A handheld surgical robotic system and control method

Abstract

The application provides a handheld surgical robot system and a control method. The system comprises: a preoperative planning module configured to generate a three-dimensional surgical plan according to preoperative image data of a patient, the three-dimensional surgical plan comprising a surgical path, a target pose and a safety boundary; an intraoperative registration module configured to register and map the three-dimensional surgical plan generated by the preoperative planning module with an actual anatomical space in surgery through a navigator and a tracer; and a handheld mechanical arm control module configured to plan a motion path of a mechanical arm based on the registered and mapped three-dimensional surgical plan, and control the mechanical arm to execute the motion path. In the application, preoperative planning and intraoperative registration are closed-loop integrated through a handheld surgical robot, which greatly improves the accuracy and safety of surgery while retaining the surgical intuition and adaptability of a surgeon.

Description

Technical Field

[0001] This application relates to the field of medical device technology, and more specifically, to a handheld surgical robot system and method. Background Technology

[0002] Traditional orthopedic surgical robotic systems typically employ bedside or floor-mounted robotic arms, with the surgeon controlling the robot from a console away from the surgical area. While this human-machine separation approach offers advantages in precision, it has several limitations: Firstly, the robotic arm system is bulky, requires extensive installation and setup, and occupies significant operating room space, making it particularly unsuitable for the intraoperative conditions in cold, rural hospitals. Secondly, the physical separation of the surgeon and patient restricts surgical intuition and on-the-spot responsiveness, as the surgeon cannot directly perceive the biomechanical interaction between the surgical instruments and bone tissue. Summary of the Invention

[0003] To address the aforementioned problems, the first aspect of this application provides a handheld surgical robot system, comprising: The preoperative planning module is used to generate a three-dimensional surgical plan based on the patient's preoperative imaging data. The three-dimensional surgical plan includes the surgical path, target pose, and safety boundaries. The intraoperative registration module is used to register and map the three-dimensional surgical plan generated by the preoperative planning module with the actual intraoperative anatomical space through a navigator and a tracer. The handheld robotic arm control module is used to plan the motion path of the robotic arm based on the registered and mapped 3D surgical plan, and control the robotic arm to execute the motion path.

[0004] A second aspect of this application provides a control method for a handheld surgical robot system, comprising: A three-dimensional surgical plan is generated based on the patient's preoperative imaging data. The three-dimensional surgical plan includes the surgical path, target pose, and safety boundaries. The 3D surgical plan generated by the preoperative planning module is registered and mapped with the actual intraoperative anatomical space using a navigator and a tracer. Based on the 3D surgical plan after registration and mapping, the motion path of the robotic arm is planned, and the robotic arm is controlled to execute the motion path.

[0005] A third aspect of this application provides an electronic device comprising: a memory and a processor; The memory is used to store programs; The processor, coupled to the memory, is used to execute the program to implement the handheld surgical robot system control method described above.

[0006] The fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, the program being executed by a processor to implement the handheld surgical robot system control method described above. Attached Figure Description

[0007] Figure 1 This is an architectural diagram of a handheld surgical robot system according to an embodiment of this application; Figure 2 This is an overall architecture diagram of a handheld surgical robot system according to an embodiment of this application; Figure 3 This is a schematic diagram of the trajectory correction module of a handheld surgical robot system according to an embodiment of this application; Figure 4 This is a schematic diagram of a handheld surgical robot system according to an embodiment of this application; Figure 5 This is a schematic diagram of the tactile feedback module of a handheld surgical robot system according to an embodiment of this application; Figure 6 This is a flowchart of a handheld surgical robot system control method according to an embodiment of this application; Figure 7 This is an architectural diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0008] To make the above-mentioned objects, features, and advantages of this application more apparent and understandable, specific embodiments of this application will be described in detail below with reference to the accompanying drawings. Although exemplary embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0009] It should be noted that, unless otherwise stated, the technical or scientific terms used in this application shall have the ordinary meaning as understood by one of ordinary skill in the art to which this application pertains.

[0010] For ease of understanding, the following terms may be used and are explained below: This application provides a handheld surgical robot system, the specific solution of which is as follows: Figures 1-5 As shown.

[0011] Combination Figure 1 , Figure 4 The diagram shown is an architectural representation of a handheld surgical robot system according to an embodiment of this application; wherein the handheld surgical robot system includes: The preoperative planning module is used to generate a three-dimensional surgical plan based on the patient's preoperative imaging data. The three-dimensional surgical plan includes the surgical path, target pose, and safety boundaries. The intraoperative registration module is used to register and map the three-dimensional surgical plan generated by the preoperative planning module with the actual intraoperative anatomical space through a navigator and a tracer. The handheld robotic arm control module is used to plan the motion path of the robotic arm based on the registered and mapped 3D surgical plan, and control the robotic arm to execute the motion path.

[0012] In this application, a handheld surgical robot is used to integrate preoperative planning and intraoperative registration in a closed loop, which greatly improves the accuracy and safety of the surgery while preserving the surgeon's surgical intuition and adaptability.

[0013] People living in cold / high-latitude environments for extended periods are prone to bone abnormalities such as localized bone loss and relatively thickened but more brittle cortex due to prolonged low temperatures and insufficient sunlight. Specifically, this manifests as harder hard areas and more porous areas within the same bone, leading to significant fluctuations in resistance during drilling or osteotomy. Furthermore, in cold regions, low temperatures induce vasoconstriction, resulting in poor bone microcirculation. Combined with the increased bone fragility associated with low temperatures, this makes occult bone fractures or drill bit edge breakage more likely.

[0014] Preferably, combined with Figure 2 As shown, the preoperative planning module is used to generate a three-dimensional surgical plan based on the patient's preoperative imaging data and cold-region-specific bone data. The three-dimensional surgical plan includes the surgical path, target pose, and safety boundary. The intraoperative registration module is used to register and map the three-dimensional surgical plan generated by the preoperative planning module with the actual intraoperative anatomical space through a navigator and a tracer. The handheld robotic arm control module is used to plan the motion path of the robotic arm based on the registered and mapped 3D surgical plan and cold-region-specific bone data, and to control the robotic arm to execute the motion path.

[0015] In this application, a handheld surgical robot is used to integrate preoperative planning and intraoperative registration in a closed loop. Combined with cold-region-specific bone data, the accuracy and safety of the surgery are greatly improved while preserving the surgeon's surgical intuition and adaptability.

