay of tactile feedback are realized, and an accurate dynamic matching result is provided for a complex operation scene.

Description

Technical Field

[0001] This invention relates to the field of medical teaching and training technology, specifically to a method and system for urological surgical simulation training based on tactile feedback. Background Technology

[0002] In the field of medical education and training, surgical simulation training is an irreplaceable tool for improving doctors' skills. This area is directly related to medical quality and patient safety, and is a key pillar of medical progress. By simulating real surgical scenarios, doctors can practice repeatedly in a risk-free environment and master complex procedures. However, despite its crucial importance, current technologies and methods still have many shortcomings and urgently need breakthroughs.

[0003] Existing surgical simulation training methods rely heavily on visual and auditory feedback, neglecting the simulation of touch, a core sensory experience. This neglect results in a lack of perception of the real feel during surgery during training, making it difficult for doctors to experience the subtle differences in actual operations through simulation training. Especially in medical scenarios requiring highly precise operations, this deficiency significantly reduces the training effectiveness and fails to meet actual clinical needs.

[0004] More importantly, the lack of tactile feedback presents two closely related technical challenges. First, there's the issue of the realism of tactile perception. Touch is a core tool for judging tissue condition and manipulative force during surgery, and its simulation needs to highly replicate the differences in tactile sensation in real-world scenarios, such as the varying degrees of softness or resistance of different tissues when touched by surgical instruments. If this problem cannot be solved, it leads to the second challenge: the dynamic matching of tactile feedback and surgical behavior. In other words, when doctors apply different forces or change the angle of manipulation during simulation training, tactile feedback cannot respond in real time and accurately, resulting in a disconnect between the training action and perception. For example, when simulating a complex surgical procedure, doctors may lack realistic tactile sensation and be unable to accurately determine whether instruments have touched critical tissues, nor can they perceive subtle changes in manipulative force. This misalignment between perception and behavior severely hinders skill development.

[0005] Therefore, how to realistically reproduce tactile feedback in simulated training and ensure the dynamic matching of tactile perception with operational behavior has become a key issue in improving training effectiveness. Solving this problem not only concerns the improvement of doctors' skills but also directly impacts the future direction of innovation in medical training technology. Summary of the Invention

[0006] To address the above technical problems, this invention provides a urological surgery simulation training method based on tactile feedback, comprising the following steps:

[0007] By collecting operational behavior data and force change data through sensors, a preliminary perception sequence is obtained;

[0008] Based on the preliminary sensing sequence, K-means clustering was used to group the force changes and resistance differences to obtain the classified tissue state clusters.

[0009] Tactile perception features are extracted from the classified tissue state clusters. If the tactile perception features match a first preset threshold, the true restoration level is determined; otherwise, the clustering parameters are adjusted to obtain an updated cluster.

[0010] For the updated clusters and operation behavior data, a support vector machine model is constructed to train the tactile feedback response, and a dynamic matching function is obtained;

[0011] The tactile feedback response data is processed by the dynamic matching function. If the response delay is determined to be lower than a set value, the resistance difference is fused to obtain an optimized feedback sequence; otherwise, the support vector machine model is iterated.

[0012] The organization state update is obtained from the optimized feedback sequence, and the sensory perception is calibrated in real time using the information processing step to obtain the calibrated sensory model.

[0013] Based on the calibrated perception model and force change data, a tactile feedback output is generated. If the output matches the operation behavior, the final dynamic matching result is determined. Based on the final dynamic matching result, it is archived to a preset database using a storage unit, and the corresponding surgical procedure trajectory is generated through a log recording system.

[0014] Preferably, the method for obtaining the classified tissue state clusters includes:

[0015] The original records of the force change and the resistance difference are obtained from the initial sensing sequence, and outliers are removed using a pre-established data cleaning process to obtain the processed basic dataset.

[0016] For the processed basic dataset, the K-means clustering method is used to group the force changes and resistance differences to obtain preliminary classification clusters;

[0017] Based on the preliminary classification of clusters, the variation characteristics and difference identification results within each cluster are analyzed to determine the organizational state category corresponding to the cluster;

[0018] If the classification of the organizational state categories does not match the preset threshold range, the classification clusters are optimized by readjusting the clustering parameters to obtain the adjusted state clusters.

[0019] Based on the adjusted state clusters, extract the state division features within each cluster, determine their matching degree with the preliminary sensing sequence, and obtain the final organizational state distribution;

[0020] Based on the final organizational state distribution, a corresponding state classification mapping table is generated, and the classification cluster category to which each sensing data point belongs is determined based on the mapping table, thus obtaining the classified organizational state cluster.

[0021] Preferably, the method for obtaining the updated cluster includes:

[0022] An initial feature set is obtained by extracting tactile perception feature vectors from the classified tissue state clusters;

[0023] The matching deviation is obtained by calculating the distance difference between the feature vector and the first preset threshold based on the initial feature set;

[0024] If the matching deviation is less than the first preset threshold, the corresponding true restoration level is output to obtain the final level label; if the matching deviation is greater than the first preset threshold, a clustering parameter modification instruction is triggered to obtain the adjusted parameter group.

[0025] The first-generation updated clusters are obtained by performing a K-clustering algorithm on the original tissue state data using the adjusted parameter set.

[0026] The second feature set is obtained by re-extracting the tactile perception feature vector from the first generation update cluster;

[0027] The new round of matching deviation is obtained by comparing the distance between the second feature set and the first preset threshold. If the new round of matching deviation is less than the first preset threshold, the true restoration level corresponding to the current cluster is determined. Otherwise, the parameter modification step is returned to form a closed loop until the condition is met, and the updated cluster is obtained.

[0028] Preferably, the method for obtaining the dynamic matching function includes:

[0029] By preprocessing the updated clusters and operational behavior data, a standardized dataset is obtained;

[0030] A stratified sampling method was used to group the standardized dataset according to time series and behavioral categories to obtain a processed structured dataset.

[0031] Based on the structured dataset, a support vector machine model is trained, and key behavioral patterns and intra-cluster variation trends are extracted to determine the classification boundary of the model.

[0032] If the classification boundary meets the second preset threshold, the trained model is applied to the haptic feedback response mechanism; if it does not meet the threshold, the structured data set is cleaned a second time, the data grouping strategy is readjusted, and new training base data is obtained.

