Intelligent control method and system of microwave radiation equipment
Patent Information
- Application Number
- CN202610949854.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-29
- Publication Date
- 2026-09-15
Smart Images
Figure CN122755680A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of automation control and robot path planning technology. More specifically, this invention relates to an intelligent control method and system for a microwave radiation device. Background Technology
[0002] In industrial and scientific research scenarios such as automated electromagnetic radiation processing, biomimetic thermal simulation, and large-area target surface heat treatment, microwave radiation equipment is often used to project discrete, multi-point spatial energy onto target objects. Taking the application of microwave radiation equipment in the medical industry as an example, due to the significant nonlinearity and individual differences in the electromagnetic absorption rate and thermal conduction dynamic response of different target media, in order to stably and quickly lock the surface thermal field of the target area within a preset isothermal range, it is usually necessary to perform multi-point, long-term cyclic radiation on the target area.
[0003] However, existing microwave radiation equipment control systems have the following significant drawbacks when dealing with large-area or multi-region target outputs:
[0004] (1) Inefficient timing and path planning: Traditional systems lack adaptive discrete point placement and multi-round timing planning mechanisms. If a single global path algorithm is used, when the total output time of the system is subject to a strict threshold constraint, it is easy to cause trajectory planning redundancy and frequent long-distance movement of equipment between different points, making it impossible to complete efficient coverage of the target area within a safe time.
[0005] (2) The cumulative error of the robotic arm is serious: When the multi-axis collaborative robotic arm carries the microwave radiation end to perform long-term, multi-point spatial roving movement, due to the influence of multi-axis linkage, mechanical clearance and temperature drift, the spatial positioning will produce obvious cumulative errors, making it difficult to guarantee the spatial normal alignment accuracy of the end effector at each discrete point.
[0006] (3) Lack of closed-loop control and safety intervention under thermal dynamic response: Traditional equipment mostly adopts a fixed power output mode, which is difficult to cope with transient thermal field anomalies. When the external environment or target state undergoes a sudden geometric change (such as micro-movement of the target object), the system cannot achieve millisecond-level power cut-off and trajectory retraction, which can easily lead to medium overheating and temperature control overshoot.
[0007] Therefore, how to solve the problems of low control efficiency and poor accuracy of current microwave radiation equipment is a technical problem that urgently needs to be solved in this field. Summary of the Invention
[0008] To address the aforementioned technical problems of low control efficiency and poor accuracy in microwave radiation equipment, the present invention provides solutions in the following aspects.
[0009] In a first aspect, the present invention provides an intelligent control method for a microwave radiation device, the method being executed by a control unit, comprising the following steps: S1: acquiring spatial point cloud data and infrared thermal maps of a target object, and performing data registration and fusion to construct a digital model including geometric structure and thermal field distribution and identify the target area; S2: calculating the total number of theoretical output sites required for full coverage based on the spatial geometric information of the target area, and calculating the total duration of a single round of full coverage based on the standard output duration of a single site and the average movement time between adjacent sites; S3: comparing the total duration of a single round of full coverage with a preset upper limit of total output time to determine the output round executed by the microwave radiation device, and allocating a corresponding number of calculation sites and output time to each output round so that the predicted total output time does not exceed the upper limit of total output time; S4: generating multi-round progressive motion trajectory planning data sequentially according to the allocated output rounds; wherein the trajectory planning of the next round takes the output end coordinate of the previous round as the starting point, and dynamically filters the target sites of the next round based on the real-time temperature feedback data after the output of the previous round, forming hierarchical control data.
[0010] Further, in step S2, the calculation of the total number of theoretical output points required for full coverage includes: extracting the surface area S of the target area; obtaining the effective coverage diameter d and overlap coefficient η of a single radiation point; and calculating the total number of theoretical output points required for full coverage. The formula for calculating the total number of theoretical output points is: N = S / (d(1 η)) 2 .
[0011] Furthermore, in step S3, the step of allocating the corresponding number of computing points to each output round specifically includes: calculating the single-round full coverage time T. single =N×t p +(N-1)×t m Where N is the total number of output sites, t p The duration of a single standard output, t m This represents the average movement time; if the output round number is 1 and T... single >T total When, according to the formula N1′= T total +t m / t p +t m Redetermine the number of output sites, T total This is the upper limit of the total output time; if the number of output rounds is equal to 2 or 3, it is calculated according to the preset attenuation factor k. i The maximum allocation for subsequent rounds is distributed in descending order, where the maximum allocation for the second round, N2, is... k2N The maximum quota for the third round is N3 = k3N And satisfy 1>k2>k3>0.
[0012] Furthermore, it also includes a timeout dynamic compression step: calculating the estimated time for each round, and when the calculated total time T... sum Exceeding the total output time limit T total At the same time, data is dynamically compressed according to the priority of canceling the third round, reducing the number of sites in the second round, and shortening the single-point output time, so as to ensure that the control sequence is completed within a safe time.