[0016] In one implementation, combined with Figure 2 As shown, the handheld robotic arm control module includes: The motion recognition module is used to collect the surgeon's hand motion signals in real time, and decompose the motion signals into intentional operation motion components and unintentional physiological tremor components, and extract the intentional operation motion components of the surgeon as effective surgical movement commands; The trajectory correction module is used to compare the effective surgical movement command with the planned path mapped by the intraoperative registration module in real time, calculate the deviation vector and generate a correction torque to guide the surgeon's hand to move along the planned path. The tactile feedback module is used to collect the interaction force information between the end effector and bone tissue, and convert the interaction force information into force feedback and vibration tactile feedback acting on the surgeon's hand.

[0017] In this application, a handheld surgical robot is used to integrate preoperative planning, intraoperative registration, motion recognition, trajectory correction, and tactile feedback in a closed loop. Combined with cold-region-specific bone data and microcrack-protective motion strategies, it eliminates physiological tremors, corrects trajectory deviations, and provides real-time perception of bone tissue biomechanical state while preserving the surgeon's surgical intuition and adaptability. It is especially suitable for orthopedic surgery on patients with osteoporosis in cold regions, greatly improving the precision and safety of the surgery.

[0018] In this application, the core concept of the handheld surgical robot is not to replace the surgeon's hand, but to enhance it. The surgeon maintains traditional handheld operation habits, while the handheld robot provides real-time enhancements in tremor elimination, trajectory correction, and tactile feedback during the surgeon's operation, forming a complete closed loop of preoperative planning → intraoperative registration → motion recognition → trajectory correction → tactile feedback.

[0019] In this application, the introduction of cold-region-specific bone data enables the system to be adapted to the bone characteristics of patients who have long been exposed to cold / high-latitude environments. People living in cold environments for extended periods are prone to localized bone loss and increased accumulation of microcracks due to low temperatures and insufficient sunlight, making them more susceptible to occult bone fractures or drill bit edge breakage during surgery. This application effectively reduces additional damage to the fragile bone tissue of patients in cold-region environments during surgery by incorporating cold-region-specific bone data in the preoperative planning stage and employing a microcrack-protective movement strategy during the intraoperative execution stage.

[0020] In one embodiment, the motion recognition module includes: The signal acquisition unit is used to acquire the acceleration and angular velocity data of the surgeon's hand, as well as the gripping force information applied by the surgeon to the handle; The tremor separation unit is used to perform frequency domain analysis on the motion signal acquired by the signal acquisition unit and extract the intentional operation motion component. The motion intention prediction unit is used to perform time-series analysis on the historical trajectory of the intentional operation motion component based on a deep learning model, and to predict the surgeon's motion intention and trajectory trend in the short term.

[0021] The signal acquisition unit is used to collect acceleration and angular velocity data of the surgeon's hand via the inertial measurement unit (IMU), and simultaneously collect information on the gripping force applied by the surgeon to the handle via a multi-dimensional force / torque sensor. The IMU typically includes a three-axis accelerometer and a three-axis gyroscope, capable of acquiring the six-degree-of-freedom motion state of the surgeon's hand at a high sampling rate (e.g., above 1000Hz). The multi-dimensional force / torque sensor is installed between the handle gripping area and the end effector, capable of simultaneously measuring force and torque in three directions.

[0022] The tremor separation unit performs frequency domain analysis on the motion signals acquired by the signal acquisition unit. It uses an adaptive filtering algorithm to identify and separate physiological tremor components within a preset frequency band, retaining low-frequency intentional manipulation components. Human physiological tremors typically manifest as high-frequency, low-amplitude vibrations in the 8-12Hz frequency band, belonging to involuntary micro-contractions of muscles. The adaptive filtering algorithm dynamically adjusts the filtering parameters based on changes in the spectral characteristics of the surgeon's hand tremors in cold environments. In low-temperature environments, due to increased muscle tension and microvascular constriction caused by cold stimulation, the frequency and amplitude characteristics of the surgeon's hand physiological tremors change. For example, the peak frequency of the tremor may shift from approximately 10Hz in normal conditions to 8-9Hz, and the tremor amplitude may increase. The adaptive filtering algorithm dynamically adjusts the center frequency and bandwidth of the band-stop filter by real-time monitoring of the peak frequency and energy distribution of the tremor spectrum, ensuring effective separation of tremor components under different temperature conditions.

[0023] The motion intention prediction unit performs temporal analysis on the historical trajectories of the intentional motion components based on a deep learning model, predicting the surgeon's motion intentions and trajectory trends in the near future, and outputting the prediction results to the trajectory correction module. The deep learning model can employ temporal prediction architectures such as Long Short-Term Memory (LSTM) or Temporal Convolutional Network (TCN), continuously analyzing recent intentional motion trajectory data using a sliding window approach to predict the surgeon's hand movement trends within the next 50-200 milliseconds. By outputting the prediction results to the trajectory correction module in advance, the correction action can begin before the actual deviation occurs, achieving predictive correction. Compared to passive correction methods, this significantly reduces the abrupt change in correction torque, improving the surgeon's operational comfort.

[0024] In one implementation, combined with Figure 3 As shown, the trajectory correction module includes: The deviation calculation unit is used to compare the effective surgical movement command with the mapped planned path in real time and output the position deviation vector and the attitude deviation vector. The correction torque generation unit is used to generate correction torque based on the multi-axis micro servo drive array of the deviation vector; wherein, when the deviation is within a small deviation range, it outputs a compliant guiding force, when the deviation tends to a large deviation range, it gradually increases the constraint force, and when the deviation reaches the safety boundary, it outputs a rigid constraint force to prevent out-of-bounds operation; at the same time, during the surgical execution, it automatically executes segmented feed, pulse propulsion and drill retraction unloading operations based on the microcrack protection motion strategy parameters determined by cold-region specific bone data.

[0025] The deviation calculation unit is used to compare the effective surgical motion command with the mapped planned path in real time under a unified coordinate system, and output the position deviation vector and the attitude deviation vector. The unified coordinate system is the world coordinate system established by the intraoperative registration module. The deviation calculation unit continuously calculates the six-degree-of-freedom deviation between the current pose of the end effector and the corresponding target pose on the planned path at a high frequency (e.g., above 500Hz).