[0033] The dynamic matching function is generated by analyzing the output of the tactile feedback response mechanism.

[0034] Preferably, the method for obtaining the optimized feedback sequence includes:

[0035] By collecting feedback response data, initial response delay information and resistance difference information are obtained;

[0036] Based on the comparison between the response delay information and the preset threshold, if the response delay is lower than the third preset threshold, the resistance difference information is integrated to obtain an initial optimized feedback sequence; if the response delay exceeds the third preset threshold, the data exceeding the delay is classified and optimized through the iterative processing of the support vector machine to obtain a refined function.

[0037] Key sequence fragments are extracted from the initial optimized feedback sequence, and the matching degree between the sequence fragments and the function is determined by combining the output of the refined function to determine the final feedback adjustment direction.

[0038] For the final feedback adjustment direction, the adjusted response delay data and resistance difference data are obtained, and the adjusted optimized feedback sequence is obtained through data analysis and comparison.

[0039] Preferably, the method for obtaining the calibrated perception model includes:

[0040] The latest data on the organizational status is obtained from the optimized feedback sequence, and the status updates are initially filtered to obtain the preliminary sorted status information.

[0041] The state information is standardized, and based on the processed data, a preset threshold is used to detect the deviation of tactile perception. If the detection result exceeds the fourth preset threshold, a real-time calibration process is triggered to obtain calibrated perception data.

[0042] Key features are extracted from the calibrated sensing data and combined with the output of the processing stage to construct an initial sensing model;

[0043] For the initial perception model, the model parameters are dynamically adjusted using the update frequency to obtain the adjusted model framework. Based on the adjusted model framework, the results of state updates and real-time calibration are fused to construct the calibrated perception model.

[0044] Preferably, the method for obtaining the surgical trajectory includes:

[0045] The intensity change data is collected and standardized to obtain a preliminary integrated input data stream.

[0046] For the input data stream, a pre-established tactile feedback generation mechanism is used to generate corresponding tactile feedback signals;

[0047] Based on the tactile feedback signal, the corresponding operation behavior record is obtained, and the two are matched by the behavior comparison module to determine whether they meet the consistency standard.

[0048] If the tactile feedback signal and the feature matching degree of the operation behavior record reach the fifth preset threshold, it is determined to be consistent and a consistency confirmation signal is generated; if it does not reach the fifth preset threshold, it is marked as inconsistent and deviation information is recorded.

[0049] Using the consistency confirmation signal and combined with dynamic matching logic, the tactile feedback signal and operation behavior record are finally correlated to obtain dynamic matching result data.

[0050] Based on the dynamic matching result data, the data is archived into a preset database using a storage unit, and the corresponding surgical procedure trajectory is generated through a log recording system.

[0051] The present invention also provides a urological surgery simulation training system based on tactile feedback. The system applies the above-mentioned method and includes: a data acquisition module, a clustering and grouping module, a tactile feature extraction module, a model building module, a feedback data acquisition module, a model calibration module, and a feedback output module.

[0052] The data acquisition module collects operational behavior data and force change data through sensors to obtain a preliminary perception sequence.

[0053] The clustering and grouping module is used to group the force changes and resistance differences according to the preliminary sensing sequence using K-means clustering to obtain the classified tissue state clusters.

[0054] The tactile feature extraction module is used to extract tactile perception features from the classified tissue state clusters. If the tactile perception features match a first preset threshold, the true restoration level is determined; otherwise, the clustering parameters are adjusted to obtain an updated cluster.

[0055] The model building module is used to build a support vector machine model to train the tactile feedback response for the updated cluster and operation behavior data, and obtain a dynamic matching function.

[0056] The feedback data acquisition module processes the tactile feedback response data through the dynamic matching function. If the response delay is determined to be lower than a set value, the module integrates the resistance difference to obtain an optimized feedback sequence; otherwise, it iterates the support vector machine model.

[0057] The model calibration module is used to obtain tissue state updates from the optimized feedback sequence, and to perform real-time calibration of tactile perception using the information processing stage to obtain the calibrated perception model.

[0058] The feedback output module generates tactile feedback output based on the calibrated perception model and force change data. If the output matches the operation behavior, the final dynamic matching result is determined.

[0059] The present invention also provides an electronic device, characterized in that it includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described tactile feedback-based urological surgery simulation training method.

[0060] The present invention also provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, which, when executed, implements the above-described tactile feedback-based urological surgery simulation training method.

[0061] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0062] This invention discloses a dynamic sensing and optimization technology based on tactile feedback, proposing an integrated solution to address the problems of insufficient realism, high response latency, and inaccurate tactile perception in business scenarios such as medical surgical simulations or remote operations. By collecting data on operational behavior and force changes, a clustering algorithm is used to classify tissue states, extract tactile features and match them with thresholds, and dynamically adjust parameters to optimize the classification results. Subsequently, a support vector machine model is constructed to train the tactile response, generating a dynamic matching function to process the feedback data, ensuring that the latency is below a set value. Finally, resistance difference information is fused, and the sensing model is calibrated in real time to generate a tactile output consistent with the operation. The core innovation of this invention lies in achieving high realism and low latency in tactile feedback through multi-level data processing and model optimization, providing accurate dynamic matching results for complex operation scenarios, and significantly improving user experience and operational precision. Attached Figure Description

[0063] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are briefly introduced below. Obviously, the 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.

[0064] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention;

[0065] Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention.

[0066] Explanation of reference numerals in the attached figures:

[0067] 1010, Processor; 1020, Memory; 1030, Input / Output Interface; 1040, Communication Interface; 1050, Bus. Detailed Implementation

[0068] 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.

[0069] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this disclosure should have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms "first," "second," and similar terms used in the embodiments of this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0070] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0071] Example 1

[0072] As the background technology shows, when simulating a complex surgical procedure, doctors may not be able to accurately judge whether the instruments have touched key tissues due to a lack of real tactile sensation, nor can they feel subtle changes in the force of operation. This misalignment between perception and behavior seriously affects the development of skills.

[0073] This invention provides a urological surgical simulation training method based on tactile feedback, such as... Figure 1 As shown, it includes the following steps:

[0074] S1. Collect operational behavior data and force change data through sensors to obtain a preliminary perception sequence.