[0013] Furthermore, the motion trajectory planning data specifically includes a first round of full-coverage trajectory planning: generating coordinates of all candidate output points evenly distributed based on the three-dimensional model of the target area; calculating the shortest circumvention path using the global traveling salesman optimization algorithm to generate a first motion control trajectory that sequentially traverses all candidate output points; and recording the endpoint coordinates of the first motion control trajectory as the starting point of the next round of trajectory planning.
[0014] Furthermore, the motion trajectory planning data also includes a second round of focused trajectory planning: based on the real-time temperature feedback data after the first round of output, target areas or core target areas whose temperature has not reached the preset benchmark are automatically selected as high-priority sites, and the coordinates of sites whose temperature has reached the benchmark are removed; using the endpoint coordinates of the previous round as the starting point, the remaining high-priority sites are re-optimized locally, and a second motion control trajectory is generated according to the principle of proximity concentration.
[0015] Furthermore, the motion trajectory planning data also includes a third round of supplementary fine-tuning trajectory planning: thermal imaging data is collected again, and only the coordinates of supplementary measurement points in areas with insufficient temperature after the output of the first two rounds are retained; the nearest neighbor greedy algorithm is used to directly plan from the endpoint coordinates of the previous round to the coordinates of the nearest supplementary measurement point to generate the third motion control trajectory; the first, second, and third motion control trajectories are sequentially connected in time series and do not return to the initial standby coordinates in the middle, and the control parameters and end effector attitude are updated based on the output effect of the previous round in each round.
[0016] Furthermore, if the output rounds are greater than 3, the steps for dynamically calculating the number of computational sites allocated to subsequent rounds include: constructing a time boundary equation and inversely solving for the maximum allowable value of the relaxation factor, wherein the time boundary equation is:
[0017] Based on the solved relaxation factor, determine the number N of dynamic compression calculation points in the i-th round. i ′= N max ×k i ×α , where ki The initial convergence factors for each round are i=1,2,…,P, and k1=1, where α is the relaxation factor.
[0018] Furthermore, it also includes the step of performing adaptive temperature closed-loop control for each target site: acquiring the temperature data of each target site and comparing it with the preset target temperature; if the temperature rises steadily to the preset target temperature, maintaining or fine-tuning the microwave output power to stabilize the temperature within the range of ±0.5℃ of the preset target temperature; if the temperature rise rate exceeds the preset safety threshold, immediately reducing the microwave output power to slow down the heating rate; if the temperature exceeds the preset safety upper limit, immediately cutting off the microwave radiation equipment output and issuing an alarm.
[0019] In a second aspect, the present invention also provides an intelligent control system for a microwave radiation device, comprising: a multimodal sensing unit, which acquires spatial point cloud data and infrared thermal images of a target object, and performs data registration and fusion to construct a digital model including geometric structure and thermal field distribution and identify the target area; a control unit, which is connected to the multimodal sensing unit and executes the intelligent control method for the microwave radiation device as described in one or more of the foregoing embodiments to generate corresponding motion control commands and power adjustment commands; and a multi-axis execution unit, which is connected to the control unit and includes a multi-degree-of-freedom cooperative robotic arm and a microwave radiation generator installed at the end of the robotic arm, configured to drive the robotic arm to move to each output position and adjust the energy output of the microwave generator according to the motion control commands and power adjustment commands generated by the control unit.
[0020] The beneficial effects of this invention are as follows: By dynamically coupling the spatial geometric information of the target area with the rigid time upper limit constraint of the system in multiple dimensions, this invention realizes the reverse adaptive configuration of calculating time first and then determining the round. By taking the end point of the previous round as the starting point of the next round and dynamically selecting the site based on real-time temperature feedback, a multi-round progressive closed-loop control chain is constructed, which greatly improves the targeting accuracy of energy projection.
[0021] Furthermore, this scheme introduces a dynamic relaxation factor and boundary integral equations to provide a high-order general iterative algorithm that supports any P rounds (P>3) without modifying the underlying software. The control unit solves for the relaxation factor through a one-dimensional search, adaptively mapping and compressing the requirements of higher-order rounds into a finite time axis, achieving a high-precision control effect that gradually approximates and converges infinitely. Attached Figure Description
[0022] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent upon reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein: Figure 1 This is a flowchart illustrating an intelligent control method for a microwave radiation device according to an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the composition of an intelligent control system for a microwave radiation device according to an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating a digital model including geometry and thermal field distribution according to an embodiment of the present invention. Detailed Implementation
[0023] 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, not all, of the embodiments of the present invention. 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.
[0024] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
[0025] In the context of this invention, the intelligent control method for microwave radiation equipment can be applied to various application scenarios with hard time constraints, nonlinear feedback, and spatial discreteness. Therefore, the solution of this invention has wide applicability in medical, engineering, and other applications. For example, in rehabilitation therapy centers, hospital pain management departments, and orthopedic departments, microwave therapy devices can be used to perform deep hyperthermia on patients' human tissues (such as muscles and joints) to relieve inflammation and promote blood circulation. It is understood that the control method of this invention is not limited to this scenario. For example, in industrial precision thermal processing and laser surface modification (such as annealing and quenching), a laser head can be used to perform multiple rounds of laser / microwave surface heat treatment on large metal workpieces or complex curved surfaces.