[0026] A correction torque generation unit is used to generate correction torque based on the deviation vector driving the multi-axis micro servo drive array. This unit employs a progressive impedance control strategy: when the deviation is within a small range (e.g., 0-1mm), it outputs a compliant guiding force, guiding the surgeon along the optimal path with almost no awareness; as the deviation approaches a large range (e.g., 1-3mm), it progressively increases the constraint force and superimposes a vibration tactile warning to alert the surgeon to the deviation; when the deviation reaches a safe boundary (e.g., above 3mm), it outputs a rigid constraint force to prevent out-of-bounds operation. Simultaneously, during surgical execution, segmented feed, pulse propulsion, and drill retraction / unloading operations are automatically executed based on microcrack protection motion strategy parameters determined from cold-region-specific bone data. The multi-axis micro servo drive array, composed of multiple micro brushless DC motors and planetary reducers, is installed inside the handheld robot handle, providing multi-degree-of-freedom correction torque output without significantly increasing the handle's weight and size.

[0027] In one implementation, combined with Figure 3 As shown, the trajectory correction module also includes a surgeon style adaptive unit, which is used to dynamically adjust the correction stiffness coefficient by continuously learning the current surgeon's operating habits and force preferences, so that the human-computer interaction characteristics gradually adapt to the individual operating styles of different surgeons.

[0028] Different practitioners exhibit individual differences in their operational force, speed, and habitual deviation direction. The practitioner style adaptive unit continuously estimates the current practitioner's operational characteristic parameters through an online learning algorithm (such as recursive least squares algorithm), and dynamically adjusts the stiffness, damping, and inertia parameters of the correction torque accordingly. For example, for practitioners with greater operational force, the correction stiffness is appropriately reduced to avoid resistance, while for practitioners with faster operational speed, the damping parameter is appropriately increased to provide more stable guidance.

[0029] In one implementation, combined with Figure 5 As shown, the haptic feedback module includes: The force feedback unit continuously collects the interaction force between the end effector and bone tissue through a multi-dimensional force / torque sensor, and converts the interaction force into motion force feedback acting on the surgeon's hand through a multi-axis micro servo drive array; The vibration tactile feedback unit is used to transmit tactile information to the surgeon's fingertips, enabling the surgeon to perceive changes in bone hardness, the feeling of drill penetration, and tissue boundaries. The bone tissue impedance sensing unit is used to combine the interaction force information with the preoperatively planned three-dimensional distribution data of bone density to calculate the mechanical impedance of the bone tissue in the current contact area in real time, and automatically trigger braking and issue an alarm when an impedance mutation is detected indicating that the bone cortex has penetrated or entered an abnormal tissue area.

[0030] The force feedback unit continuously collects the interaction force between the end effector and bone tissue using a multi-dimensional force / torque sensor, and converts this interaction force into kinematic force feedback acting on the surgeon's hand via a multi-axis micro servo drive array. This force feedback employs impedance mapping to convert the end effector interaction force into a sense of resistance acting on the surgeon's hand according to a preset proportional relationship, enabling the surgeon to perceive the hardness and resistance characteristics of the current bone tissue through kinematic perception.

[0031] The vibration-tactile feedback unit transmits tactile information to the surgeon's fingertips via a miniature vibration-tactile actuator array embedded in the handle, enabling the surgeon to perceive changes in bone hardness, drill penetration, and tissue boundaries. The miniature vibration-tactile actuator array, which can employ linear resonant actuators (LRAs) or piezoelectric actuators, is distributed at multiple locations within the handle's grip area and can generate vibration patterns of different frequencies and amplitudes. For example, a short, high-frequency vibration indicates contact with the cortical bone, a gradually increasing low-frequency vibration indicates entry into the cancellous bone region, and a sudden, strong vibration indicates a warning of cortical bone penetration.

[0032] The bone tissue impedance sensing unit combines the interaction force information with the preoperatively planned three-dimensional bone density distribution data to calculate the mechanical impedance of the bone tissue in the current contact area in real time. This mechanical impedance reflects the comprehensive resistance characteristics of bone tissue to the feeding motion of the end effector, including elastic and viscous impedance components. By comparing the real-time measured interaction force information with the expected impedance range in the preoperative three-dimensional bone density distribution data, impedance abrupt events can be detected. When an impedance abrupt event indicating cortical bone penetration (a sharp drop in impedance) or entry into an abnormal tissue area (impedance abnormally deviating from the expected range) is detected, braking is automatically triggered and an alarm is issued, forming the last line of defense for surgical safety.

[0033] In one embodiment, the cold-region-specific bone data includes at least two of the following: Epidemiological data on osteoporosis in cold regions include age-sex-bone mineral density distribution models based on large-scale population surveys in cold regions and the incidence of fragility fractures in various parts of the body in cold regions. Cold regions typically refer to high-latitude areas with an average annual temperature below 0°C or winters with prolonged periods of low temperatures exceeding 5 months. Statistical modeling was performed on large-scale bone mineral density screening data from populations in cold regions to establish reference curves for bone mineral density distribution across different age groups and sexes. The incidence of fragility fractures in various parts of the body (hip, vertebrae, distal radius, etc.) was also statistically analyzed, forming a bone risk assessment benchmark specific to cold regions.

[0034] Data on vitamin D deficiency and bone metabolism in cold regions, including the distribution characteristics of serum 25(OH)D levels and seasonal bone loss patterns in high-latitude areas. Shorter daylight hours and lower ultraviolet radiation in winter in high-latitude regions lead to insufficient vitamin D synthesis in the skin, resulting in generally low serum 25(OH)D levels. Vitamin D deficiency can cause secondary hyperparathyroidism, accelerating bone resorption and leading to seasonal bone loss. This data is helpful in assessing a patient's current bone metabolic status and determining the timing of surgery.

[0035] Bone microstructural characteristics data include thinning of trabeculae, increased intertrabecular spacing, and cumulative microcrack density parameters in patients from cold-climate regions. Due to long-term bone metabolic disorders, patients from cold-climate regions exhibit characteristic changes in bone microstructure, manifested as a decrease in the number of trabeculae, increased intertrabecular spacing, decreased trabecular connectivity, and increased cumulative microcrack density. The cumulative microcrack density parameter is particularly important, as it directly affects the bone tissue's tolerance to mechanical loads during surgery and is a key input for determining parameters for microcrack-protective movement strategies.