[0075] In this embodiment, a force feedback glove is used as a data acquisition device. The sensor therein acquires data at a frequency of 100Hz, obtaining a set of sequence data including timestamp, coordinates X / Y, and pressure value F, thus obtaining a preliminary sensing sequence.

[0076] S2. Based on the preliminary sensing sequence, K-means clustering is used to group the force changes and resistance differences to obtain the classified tissue state clusters.

[0077] The method for obtaining the classified tissue state clusters includes: obtaining the original records of force changes and resistance differences from the initial sensing sequence, and removing outliers using a pre-established data cleaning process to obtain a processed basic dataset; grouping the force changes and resistance differences into preliminary classification clusters using the K-means clustering method for the processed basic dataset; analyzing the change characteristics and difference identification results within each cluster based on the preliminary classification clusters to determine the corresponding tissue state category; if the tissue state category division does not match the preset threshold range, optimizing the classification clusters by readjusting the clustering parameters to obtain adjusted state clusters; extracting the state division features within each cluster based on the adjusted state clusters, judging their matching degree with the initial sensing sequence to obtain the final tissue state distribution; generating a corresponding state division mapping table based on the final tissue state distribution, and determining the classification cluster category to which each sensing data point belongs based on the mapping table to obtain the classified tissue state clusters.

[0078] In this embodiment, the original sensing sequence is first cleaned and its features are extracted. The sequence acquired by the sensor at 100Hz contains timestamps, coordinates (X, Y), and pressure values ​​F. Feature dimensions for clustering need to be extracted from these. Typically, the pressure gradient over time (ΔF / Δt) is selected as the force change feature, and the pressure change per unit displacement is calculated in conjunction with the coordinate displacement as an approximate representation of the resistance difference. Data cleaning employs statistical outlier removal methods, such as using box plots or Z-score standardization combined with threshold filtering for the pressure sequence to remove data points exceeding the mean ± 3 standard deviations, ensuring the stability of the input data.

[0079] Subsequently, K-means clustering is used to group the processed feature vectors. Let the feature vector set be... Each It includes two types of features: changes in force and differences in resistance. The algorithm iteratively optimizes the cluster centers by minimizing the sum of squared errors within the cluster (SSE), and its objective function is:

[0080]

[0081] Where K represents the preset number of clusters, This represents the k-th cluster. This represents the centroid of the cluster. The selection of the K value is usually determined by combining the Elbow Method and the Silhouette Coefficient to ensure that the clustering results are both discriminative and not overly fragmented.

[0082] The initial centroids are initialized using the K-means++ strategy to improve convergence efficiency and stability. After obtaining the initial clusters, each cluster needs to be classified into biomechanical states. By analyzing the statistical characteristics of the data within each cluster (such as mean pressure, variance of fluctuation, and rate of change distribution), combined with expert prior knowledge (such as typical resistance ranges for different tissue types), the clusters are mapped to state categories such as "soft tissue," "blood vessel wall," "ligament," and "bone." If the characteristics of some clusters deviate significantly from the preset physiological threshold range, the parameter adjustment process is triggered. Adjustment methods include resetting the K value, introducing feature weights (such as assigning higher importance to resistance differences), or using feature scaling to optimize the distance metric.

[0083] Finally, a state partitioning mapping table is generated based on the optimized clustering results. This table records the centroid coordinates, feature range, and corresponding tissue state label of each cluster, and all sensory data points are classified based on this. This output is the classified tissue state cluster.

[0084] S3. Extract tactile perception features from the classified tissue state clusters. If the tactile perception features match the first preset threshold, determine the true restoration level; otherwise, adjust the clustering parameters to obtain updated clusters.

[0085] The method for obtaining the updated clusters includes: extracting tactile feature vectors from the classified tissue state clusters to obtain an initial feature set; calculating the distance difference between the feature vectors and a first preset threshold based on the initial feature set to obtain a matching deviation; if the matching deviation is less than the first preset threshold, outputting the corresponding true restoration level to obtain the final level label; if the matching deviation is greater than the first preset threshold, triggering a clustering parameter modification instruction to obtain an adjusted parameter set; performing a K-clustering algorithm on the original tissue state data using the adjusted parameter set to obtain the first-generation updated clusters; re-extracting tactile feature vectors from the first-generation updated clusters to obtain a second feature set; performing a distance comparison between the second feature set and the first preset threshold to obtain a new round of matching deviation; if the new round of matching deviation is less than the first preset threshold, determining the true restoration level corresponding to the current cluster; otherwise, continuing to return to the parameter modification step to form a closed loop until the conditions are met, thus obtaining the updated clusters.

[0086] In this embodiment, firstly, quantitative tactile sensory feature vectors are extracted from the classified tissue state clusters. Each state cluster can be considered as a data distribution, from which a set of statistical features are extracted to constitute the sensory feature vector of that cluster. Key feature f iTypically includes: the average value of intra-cluster pressure readings with standard deviation (Reflecting basic strength and stability), the peak value and entropy of the pressure change rate sequence (characterizing dynamic resistance characteristics), and the covariance of coordinate displacement and pressure change (depicting the viscosity in motion). These features together constitute the initial feature set for evaluating the realism of the feel.

[0087] Next, the matching degree between the feature vector of each cluster and a first preset threshold (i.e., a preset ideal range of tactile features defined by expert experience or a physical model) is calculated. The matching degree is determined by calculating the match between the feature vector and the threshold vector. The distance difference between them (i.e., the matching deviation) This is used to quantify distances. Commonly used distance metrics include weighted Euclidean distance:

[0088]

[0089] in, This represents the weight assigned to the importance of subjective feel based on different features. If a certain cluster of... Less than the preset fault tolerance threshold If the tactile reproduction of the cluster is satisfactory, it will be marked with a true reproduction level (such as high fidelity, standard, etc.).

[0090] If clusters exist This triggers a clustering parameter modification command. The adjustment is not done blindly, but based on bias analysis: if the characteristic means of multiple clusters are too similar, it may be necessary to increase the number of clusters K; if the standard deviation within a cluster is too close, it may be necessary to adjust the clustering parameters. If the value is too large, it can lead to ambiguity in the definition of touch. In such cases, it may be necessary to adjust the distance metric weights in the K-means algorithm or perform a non-linear transformation on the original features to better separate the categories. This generates an "adjusted parameter set".