[0036] Intraoperative tissue characteristic data in a hypothermic environment includes temperature correction factors for the elastic modulus and hardness of the patient's bone tissue under hypothermic conditions. The mechanical properties of bone tissue change under hypothermic conditions, typically manifesting as increased elastic modulus and increased brittleness. These temperature correction factors are used to adjust the temperature environment of the three-dimensional bone density distribution data established during the preoperative planning phase, ensuring that intraoperative bone impedance sensing and microcrack-protective motion strategies accurately adapt to the actual intraoperative bone tissue mechanical state.

[0037] In one embodiment, the microcrack-protected motion strategy includes: The segmented feeding strategy divides the surgical execution path into multiple execution segments based on the patient's cumulative bone microcrack density parameters. After each segment is completed, the pre-set time is actively paused to release the accumulated stress in the bone tissue. The pulse propulsion strategy uses intermittent pulse force output to replace continuous constant force propulsion for the feed motion of the end effector, thereby reducing continuous load damage to the trabecular structure. The drill retraction and unloading strategy periodically withdraws the end effector at preset depth intervals to release the deformation and thermal stress accumulated in the bone tissue around the borehole. The withdrawal depth and frequency are automatically adjusted based on the cumulative density parameter of bone microcracks and the risk stratification level of bone in cold regions.

[0038] The segmented feeding strategy divides the surgical path into multiple segments based on the patient's cumulative bone microcrack density parameters. After each segment is completed, a preset pause time is initiated to release accumulated stress in the bone tissue. In areas with high cumulative bone microcrack density in patients from cold climates, the segment length should be shortened and the pause time extended to fully release local stress concentration and prevent microcracks from expanding into macroscopic fractures. For example, in high-density areas of bone microcracks, the segment length can be set to 1-2 mm and the pause time to 0.5-1 second; in low-density areas of bone microcracks, the segment length can be set to 3-5 mm and the pause time can be shortened to 0.2-0.3 seconds.

[0039] The pulse propulsion strategy replaces continuous constant force propulsion with intermittent pulse force output for the end effector's feed motion, reducing continuous load damage to the trabecular bone structure. This pulse propulsion strategy uses a high-frequency, low-amplitude pulse force mode (e.g., pulse frequency 20-50Hz, duty cycle 30%-50%) to allow bone tissue a small elastic recovery time within each pulse interval, avoiding continuous trabecular bone fracture and bone debris accumulation caused by continuous constant force propulsion.

[0040] The drill retraction and unloading strategy periodically withdraws the end effector at preset depth intervals to release accumulated deformation and thermal stress in the bone tissue surrounding the borehole. During drilling or pin placement, friction between the drill bit and bone tissue generates a localized thermal effect, which, combined with mechanical stress, can lead to thermal necrosis of bone tissue and microcrack propagation. The drill retraction and unloading strategy allows the bone tissue surrounding the borehole stress release and heat dissipation time by periodically withdrawing the drill bit (e.g., withdrawing 0.5-1 mm for every 2-3 mm of feed). The withdrawal depth and frequency are automatically adjusted based on the cumulative density parameter of bone microcracks and the risk stratification level of bone lesions in cold regions. For high-risk patients in cold regions, the withdrawal frequency and depth are increased to provide more adequate stress release protection.

[0041] In one implementation, combined with Figure 2 As shown, the handheld surgical robot system further includes: The intelligent assessment module is used to conduct targeted disease assessments on patients based on their medical data and cold-region-specific data, generating personalized auxiliary assessment results for each patient. The auxiliary rehabilitation module is used to generate personalized preoperative rehabilitation plans for patients based on their medical data, surgical plans, and cold-region-specific data; and to generate personalized postoperative rehabilitation plans for patients based on their medical data, postoperative assessment results, and cold-region-specific data.

[0042] In this application, the core of the intelligent assessment module is an assessment model based on a multimodal deep convolutional neural network. This model can comprehensively process imaging data and structured clinical data to output personalized disease risk assessment results.

[0043] Model Architecture: This convolutional neural network model employs a multi-branch fusion architecture. The first branch is the image processing branch, which uses a 3D convolutional neural network (3D-CNN) to process the patient's CT image data. This 3D-CNN uses a 3D extended version of ResNet-50 as its backbone network, and the input is standardized CT image volume data (size 128×128×64 voxels). It contains four residual convolutional blocks, each containing two 3×3×3 3D convolutional layers, a batch normalization layer, and a ReLU activation function, and residual learning is achieved through skip connections. Each residual convolutional block is followed by a 3D max pooling layer to achieve spatial downsampling. After global average pooling, the image branch outputs a 512-dimensional image feature vector. The second branch is the clinical data processing branch, which uses a fully connected network to process structured clinical and cold-region-specific data. The input is a standardized numerical feature vector (containing quantitative indicators from the aforementioned basic medical data and cold-region-specific data, with a dimension of approximately 30-50). This vector is passed through three fully connected layers (128, 256, and 128 nodes respectively), each followed by batch normalization, ReLU activation, and Dropout (with a dropout rate of 0.3), outputting a 128-dimensional clinical feature vector. The output feature vectors from the two branches are concatenated in a fusion layer to generate a 640-dimensional fused feature vector. This fused feature vector is then passed through two fully connected layers (256 and 128 nodes respectively) and a Softmax output layer to output multi-task assessment results, including osteoporosis risk grading (normal, osteopenia, osteoporosis, and severe osteoporosis), intraoperative fracture risk assessment (low, intermediate, and high risk), cold-region bone risk stratification (grades I to IV), and expected healing period assessment.

[0044] Training Process: The model training employs a multi-task learning strategy, with the total loss function being the weighted sum of the losses from each task. Classification tasks (osteoporosis risk grading, intraoperative fracture risk assessment, and cold-region bone risk stratification) use the cross-entropy loss function, while regression tasks (expected healing period assessment) use the smooth L1 loss function. The weights of each task's loss are automatically learned and determined using an uncertainty weighting method. Training utilizes the Adam optimizer with an initial learning rate of 1×10⁻. 4 A cosine annealing learning rate scheduling strategy was employed, with a batch size of 16, and training lasted for 200 epochs. Data augmentation strategies included random rotation (±15 degrees), random translation (±10 voxels), random scaling (0.9-1.1x), and random elastic deformation. The training, validation, and test sets were divided in a 7:1.5:1.5 ratio to ensure that data from each center was representative in each subset. The model's optimal performance on the validation set was used as the final model parameters.

[0045] In this application, the specific implementation of the personalized rehabilitation plan is as follows: A rehabilitation plan generation model based on a multi-layer convolutional neural network is embedded in the auxiliary rehabilitation module. The architecture of this model is an encoder-decoder structure. (Taking the preoperative rehabilitation plan model as an example) The encoder part receives multi-dimensional input data from the patient, including personalized assessment results output by the intelligent assessment module (osteoporosis risk level, intraoperative fracture risk level, cold-region bone risk stratification level, expected healing period), patient basic medical data vector, surgical planning plan feature vector (including surgical type encoding, surgical site encoding, estimated operation time, implant type encoding, etc.), and cold-region-specific data vector. The encoder consists of 4 one-dimensional convolutional layers with a kernel size of 3, and the number of channels in each layer are 64, 128, 256, and 512, respectively. Each layer is followed by a batch normalization layer, a GELU activation function, and residual connections. The high-dimensional features output by the encoder are obtained as 512-dimensional encoded features after global average pooling.

[0046] The decoder employs a combination of multi-head attention mechanisms and fully connected layers to output structured rehabilitation protocol parameters. The decoder consists of two multi-head attention layers (8 attention heads, 256 hidden dimensions), followed by three fully connected layers (256, 128 nodes, and output dimension, respectively). The final output layer uses different activation functions based on the type of each output parameter (e.g., Sigmoid for intensity coefficients in the 0-1 range, and Softmax for classification selection).

[0047] The training of this rehabilitation program generation model involves inputting multi-dimensional preoperative data of the patient and outputting the actual preoperative rehabilitation program received by the patient (a program developed by a team of rehabilitation medicine experts based on guidelines and clinical experience and validated for efficacy). Training employs a hybrid loss function: mean squared error loss is used for continuous output parameters (such as dosage and duration); cross-entropy loss is used for categorical output parameters (such as exercise type recommendations). The optimizer uses AdamW with an initial learning rate of 5×10⁻⁻⁻⁶. 4 A combined learning rate scheduling strategy of linear preheating and cosine annealing was employed, with training for 150 epochs and a batch size of 32. Model performance was evaluated using 5-fold cross-validation, and the model parameters with the best average validation performance were selected.

[0048] In one implementation, the surgical planning module is specifically used for: A three-dimensional skeletal model is constructed based on the patient's medical images; Based on a 3D skeletal model, targeted surgical planning is performed for patients, generating personalized 3D surgical plans. Acquire cold-region-specific data of the patient, which is generated based on the patient's medical images; The three-dimensional surgical plan was revised based on cold-region-specific data.

[0049] The modified surgical planning scheme described in this application may include mechanical modifications, such as modifications to ensure that the screw placement channel avoids areas of cancellous bone, or modifications to ensure that the screw placement channel penetrates areas of cancellous bone.

[0050] In one implementation, in constructing a three-dimensional skeletal model based on a patient's medical images, the skeletal medical images are segmented using a segmentation model, and then a three-dimensional skeletal model is constructed based on the bone edges in the segmentation results.

[0051] In one implementation, the segmentation process of the segmentation model includes: The image to be segmented is downsampled sequentially to obtain downsampled images at multiple levels; Upsampling based on multi-scale extraction and multi-attention extraction is performed sequentially on downsampled images at multiple levels to obtain upsampled images at multiple levels. The highest-level upsampled image is processed to obtain the segmentation result.

[0052] Specifically, in the segmentation process: The image to be segmented is downsampled sequentially to obtain the first downsampled image, the second downsampled image, the third downsampled image, the fourth downsampled image, and the fifth downsampled image. Multi-scale extraction processing is performed on the fifth downsampled image to obtain the fifth upsampled image; Multi-attention extraction is performed on the first downsampled image, the second downsampled image, the third downsampled image, and the fourth downsampled image respectively to obtain the first attention map, the second attention map, the third attention map, and the fourth attention map; Upsampling is performed on the fifth upsampled image, the fourth attention image, the third attention image, the second attention image, and the first attention image to obtain the fourth upsampled image, the third upsampled image, the second upsampled image, and the first upsampled image in sequence. The first upsampled image, after convolution processing, becomes the segmentation result.

[0053] The technical details of upsampling are as follows: The fifth upsampled image and the fourth attention image are upsampled to obtain the fourth upsampled image.

[0054] The upsampling and downsampling processes can be referred to existing similar processes and will not be described in detail in this application.

[0055] The process of multi-attention extraction is as follows: The input feature map is divided into three branches, and convolutions of different sizes are performed to extract the feature maps. The feature maps of the three branches are then concatenated to obtain the concatenated feature map. The spliced ​​feature map is divided into two branches; Within the first branch, the concatenated feature map is convolved and positional information is added. Then, it is multiplied with the original concatenated feature map to obtain the multiplied feature map. The multiplied feature map is added to the original concatenated feature map and then convolved to obtain the branch feature map of the first branch. In the second branch, the concatenated feature map is convolved and channel features are added. Then it is multiplied with the original concatenated feature map to obtain the multiplied feature map. The multiplied feature map is added to the original concatenated feature map and then convolved to obtain the branch feature map of the second branch. By concatenating the branch feature maps of the first and second branches, the output feature map of multi-attention extraction is obtained.

[0056] In this way, by extracting positional attention features and channel attention features through branching, more contextual information can be captured using positional and channel attention at different scales, and the importance of each channel can be selectively weighted to produce the best output characteristics.

[0057] The multi-scale extraction process is as follows: The input feature map is subjected to convolution, normalization, and activation processing to obtain the first convolutional map; The first convolutional image is convolved, normalized, and activated to obtain the second convolutional image. The input feature map is subjected to feature extraction in multiple branches to obtain the corresponding feature maps; the convolution kernel of each branch is different. After concatenating and merging the feature maps of multiple branches, a concatenated convolutional map is obtained. After combining the concatenated convolutional map and the input feature map, multi-head attention, normalization, and multi-layer perception processing are performed to obtain a multi-layer perception map. By combining the second convolutional map and the multilayer perceptron map, we obtain the output feature map extracted at multiple scales.