[0091] Finally, K-means clustering is re-executed on the original tissue state data of S2 using the adjusted parameters to generate the first-generation updated clusters. A closed-loop optimization process then begins: feature vectors are re-extracted from the new clusters, a new round of matching bias is calculated, and compared with a threshold. This process iterates until the matching bias of all clusters meets the threshold. The maximum number of iterations may be reached. The final output is the updated cluster that has passed the feel authenticity verification.

[0092] S4. For the updated clusters and operation behavior data, construct a support vector machine model to train the tactile feedback response and obtain the dynamic matching function.

[0093] The method for obtaining the dynamic matching function includes: preprocessing the updated clusters and operational behavior data to obtain a standardized dataset; using stratified sampling, grouping the standardized dataset according to time series and behavior categories to obtain a processed structured dataset; training a support vector machine model based on the structured dataset, extracting key behavior patterns and intra-cluster change trends to determine the model's classification boundary; if the classification boundary meets a second preset threshold, applying the trained model to the haptic feedback response mechanism; if not, performing secondary cleaning of the structured dataset, readjusting the data grouping strategy, and obtaining new training base data; and generating the dynamic matching function based on the output of the haptic feedback response mechanism.

[0094] In this embodiment, the input data is first preprocessed and structured. The operational behavior data includes a timestamp and coordinate X. t Y t The sequence of data, and the updated cluster data, are essentially a set of pressure-resistance features with labels (such as "soft tissue_high fidelity"). This step requires aligning the two on a timeline to construct a sample set for SVM training. The feature vector for each sample... It consists of two parts: one part is p temporal features b extracted from the sequence of operational actions. i One part consists of instantaneous velocity v, acceleration a, and trajectory curvature; the other part consists of q state features s extracted from the update cluster to which the corresponding time point belongs. i For example, the average pressure of this cluster and the standard deviation of the gradient Tag y j This is a quantified value of the expected tactile feedback intensity at this moment, which may be derived from a pre-defined force-displacement physics model or expert scoring. All features need to be Z-score standardized to form a standardized dataset.

[0095] Next, a stratified sampling method is used to divide the training and test sets, and SVM model training is then performed. Considering the temporal continuity and class balance of the data, stratification by time block and organizational state category is necessary to ensure that behavioral patterns in each state are fully learned. The goal of the support vector machine is to find an optimal hyperplane that clearly distinguishes the feature vectors of different tactile feedback levels. For non-linearly separable tactile mapping problems, the radial basis function (RBF) kernel function is typically used.

[0096]

[0097] in, Let be the kernel parameters. The optimization objective is to minimize the structural risk, and its decision function has the form: ,in, For Lagrange multipliers, This is the bias term. The training process involves adjusting parameters C (penalty coefficient) and γ to find the model that provides the clearest classification boundary and the strongest generalization ability.

[0098] Then, the training results are validated and iteratively optimized. The model is evaluated using a test set, and its classification accuracy, precision, and recall are calculated. Here, the classification boundary meets the second preset threshold, meaning that the model's overall performance metric M (such as F1-score) on the test set must be higher than the set threshold η. If... This indicates that the model has failed to fully learn the complex mapping from behavior-state to haptic feedback, triggering the optimization process. Optimization measures include: performing secondary cleaning on the structured dataset to check and correct feature alignment errors; or readjusting the data grouping strategy, such as changing the window size for extracting temporal features from operational behavior to capture more effective dynamic patterns, thereby generating new training data for retraining.

[0099] Finally, the SVM model that meets the performance requirements is encapsulated into a dynamic matching function. This function can receive feature vectors that are acquired and constructed in real time. It outputs a precise, continuous or discrete haptic feedback command value.

[0100] S5. Process the tactile feedback response data through a dynamic matching function. If the response delay is determined to be lower than the set value, then the resistance difference is fused to obtain an optimized feedback sequence; otherwise, the support vector machine model is iterated.

[0101] The method for obtaining the optimized feedback sequence includes: acquiring initial response delay information and resistance difference information by collecting feedback response data; comparing the response delay information with a preset threshold, if the response delay is lower than a third preset threshold, integrating the resistance difference information to obtain the initial optimized feedback sequence; if the response delay exceeds the third preset threshold, iterative processing using a support vector machine is used to classify and optimize the data exceeding the delay to obtain a refined function; extracting key sequence fragments from the initial optimized feedback sequence, and combining the output of the refined function to determine the matching degree between the sequence fragments and the function, thus determining the final feedback adjustment direction; and for the final feedback adjustment direction, acquiring the adjusted response delay data and resistance difference data, and obtaining the adjusted optimized feedback sequence through data analysis and comparison.

[0102] In this embodiment, firstly, initial feedback response data is collected and analyzed. When the dynamic matching function... The target tactile feedback command is calculated based on the real-time feature vector. Then, the actuator (such as the motor in a force feedback glove) will generate an actual force signal. The time interval between the issuance of the command and the actual force signal reaching a stable value (e.g., 95% of the target value) is recorded and defined as the response delay. Simultaneously, resistance difference information R is extracted from the current tissue state cluster. This information is one of the key features of the S2 clustering output, such as the instantaneous rate of change of the pressure gradient. With the third preset threshold This is the critical latency (typically less than 50 milliseconds) that ensures real-time haptic feedback and immersion.

[0103] like This indicates a rapid response, initiating the optimized feedback sequence generation process. The core idea is to fuse the discrete or step-like commands output by the dynamic matching function with continuous resistance difference information R to generate a smoother, more natural force feedback curve. Specifically, a first-order low-pass filtering and weighted interpolation method is used: let the target command at time n be... The actual optimized output at the previous time step was Then the optimized output at this moment Calculated by the following formula:

[0104]

[0105] in, Indicates the inertial smoothing coefficient. , This represents the resistance difference gain coefficient. This formula ensures that the change in feedback force follows the model decision while also taking into account the physical inertia of the mechanical response and the organization's real-time resistance perception, thus obtaining the initial optimized feedback sequence.