[0058] In this way, the context extracted by larger convolutional kernels is integrated with deeper information flow, and multi-scale features are formed by integrating convolutional kernels of different depths and sizes. Multi-head attention is then used to fuse multi-scale features, thereby achieving further integration of features.

[0059] In this application, it should be noted that if inconsistent sizes occur during the specific feature extraction process, they can be unified by reshaping. The specific location for this reshaping can be determined based on the actual situation, and will not be elaborated upon in this application.

[0060] In this application, the training process of the above-mentioned segmentation model involves obtaining training samples, which include the input image to be segmented and the labeled segmentation results; inputting the image to be segmented from the training samples into the segmentation model to obtain a predicted segmentation result; calculating a loss function, namely the DiceLoss function, based on the predicted segmentation result and the sample segmentation result; and iterating the parameters of the entire segmentation model based on the loss function until the loss function converges. During training, the learning rate ranges from 1e-4 to 1e-3, the weight decay is 1e-4, and the total number of training epochs is 200-500.

[0061] It should be noted that, unless otherwise specified, the personalized processing model in this application can be obtained by targeted fine-tuning of an existing large model. The specific fine-tuning process may include: acquiring the patient's multimodal features and output information as sample data for the model based on its input and output requirements; modifying the input and output layers of the pre-trained large model to match its input and output; adding a low-rank adapter module next to the model's key layer (attention mechanism) so that only these few new parameters are trained during training; training the large model based on the sample data and updating the parameters within the low-rank adapter module and the modified input and output layer parameters until the loss converges. Further details can be found in the training requirements of existing models and will not be elaborated upon in this application.

[0062] This application provides a navigation and positioning method for the handheld surgical robot system described above. The specific solution of this method is as follows: Figure 6 As shown below, the control method of the handheld surgical robot system will be described in detail.

[0063] Combination Figure 6 As shown, the control method for the handheld surgical robot system includes: S101, Generate a three-dimensional surgical plan based on the patient's preoperative image data. The three-dimensional surgical plan includes the surgical path, target pose, and safety boundary. S102, the three-dimensional surgical plan generated by the preoperative planning module is registered and mapped with the actual intraoperative anatomical space through the navigator and tracer; S103, based on the 3D surgical plan after registration and mapping, plans the motion path of the robotic arm and controls the robotic arm to execute the motion path.

[0064] Preferably, it includes: S101, Generate a three-dimensional surgical plan based on the patient's preoperative imaging data and cold-region-specific bone data. The three-dimensional surgical plan includes the surgical path, target pose, and safety boundary. S102, the three-dimensional surgical plan generated by the preoperative planning module is registered and mapped with the actual intraoperative anatomical space through the navigator and tracer; S103, based on the registered and mapped 3D surgical plan and cold-region-specific bone data, plans the motion path of the robotic arm and controls the robotic arm to execute the motion path.

[0065] In one implementation, S103, based on the registered and mapped three-dimensional surgical plan and cold-region-specific bone data, the motion path of the robotic arm is planned, and the robotic arm is controlled to execute the motion path, including: The surgeon's hand movement signals are collected in real time and decomposed into intentional operation movement components and unintentional physiological tremor components. The intentional operation movement components of the surgeon are extracted as effective surgical movement commands. The effective surgical movement command is compared with the planned path mapped by the intraoperative registration module in real time, the deviation vector is calculated and the correction torque is generated to guide the surgeon's hand to move along the planned path. The interaction force information between the end effector and bone tissue is collected, and the interaction force information is converted into force feedback and vibration tactile feedback acting on the surgeon's hand.

[0066] In one implementation, the surgeon's hand movement signals are acquired in real time, and the movement signals are decomposed into intentional operational movement components and unintentional physiological tremor components. The intentional operational movement components are extracted as effective surgical movement commands; including: Collect data on the acceleration and angular velocity of the surgeon's hand, and simultaneously collect information on the gripping force applied by the surgeon to the handle; The motion signal acquired by the signal acquisition unit is subjected to frequency domain analysis to extract the intentional operation motion component; Based on a deep learning model, the historical trajectory of the intentional movement components is analyzed over time to predict the surgeon's movement intentions and trajectory trends in the near future.

[0067] In one implementation, the effective surgical movement command is compared in real time with the planned path mapped by the intraoperative registration module, a deviation vector is calculated, and a corrective torque is generated to guide the surgeon's hand to move along the planned path; including: The effective surgical movement command is compared with the mapped planned path in real time, and the position deviation vector and attitude deviation vector are output. The deviation vector multi-axis micro servo drive array generates a correction torque; wherein, when the deviation is within a small deviation range, a compliant guiding force is output, the constraint force is progressively enhanced when the deviation tends to be within a large deviation range, and a rigid constraint force is output to prevent out-of-bounds operation when the deviation reaches the safety boundary; at the same time, during the surgical execution, segmented feed, pulse propulsion and drill retraction unloading operations are automatically executed based on the microcrack protection motion strategy parameters determined by cold-region specific bone data.

[0068] In one implementation, the interaction force information between the end effector and bone tissue is acquired, and the interaction force information is converted into force feedback and vibration tactile feedback acting on the surgeon's hand, including: The interaction force between the end effector and bone tissue is continuously collected by a multi-dimensional force / torque sensor, and the interaction force is converted into kinematic force feedback acting on the surgeon's hand by a multi-axis micro servo drive array. It transmits tactile information to the surgeon's fingertips, enabling the surgeon to perceive changes in bone hardness, the feeling of drill penetration, and tissue boundaries; By combining the interaction force information with the preoperatively planned three-dimensional distribution data of bone density, the mechanical impedance of bone tissue in the current contact area is calculated in real time. When an impedance mutation is detected indicating that the bone cortex has penetrated or entered an abnormal tissue area, braking is automatically triggered and an alarm is issued.