[0106] like If the latency exceeds the limit, the SVM model needs to be iteratively refined. High latency often stems from the dynamic matching function. Excessive computational complexity or insufficient clarification of novel operational modes (not adequately included in the training set) leads to decision hesitation. In this case, the current outlier data pairs causing delays (i.e., feature vectors) are considered... and its actual tactile feedback results This data is stored in a temporary buffer pool. When such data accumulates to a certain scale, it triggers incremental learning or parameter retuning of the support vector machine. For example, adjusting the bandwidth parameter of the RBF kernel function. or penalty factor By retraining using the existing training set plus new data, a refined function with higher computational efficiency or clearer decision boundaries can be obtained. This process aims to reduce decision-making delays when encountering similar scenarios in the future.

[0107] Finally, a verification loop is initiated. Both the initially generated optimized sequence and the new instructions generated after model iterations will re-enter the latency detection phase. Key sequence segments (such as intervals of rapid force changes) are extracted from the optimized sequence, and their average latency is calculated. and the degree of matching with the target curve This determines the final direction of the feedback adjustment. This direction may guide further fine-tuning of the inertial smoothing coefficient. Gain coefficient with resistance difference This may also confirm that the model iteration has met the requirements. After multiple rounds of adjustments and verifications, the final output is an adjusted optimized feedback sequence that meets both real-time requirements and deeply integrates the characteristics of tissue resistance changes.

[0108] S6. Obtain the organization status update from the optimized feedback sequence, and use the information processing step to perform real-time calibration of the sensory perception to obtain the calibrated perception model.

[0109] The method for obtaining the calibrated perception model includes: acquiring the latest data on tissue status from the optimized feedback sequence, performing preliminary screening on the status updates to obtain preliminary sorted status information; standardizing the status information, and detecting deviations in tactile perception using a preset threshold based on the processed data; if the detection result exceeds a fourth preset threshold, triggering a real-time calibration process to obtain calibrated perception data; extracting key features from the calibrated perception data, and constructing an initial perception model by combining the output of the processing steps; dynamically adjusting the model parameters using the update frequency to obtain the adjusted model framework; and constructing the calibrated perception model by fusing the results of status updates and real-time calibration based on the adjusted model framework.

[0110] In this embodiment, firstly, the implicit organizational state update is parsed from the optimized feedback sequence. This sequence contains force feedback data that has been smoothed and delayed, along with its corresponding timestamps and coordinates. The system calculates the force curve features within each short time window, which constitute the state information describing the current operation. Subsequently, these features are standardized in the same manner as in stage S2, forming a standardized feature vector that can be compared with historical cluster centers. .

[0111] Next, the system performs deviation detection and real-time calibration based on tactile feedback. The input is fed into the currently running perception model, which outputs a predicted tissue state label (such as "blood vessel wall") and its confidence level. Simultaneously, The central feature vectors of various types of organizational state clusters generated by S2 Perform similarity calculation. If the predicted label matches the label of the most similar cluster, and the confidence level is higher than the fourth preset threshold... If the confidence level is below the threshold, the perception is considered accurate and no calibration is required. If label inconsistencies occur, a real-time calibration process will be triggered. Calibration will be performed by... With a certain weight This is achieved by merging the vectors into the center vectors of the most similar clusters: This incremental update method allows the cluster center to slowly adapt to the new data pattern and output calibrated sensing data.

[0112] Then, based on the updated cluster centers, the perception model is reconstructed or adjusted. The system extracts the latest feature distribution parameters from all updated clusters and, based on these, recalculates the discrimination boundary used for fast classification. This typically manifests as a parameter update of a lightweight classifier, forming the initial perception model. This model is able to more accurately map real-time perceived features to the latest organizational state category.

[0113] Finally, the model is dynamically fine-tuned based on the data update frequency to generate the final calibrated perception model. The frequency and magnitude of calibration triggers are monitored. If the center of a certain cluster continuously moves in the same direction for multiple consecutive periods, the system will increase the learning rate for that category to accelerate its adaptation; conversely, it will decrease the learning rate to maintain stability. Through this adaptive mechanism, the model parameters are dynamically adjusted, and the final calibrated perception model is a lightweight online learning system capable of incorporating the latest feedback information and correcting its internal state representation in real time.

[0114] S7. Based on the calibrated perception model and force change data, generate tactile feedback output. If the output is consistent with the operation behavior, determine the final dynamic matching result. Based on the final dynamic matching result, archive it to the preset database using the storage unit, and generate the corresponding surgical processing trajectory through the log recording system.

[0115] The method for obtaining the surgical procedure trajectory includes: collecting and standardizing force change data to obtain a pre-integrated input data stream; generating corresponding tactile feedback signals using a pre-established tactile feedback generation mechanism for the input data stream; acquiring corresponding operation behavior records based on the tactile feedback signals, and performing feature matching between the two through a behavior comparison module to determine whether they meet the consistency standard; if the feature matching degree between the tactile feedback signal and the operation behavior record reaches a fifth preset threshold, they are determined to be consistent, and a consistency confirmation signal is generated; if the fifth preset threshold is not reached, they are marked as inconsistent and deviation information is recorded; using the consistency confirmation signal and dynamic matching logic, the tactile feedback signal and the operation behavior record are finally correlated to obtain dynamic matching result data; based on the dynamic matching result data, it is archived in a preset database using a storage unit, and the corresponding surgical procedure trajectory is generated through a log recording system.

[0116] In this embodiment, firstly, data integration and tactile feedback signal generation are performed: real-time collected force change data and coordinate data are fed into a calibrated perception model. This model quickly determines the current tissue state category and its confidence level, and combines this state label with operational behavior features (such as speed) to form a composite feature vector. Subsequently, this vector is input into a dynamically matched function trained by S4, which calculates the theoretically appropriate ideal tactile feedback signal that best matches the current scene. This signal is immediately sent to the force feedback actuator, driving it to generate actual physical feedback, thereby completing the entire closed-loop process from perception to execution.

[0117] Next, a rigorous output consistency verification is performed: the actual tactile feedback signal and the original operation behavior sequence from the glove sensor are recorded simultaneously. Feature matching analysis is then performed on both through behavior comparison. The extracted temporal features (such as peak values ​​and durations) and the corresponding action features of the original operation behavior sequence (such as the direction of applied force and displacement) are compared to calculate the feature matching degree. The matching degree is calculated using a predefined similarity metric (such as Dynamic Time Warping (DTW) distance or the correlation coefficient of specific features). If the feature matching degree reaches a fifth preset threshold... If the tactile feedback is highly consistent with the operational behavior, a consistency confirmation signal is generated, indicating that the system simulation is realistic and effective.