[0069] In one embodiment, the microcrack-protected motion strategy includes: The segmented feeding strategy divides the surgical execution path into multiple execution segments based on the patient's cumulative bone microcrack density parameters. After each segment is completed, the pre-set time is actively paused to release the accumulated stress in the bone tissue. The pulse propulsion strategy uses intermittent pulse force output to replace continuous constant force propulsion for the feed motion of the end effector, thereby reducing continuous load damage to the trabecular structure. The drill retraction and unloading strategy periodically withdraws the end effector at preset depth intervals to release the deformation and thermal stress accumulated in the bone tissue around the borehole. The withdrawal depth and frequency are automatically adjusted based on the cumulative density parameter of bone microcracks and the risk stratification level of bone in cold regions.

[0070] In one embodiment, the method further includes: Based on the patient's medical data and cold-region-specific data, a targeted disease assessment is conducted to generate personalized auxiliary assessment results for the patient. Based on the patient's medical data, surgical plan, and cold-region-specific data, a personalized preoperative rehabilitation plan is generated for the patient; and based on the patient's medical data, postoperative assessment results, and cold-region-specific data, a personalized postoperative rehabilitation plan is generated for the patient.

[0071] The handheld surgical robot system control method provided in the above embodiments of this application corresponds to the handheld surgical robot system provided in the embodiments of this application. Therefore, the specific content of the method corresponds to the handheld surgical robot system. The specific content can be referred to the records in the handheld surgical robot system, and will not be repeated in this application.

[0072] The handheld surgical robot system control method provided in the above embodiments of this application is based on the same inventive concept as the handheld surgical robot system provided in the embodiments of this application, and has the same beneficial effects as the methods adopted, run or implemented by the application stored therein.

[0073] Based on the same inventive concept, another embodiment of the present invention provides an electronic device for implementing the handheld surgical robot system control method described in the above embodiments. Figure 7 As shown, the electronic device includes a memory 301 and a processor 303.

[0074] Memory 301 can be configured to store a program.

[0075] Additionally, memory 301 can also be configured to store various other data to support operation on the electronic device. Examples of this data include instructions for any application or method used to operate on the electronic device, contact data, phonebook data, messages, pictures, videos, etc.

[0076] Memory 301 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. Processor 303, coupled to memory 301, is used to execute programs in memory 301 for: A three-dimensional surgical plan is generated based on the patient's preoperative imaging data. The three-dimensional surgical plan includes the surgical path, target pose, and safety boundaries. The 3D surgical plan generated by the preoperative planning module is registered and mapped with the actual intraoperative anatomical space using a navigator and a tracer. Based on the 3D surgical plan after registration and mapping, the motion path of the robotic arm is planned, and the robotic arm is controlled to execute the motion path.

[0077] In one implementation, the processor 303 is further configured to: The surgeon's hand movement signals are collected in real time and decomposed into intentional operation movement components and unintentional physiological tremor components. The intentional operation movement components of the surgeon are extracted as effective surgical movement commands. The effective surgical movement command is compared with the planned path mapped by the intraoperative registration module in real time, the deviation vector is calculated and the correction torque is generated to guide the surgeon's hand to move along the planned path. The interaction force information between the end effector and bone tissue is collected, and the interaction force information is converted into force feedback and vibration tactile feedback acting on the surgeon's hand.

[0078] In one implementation, the processor 303 is further configured to: Collect data on the acceleration and angular velocity of the surgeon's hand, and simultaneously collect information on the gripping force applied by the surgeon to the handle; The motion signal acquired by the signal acquisition unit is subjected to frequency domain analysis to extract the intentional operation motion component; Based on a deep learning model, the historical trajectory of the intentional movement components is analyzed over time to predict the surgeon's movement intentions and trajectory trends in the near future.

[0079] In one implementation, the processor 303 is further configured to: The effective surgical movement command is compared with the mapped planned path in real time, and the position deviation vector and attitude deviation vector are output. The deviation vector multi-axis micro servo drive array generates a correction torque; wherein, when the deviation is within a small deviation range, a compliant guiding force is output, the constraint force is progressively enhanced when the deviation tends to be within a large deviation range, and a rigid constraint force is output to prevent out-of-bounds operation when the deviation reaches the safety boundary; at the same time, during the surgical execution, segmented feed, pulse propulsion and drill retraction unloading operations are automatically executed based on the microcrack protection motion strategy parameters determined by cold-region specific bone data.

[0080] In one implementation, the processor 303 is further configured to: The interaction force between the end effector and bone tissue is continuously collected by a multi-dimensional force / torque sensor, and the interaction force is converted into kinematic force feedback acting on the surgeon's hand by a multi-axis micro servo drive array. It transmits tactile information to the surgeon's fingertips, enabling the surgeon to perceive changes in bone hardness, the feeling of drill penetration, and tissue boundaries; By combining the interaction force information with the preoperatively planned three-dimensional distribution data of bone density, the mechanical impedance of bone tissue in the current contact area is calculated in real time. When an impedance mutation is detected indicating that the bone cortex has penetrated or entered an abnormal tissue area, braking is automatically triggered and an alarm is issued.

[0081] In one embodiment, the microcrack-protected motion strategy includes: The segmented feeding strategy divides the surgical execution path into multiple execution segments based on the patient's cumulative bone microcrack density parameters. After each segment is completed, the pre-set time is actively paused to release the accumulated stress in the bone tissue. The pulse propulsion strategy uses intermittent pulse force output to replace continuous constant force propulsion for the feed motion of the end effector, thereby reducing continuous load damage to the trabecular structure. The drill retraction and unloading strategy periodically withdraws the end effector at preset depth intervals to release the deformation and thermal stress accumulated in the bone tissue around the borehole. The withdrawal depth and frequency are automatically adjusted based on the cumulative density parameter of bone microcracks and the risk stratification level of bone in cold regions.

[0082] In one implementation, the processor 303 is further configured to: Based on the patient's medical data and cold-region-specific data, a targeted disease assessment is conducted to generate personalized auxiliary assessment results for the patient. Based on the patient's medical data, surgical plan, and cold-region-specific data, a personalized preoperative rehabilitation plan is generated for the patient; and based on the patient's medical data, postoperative assessment results, and cold-region-specific data, a personalized postoperative rehabilitation plan is generated for the patient.

[0083] In this application, Figure 7 The diagram only shows some components and does not mean that the electronic device includes only these components. Figure 7The components shown.

[0084] The electronic device provided in this embodiment is based on the same inventive concept as the force-optimized robotic arm osteotomy control method provided in the embodiments of this application, and has the same beneficial effects as the methods adopted, run or implemented by the application stored therein.