[0118] Then, based on the consistency confirmation, the final dynamic matching result is generated. This result is a structured data packet containing information such as timestamps, operation behavior codes, perception state labels, ideal feedback signals, actual feedback signals, and matching degrees. This signifies the successful matching and verification of the entire dynamic link from perception and decision-making to feedback at the current moment.

[0119] Finally, the system initiates data archiving and trajectory generation: all dynamic matching result data is archived in real time to a preset relational or time-series database through storage units. Each record serves as a basic event in the surgical simulation. The logging system then links these events together chronologically, adding contextual information (such as the tissue state cluster mapping table version in S2 and the dynamic matching function ID used in S4) to generate a structured surgical processing trajectory. This trajectory fully reproduces the tactile interaction logic of the entire simulated surgery, providing an immutable data foundation for subsequent performance evaluation, skill analysis, or model iteration.

[0120] This invention discloses a dynamic perception and optimization technology based on tactile feedback. It proposes an integrated solution to address the problems of insufficient realism, high response latency, and inaccurate tactile perception in business scenarios such as medical surgical simulations or remote operations. By collecting data on operational behavior and force changes, a clustering algorithm is used to classify tissue states, extract tactile features, and match them with thresholds. Parameters are dynamically adjusted to optimize the classification results. Subsequently, a support vector machine model is constructed to train the tactile response, generating a dynamic matching function to process the feedback data, ensuring that the latency is below a set value. Finally, resistance difference information is fused, and the perception model is calibrated in real time to generate a tactile output consistent with the operation. The core innovation of this invention lies in achieving high realism and low latency in tactile feedback through multi-level data processing and model optimization, providing accurate dynamic matching results for complex operation scenarios, and significantly improving user experience and operational precision.

[0121] It should be noted that the method of this disclosure embodiment can be executed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method of this disclosure embodiment, and the multiple devices will interact with each other to complete the method described.

[0122] It should be noted that the above description describes some embodiments of this disclosure. Other embodiments are within the scope of the appended claims. In some cases, it should be understood that the sequence number of each step in the above embodiments does not imply the order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention. The actions or steps recorded in the claims can be performed in a different order than that in the above embodiments and can still achieve the desired result. In addition, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0123] Example 2

[0124] Based on the same inventive concept, corresponding to any of the above embodiments, the present invention also provides a urological surgical simulation training system based on tactile feedback, including: a data acquisition module, a clustering and grouping module, a tactile feature extraction module, a model building module, a feedback data acquisition module, a model calibration module, and a feedback output module.

[0125] The data acquisition module collects operational behavior data and force change data through sensors to obtain a preliminary perception sequence.

[0126] The clustering and grouping module is used to group the force changes and resistance differences based on the preliminary sensing sequence using K-means clustering, resulting in classified tissue state clusters.

[0127] The workflow of the clustering and grouping module includes: obtaining raw records of force changes and resistance differences from the initial sensing sequence, and removing outliers using a pre-established data cleaning process to obtain a processed basic dataset; grouping the force changes and resistance differences into preliminary classification clusters using the K-means clustering method for the processed basic dataset; analyzing the change characteristics and difference identification results within each cluster based on the preliminary classification clusters to determine the corresponding organizational state category; optimizing the classification clusters by readjusting the clustering parameters if the organizational state category classification does not match the preset threshold range to obtain adjusted state clusters; extracting the state classification features within each cluster based on the adjusted state clusters, judging their matching degree with the initial sensing sequence to obtain the final organizational state distribution; generating a corresponding state classification mapping table based on the final organizational state distribution, and determining the classification cluster category to which each sensing data point belongs based on the mapping table to obtain the classified organizational state clusters.

[0128] The tactile feature extraction module is used to extract tactile perception features from the classified tissue state clusters. If the tactile perception features match the first preset threshold, the true restoration level is determined; otherwise, the clustering parameters are adjusted to obtain updated clusters.

[0129] The workflow of the tactile feature extraction module includes: extracting tactile perception feature vectors from the classified tissue state clusters to obtain an initial feature set; calculating the distance difference between the feature vectors and a first preset threshold based on the initial feature set to obtain a matching deviation; if the matching deviation is less than the first preset threshold, outputting the corresponding true restoration level to obtain the final level label; if the matching deviation is greater than the first preset threshold, triggering a clustering parameter modification instruction to obtain an adjusted parameter set; performing a K-clustering algorithm on the original tissue state data using the adjusted parameter set to obtain the first-generation updated cluster; re-extracting tactile perception feature vectors from the first-generation updated cluster to obtain a second feature set; performing a distance comparison between the second feature set and the first preset threshold to obtain a new round of matching deviation; if the new round of matching deviation is less than the first preset threshold, determining the true restoration level corresponding to the current cluster; otherwise, continuing to return to the parameter modification step to form a closed loop until the conditions are met to obtain the updated cluster.

[0130] The model building module is used to build a support vector machine model to train the tactile feedback response based on the updated cluster and operation behavior data, and obtain the dynamic matching function.

[0131] The workflow of the model building module includes: preprocessing the updated clusters and operational behavior data to obtain a standardized dataset; using stratified sampling, grouping the standardized dataset according to time series and behavior categories to obtain a processed structured dataset; training a support vector machine model based on the structured dataset, extracting key behavior patterns and intra-cluster change trends, and determining the model's classification boundary; if the classification boundary meets a second preset threshold, applying the trained model to the haptic feedback response mechanism; if not, performing secondary cleaning of the structured dataset, readjusting the data grouping strategy, and obtaining new training base data; and generating a dynamic matching function based on the output of the haptic feedback response mechanism.

[0132] The feedback data acquisition module processes the tactile feedback response data through a dynamic matching function. If the response delay is determined to be lower than a set value, the resistance difference is fused to obtain an optimized feedback sequence; otherwise, the support vector machine model is iterated.

[0133] The workflow of the feedback data acquisition module includes: acquiring initial response delay and resistance difference information by collecting feedback response data; comparing the response delay information with a preset threshold, if the response delay is lower than a third preset threshold, integrating the resistance difference information to obtain an initial optimized feedback sequence; if the response delay exceeds the third preset threshold, iterative processing using a support vector machine is used to classify and optimize the data exceeding the delay to obtain a refined function; extracting key sequence fragments from the initial optimized feedback sequence, and combining the output of the refined function to determine the matching degree between the sequence fragments and the function, thus determining the final feedback adjustment direction; and acquiring the adjusted response delay and resistance difference data based on the final feedback adjustment direction, and obtaining the adjusted optimized feedback sequence through data analysis and comparison.

[0134] The model calibration module is used to obtain tissue state updates from the optimization feedback sequence and uses the information processing step to perform real-time calibration of the sensory perception to obtain the calibrated sensory model.

[0135] The workflow of the model calibration module includes: obtaining the latest organizational status data from the optimization feedback sequence; performing preliminary screening on the status updates to obtain preliminary organized status information; standardizing the status information and detecting deviations in tactile perception using preset thresholds based on the processed data; if the detection result exceeds a fourth preset threshold, triggering the real-time calibration process to obtain calibrated perception data; extracting key features from the calibrated perception data and constructing an initial perception model by combining the output of the processing steps; dynamically adjusting the model parameters using the update frequency to obtain the adjusted model framework; and constructing a calibrated perception model by fusing the status update and real-time calibration results based on the adjusted model framework.

[0136] The feedback output module generates tactile feedback output based on the calibrated perception model and force change data. If the output matches the operation behavior, the final dynamic matching result is determined.

[0137] The method for obtaining the surgical procedure trajectory includes: collecting and standardizing force change data to obtain a pre-integrated input data stream; generating corresponding tactile feedback signals using a pre-established tactile feedback generation mechanism for the input data stream; acquiring corresponding operation behavior records based on the tactile feedback signals, and performing feature matching between the two through a behavior comparison module to determine whether they meet the consistency standard; if the feature matching degree between the tactile feedback signal and the operation behavior record reaches a fifth preset threshold, they are determined to be consistent, and a consistency confirmation signal is generated; if the fifth preset threshold is not reached, they are marked as inconsistent and deviation information is recorded; using the consistency confirmation signal and dynamic matching logic, the tactile feedback signal and the operation behavior record are finally correlated to obtain dynamic matching result data; based on the dynamic matching result data, it is archived in a preset database using a storage unit, and the corresponding surgical procedure trajectory is generated through a log recording system.

[0138] The system described above is used to implement the corresponding tactile feedback-based urological surgery simulation training method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0139] It should be noted that the aforementioned tactile feedback-based urological surgical simulation training system is embodied in the form of functional units. The term "module" here can be implemented in software and / or hardware, without specific limitations.

[0140] For example, a "module" can be a software program, hardware circuit, or a combination of both that implements the above functions. Hardware circuits may include application-specific integrated circuits (ASICs), electronic circuits, processors (e.g., shared processors, proprietary processors, or group processors) and memory for executing one or more software or firmware programs, combined logic circuits, and / or other suitable components that support the described functions.

[0141] Example 3

[0142] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the tactile feedback-based urological surgery simulation training method described in any of the above embodiments.

[0143] Figure 2This embodiment illustrates a more specific hardware structure of an electronic device, which may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.

[0144] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0145] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.

[0146] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0147] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB (Universal Serial Bus), network cable, etc.) or wireless means (such as mobile network, WIFI (Wireless Fidelity), Bluetooth, etc.).

[0148] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0149] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0150] The system described above is used to implement the corresponding tactile feedback-based urological surgery simulation training method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0151] Example 4

[0152] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, the present invention also provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the tactile feedback-based urological surgery simulation training method as described in any of the above embodiments.

[0153] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0154] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the urological surgery simulation training method based on tactile feedback as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0155] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this disclosure (including the claims) is limited to these examples; within the framework of this disclosure, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this disclosure as described above, which are not provided in detail for the sake of brevity.

[0156] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this disclosure, the provided drawings may or may not show well-known power / ground connections to integrated circuit (IC) chips and other components. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this disclosure, and this also takes into account the fact that the details of implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this disclosure will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuitry) have been set forth to describe exemplary embodiments of this disclosure, it will be apparent to those skilled in the art that the embodiments of this disclosure may be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0157] Although this disclosure has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0158] Therefore, the units of the various examples described in the embodiments of this application can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0159] This disclosure is intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A urological surgical simulation training method based on tactile feedback, characterized in that, Includes the following steps: By collecting operational behavior data and force change data through sensors, a preliminary perception sequence is obtained; Based on the preliminary sensing sequence, K-means clustering was used to group the force changes and resistance differences to obtain the classified tissue state clusters. Tactile perception features are extracted from the classified tissue state clusters. If the tactile perception features match a first preset threshold, the true restoration level is determined; otherwise, the clustering parameters are adjusted to obtain an updated cluster. For the updated clusters and operation behavior data, a support vector machine model is constructed to train the tactile feedback response, and a dynamic matching function is obtained; The tactile feedback response data is processed by the dynamic matching function. If the response delay is determined to be lower than a set value, the resistance difference is fused to obtain an optimized feedback sequence; otherwise, the support vector machine model is iterated. The organization state update is obtained from the optimized feedback sequence, and the sensory perception is calibrated in real time using the information processing step to obtain the calibrated sensory model. Based on the calibrated perception model and force change data, a tactile feedback output is generated. If the output matches the operation behavior, the final dynamic matching result is determined. Based on the final dynamic matching result, it is archived to a preset database using a storage unit, and the corresponding surgical procedure trajectory is generated through a log recording system.

2. The urological surgical simulation training method based on tactile feedback according to claim 1, characterized in that, The method for obtaining the classified tissue state clusters includes: The original records of the force change and the resistance difference are obtained from the initial sensing sequence, and outliers are removed using a pre-established data cleaning process to obtain the processed basic dataset. For the processed basic dataset, the K-means clustering method is used to group the force changes and resistance differences to obtain preliminary classification clusters; Based on the preliminary classification of clusters, the variation characteristics and difference identification results within each cluster are analyzed to determine the organizational state category corresponding to the cluster; If the classification of the organizational state categories does not match the preset threshold range, the classification clusters are optimized by readjusting the clustering parameters to obtain the adjusted state clusters. Based on the adjusted state clusters, extract the state division features within each cluster, determine their matching degree with the preliminary sensing sequence, and obtain the final organizational state distribution; Based on the final organizational state distribution, a corresponding state classification mapping table is generated, and the classification cluster category to which each sensing data point belongs is determined based on the mapping table, thus obtaining the classified organizational state cluster.