[0085] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0086] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0087] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0088] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory. Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0089] This application also provides a computer-readable storage medium corresponding to the force-optimized robotic arm osteotomy control method provided in the foregoing embodiments, wherein a computer program (i.e., a program product) is stored thereon. When the computer program is run by a processor, it executes the interactive image analysis assistance method for 3D aerial imaging provided in any of the foregoing embodiments.

[0090] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, 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, optical disc read-only memory (CDROM), digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0091] The computer-readable storage medium provided in the above embodiments of this application and the handheld surgical robot system control method provided in the embodiments of this application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application stored therein.

[0092] It should be noted that numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. In some instances, well-known structures and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0093] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, system, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, system, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, system, article, or apparatus that includes said element.

[0094] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A handheld surgical robot system, characterized in that, include: The preoperative planning module is used to generate a three-dimensional surgical plan based on the patient's preoperative imaging data. The three-dimensional surgical plan includes the surgical path, target pose, and safety boundaries. The intraoperative registration module is used to register and map the three-dimensional surgical plan generated by the preoperative planning module with the actual intraoperative anatomical space through a navigator and a tracer. The handheld robotic arm control module is used to plan the motion path of the robotic arm based on the registered and mapped 3D surgical plan, and control the robotic arm to execute the motion path.

2. The handheld surgical robot system according to claim 1, characterized in that, The handheld robotic arm control module includes: The motion recognition module is used to collect the surgeon's hand motion signals in real time, and decompose the motion signals into intentional operation motion components and unintentional physiological tremor components, and extract the intentional operation motion components of the surgeon as effective surgical movement commands; The trajectory correction module is used to compare the effective surgical movement command with the planned path mapped by the intraoperative registration module in real time, calculate the deviation vector and generate a correction torque to guide the surgeon's hand to move along the planned path. The tactile feedback module is used to collect the interaction force information between the end effector and bone tissue, and convert the interaction force information into force feedback and vibration tactile feedback acting on the surgeon's hand.

3. The handheld surgical robot system according to claim 2, characterized in that, The motion recognition module includes: The signal acquisition unit is used to acquire the acceleration and angular velocity data of the surgeon's hand, as well as the gripping force information applied by the surgeon to the handle; The tremor separation unit is used to perform frequency domain analysis on the motion signal acquired by the signal acquisition unit and extract the intentional operation motion component. The motion intention prediction unit is used to perform time-series analysis on the historical trajectory of the intentional operation motion component based on a deep learning model, and to predict the surgeon's motion intention and trajectory trend in the short term.

4. The handheld surgical robot system according to claim 2, characterized in that, The trajectory correction module includes: The deviation calculation unit is used to compare the effective surgical movement command with the mapped planned path in real time and output the position deviation vector and the attitude deviation vector. The correction torque generation unit is used to generate correction torque based on the multi-axis micro servo drive array of the deviation vector; wherein, when the deviation is within a small deviation range, it outputs a compliant guiding force, when the deviation tends to a large deviation range, it gradually increases the constraint force, and when the deviation reaches the safety boundary, it outputs a rigid constraint force to prevent out-of-bounds operation; at the same time, during the surgical execution, it automatically executes segmented feed, pulse propulsion and drill retraction unloading operations based on the microcrack protection motion strategy parameters determined by cold-region specific bone data.

5. The handheld surgical robot system according to claim 2, characterized in that, The haptic feedback module includes: The force feedback unit continuously collects the interaction force between the end effector and bone tissue through a multi-dimensional force / torque sensor, and converts the interaction force into motion force feedback acting on the surgeon's hand through a multi-axis micro servo drive array; The vibration tactile feedback unit is used to transmit tactile information to the surgeon's fingertips, enabling the surgeon to perceive changes in bone hardness, the feeling of drill penetration, and tissue boundaries. The bone tissue impedance sensing unit is used to combine the interaction force information with the preoperatively planned three-dimensional distribution data of bone density to calculate the mechanical impedance of the bone tissue in the current contact area in real time, and automatically trigger braking and issue an alarm when a sudden impedance change is detected.

6. The handheld surgical robot system according to claim 4, characterized in that, The microcrack-protected motion strategy includes: The segmented feeding strategy divides the surgical execution path into multiple execution segments based on the patient's cumulative bone microcrack density parameters. After each segment is completed, the pre-set time is actively paused to release the accumulated stress in the bone tissue. The pulse propulsion strategy uses intermittent pulse force output to replace continuous constant force propulsion for the feed motion of the end effector, thereby reducing continuous load damage to the trabecular structure. The drill retraction and unloading strategy periodically withdraws the end effector at preset depth intervals to release the deformation and thermal stress accumulated in the bone tissue around the borehole. The withdrawal depth and frequency are automatically adjusted based on the cumulative density parameter of bone microcracks and the risk stratification level of bone in cold regions.

7. The handheld surgical robot system according to claim 1, characterized in that, Also includes: The intelligent assessment module is used to conduct targeted disease assessments on patients based on their medical data and cold-region-specific data, generating personalized auxiliary assessment results for each patient. The auxiliary rehabilitation module is used to generate personalized preoperative rehabilitation plans for patients based on their medical data, surgical plans, and cold-region-specific data. Furthermore, based on the patient's medical data, postoperative assessment results, and cold-region-specific data, a personalized postoperative rehabilitation plan is generated for the patient.

8. A control method for a handheld surgical robot system, applied to the handheld surgical robot system according to any one of claims 1 to 7, characterized in that, Includes the following steps: A three-dimensional surgical plan is generated based on the patient's preoperative imaging data. The three-dimensional surgical plan includes the surgical path, target pose, and safety boundaries. The 3D surgical plan generated by the preoperative planning module is registered and mapped with the actual intraoperative anatomical space using a navigator and a tracer. Based on the 3D surgical plan after registration and mapping, the motion path of the robotic arm is planned, and the robotic arm is controlled to execute the motion path.

9. An electronic device, characterized in that, It includes a processor and a memory, wherein the memory stores computer-executable instructions, and the processor executes the computer-executable instructions to implement the steps of the control method of the handheld surgical robot system of claim 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, implement the steps of the control method for the handheld surgical robot system of claim 8.

Images