3. The urological surgical simulation training method based on tactile feedback according to claim 1, characterized in that, The method for obtaining the updated cluster includes: An initial feature set is obtained by extracting tactile perception feature vectors from the classified tissue state clusters; The matching deviation is obtained by calculating the distance difference between the feature vector and the first preset threshold based on the initial feature set; If the matching deviation is less than the first preset threshold, the corresponding true restoration level is output to obtain the final level label; if the matching deviation is greater than the first preset threshold, a clustering parameter modification instruction is triggered to obtain the adjusted parameter group. The first-generation updated clusters are obtained by performing a K-clustering algorithm on the original tissue state data using the adjusted parameter set. The second feature set is obtained by re-extracting the tactile perception feature vector from the first generation update cluster; The new round of matching deviation is obtained by comparing the distance between the second feature set and the first preset threshold. If the new round of matching deviation is less than the first preset threshold, the true restoration level corresponding to the current cluster is determined. Otherwise, the parameter modification step is returned to form a closed loop until the condition is met, and the updated cluster is obtained.

4. The urological surgical simulation training method based on tactile feedback according to claim 1, characterized in that, The methods for obtaining the dynamic matching function include: By preprocessing the updated clusters and operational behavior data, a standardized dataset is obtained; A stratified sampling method was used to group the standardized dataset according to time series and behavioral categories to obtain a processed structured dataset. Based on the structured dataset, a support vector machine model is trained, and key behavioral patterns and intra-cluster variation trends are extracted to determine the classification boundary of the model. If the classification boundary meets the second preset threshold, the trained model is applied to the haptic feedback response mechanism; if it does not meet the threshold, the structured data set is cleaned a second time, the data grouping strategy is readjusted, and new training base data is obtained. The dynamic matching function is generated by analyzing the output of the tactile feedback response mechanism.

5. The urological surgical simulation training method based on tactile feedback according to claim 1, characterized in that, The methods for obtaining the optimized feedback sequence include: By collecting feedback response data, initial response delay information and resistance difference information are obtained; Based on the comparison between the response delay information and the preset threshold, if the response delay is lower than the third preset threshold, the resistance difference information is integrated to obtain an initial optimized feedback sequence; if the response delay exceeds the third preset threshold, the data exceeding the delay is classified and optimized through the iterative processing of the support vector machine to obtain a refined function. Key sequence fragments are extracted from the initial optimized feedback sequence, and the matching degree between the sequence fragments and the function is determined by combining the output of the refined function to determine the final feedback adjustment direction. For the final feedback adjustment direction, the adjusted response delay data and resistance difference data are obtained, and the adjusted optimized feedback sequence is obtained through data analysis and comparison.

6. The urological surgical simulation training method based on tactile feedback according to claim 1, characterized in that, The method for obtaining the calibrated perception model includes: The latest data on the organizational status is obtained from the optimized feedback sequence, and the status updates are initially filtered to obtain the preliminary sorted status information. The state information is standardized, and based on the processed data, a preset threshold is used to detect the deviation of tactile perception. If the detection result exceeds the fourth preset threshold, a real-time calibration process is triggered to obtain calibrated perception data. Key features are extracted from the calibrated sensing data and combined with the output of the processing stage to construct an initial sensing model; For the initial perception model, the model parameters are dynamically adjusted using the update frequency to obtain the adjusted model framework. Based on the adjusted model framework, the results of state updates and real-time calibration are fused to construct the calibrated perception model.

7. The urological surgical simulation training method based on tactile feedback according to claim 1, characterized in that, The methods for obtaining the surgical procedure trajectory include: The intensity change data is collected and standardized to obtain a preliminary integrated input data stream. For the input data stream, a pre-established tactile feedback generation mechanism is used to generate corresponding tactile feedback signals; Based on the tactile feedback signal, the corresponding operation behavior record is obtained, and the two are matched by the behavior comparison module to determine whether they meet the consistency standard. If the tactile feedback signal and the feature matching degree of the operation behavior record reach the fifth preset threshold, it is determined to be consistent and a consistency confirmation signal is generated; if it does not reach the fifth preset threshold, it is marked as inconsistent and deviation information is recorded. Using the consistency confirmation signal and combined with dynamic matching logic, the tactile feedback signal and operation behavior record are finally correlated to obtain dynamic matching result data. Based on the dynamic matching result data, the data is archived into a preset database using a storage unit, and the corresponding surgical procedure trajectory is generated through a log recording system.

8. A urological surgical simulation training system based on tactile feedback, wherein the system applies the method described in any one of claims 1-7, characterized in that, include: The module includes a data acquisition module, a clustering and grouping module, a tactile feature extraction module, a model building module, a feedback data acquisition module, a model calibration module, and a feedback output module. The data acquisition module collects operational behavior data and force change data through sensors to obtain a preliminary perception sequence. The clustering and grouping module is used to group the force changes and resistance differences according to the preliminary sensing sequence using K-means clustering to obtain the classified tissue state clusters. The tactile feature extraction module is used to extract tactile perception features from the classified tissue state clusters. If the tactile perception features match a first preset threshold, the true restoration level is determined; otherwise, the clustering parameters are adjusted to obtain an updated cluster. The model building module is used to build a support vector machine model to train the tactile feedback response for the updated cluster and operation behavior data, and obtain a dynamic matching function. The feedback data acquisition module processes the tactile feedback response data through the dynamic matching function. If the response delay is determined to be lower than a set value, the module integrates the resistance difference to obtain an optimized feedback sequence; otherwise, it iterates the support vector machine model. The model calibration module is used to obtain tissue state updates from the optimized feedback sequence, and to perform real-time calibration of tactile perception using the information processing stage to obtain the calibrated perception model. The feedback output module generates tactile feedback output based on the calibrated perception model and force change data. If the output matches the operation behavior, the final dynamic matching result is determined.

9. An electronic device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the tactile feedback-based urological surgical simulation training method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed, implements the urological surgery simulation training method based on tactile feedback as described in any one of claims 1 to 7.

Images