[0026] Intelligent control method embodiment Figure 1 This is a flowchart illustrating an intelligent control method for a microwave radiation device according to an embodiment of the present invention.
[0027] like Figure 1 As shown, the intelligent control method for the microwave radiation device of the present invention mainly includes the following steps: S1: Acquire spatial point cloud data and infrared thermal map of the target object, and perform data registration and fusion to construct a digital model containing geometric structure and thermal field distribution and identify the target area.
[0028] Taking the control process of microwave physiotherapy equipment in a medical setting as an example, a ToF depth camera integrated into the end effector of a robotic arm or a stand-alone device is used to scan the patient's treatment area, generating a high-precision 3D point cloud model. By identifying preset skeletal landmarks, the three-dimensional coordinates of the target treatment area are calculated. An infrared thermal imager, spatially registered and time-synchronized with the depth camera, collects real-time temperature distribution data of the treatment area, generating a thermal map.
[0029] A fusion algorithm is used to map heatmaps onto a 3D model, automatically identifying target areas (temperature anomaly areas) and calculating their three-dimensional coordinates. Specifically, point cloud data and heatmap data can be aligned in a three-dimensional space through coordinate transformation to achieve data fusion.
[0030] S2: Calculate the total number of theoretical output points required for full coverage based on the spatial geometry information of the target area, and calculate the total duration of a single round of full coverage based on the standard output time of a single point and the average movement time between adjacent points. The calculation steps for the total number of theoretical output points required for full coverage include: extracting the surface area S of the target area; obtaining the effective coverage diameter d and overlap coefficient η of a single radiation point; and calculating the total number of theoretical output points required for full coverage. The formula for calculating the total number of theoretical output points is: N = S / (d(1 η)) 2 .
[0031] S3: Compare the total duration of a single full-coverage cycle with the preset upper limit of total output time to determine the output cycle executed by the microwave radiation device, and assign the corresponding number of calculation points and output time to each output cycle so that the predicted total output time does not exceed the upper limit of total output time.
[0032] The step of allocating the corresponding number of calculation points to each output round specifically includes different cases where the output round is 1, 2, or 3 times, or more than 3 times.
[0033] Specifically, calculate the single-round full coverage time T. single =N×t p +(N-1)×t m Where N is the total number of output sites, t p The duration of a single standard output, t m This represents the average travel time.
[0034] If the output round is 1 and T single >T totalWhen, according to the formula N1′= T total +t m / t p +t m Redetermine the number of output sites, T total This represents the maximum total output time.
[0035] If the output round is 2 or 3, the preset attenuation factor k will be applied. i The maximum allocation for subsequent rounds is distributed in descending order, where the maximum allocation for the second round, N2, is... k2N The maximum quota for the third round is N3 = k3N And satisfy 1>k2>k3>0.
[0036] If the output round is greater than 3, the steps for dynamically calculating the number of calculation points allocated to subsequent rounds include: Construct the time boundary equations and solve in reverse for the maximum allowable value of the relaxation factor. The time boundary equations are as follows:
[0037] Based on the solved relaxation factor, determine the number N of dynamic compression calculation points in the i-th round. i ′= N×k i ×α , where k i The initial convergence factors for each round are i=1,2,…,P, and k1=1, where α is the relaxation factor.
[0038] Furthermore, it also includes a timeout dynamic compression step: calculating the estimated time for each round, and when the calculated total time T... sum Exceeding the total output time limit T total At the same time, data is dynamically compressed according to the priority of canceling the third round, reducing the number of sites in the second round, and shortening the single-point output time, so as to ensure that the control sequence is completed within a safe time.
[0039] S4: Based on the assigned output rounds, generate multi-round progressive motion trajectory planning data in sequence; wherein, the trajectory planning of the next round takes the output end coordinates of the previous round as the starting point, and dynamically selects the target position of the next round based on the real-time temperature feedback data after the output of the previous round, forming hierarchical control data.
[0040] The aforementioned motion trajectory planning data specifically includes the first round of full-coverage trajectory planning: generating coordinates of all candidate output points evenly distributed based on the 3D model of the target area; calculating the shortest circumvention path using the global traveling salesman optimization algorithm to generate the first motion control trajectory that sequentially traverses all candidate output points; and recording the endpoint coordinates of the first motion control trajectory as the starting point for the next round of trajectory planning.
[0041] When there are 2 output rounds, the motion trajectory planning data also includes a second round of focused trajectory planning: based on the real-time temperature feedback data after the first round of output, the target area or core target area whose temperature has not reached the preset benchmark is automatically selected as a high-priority site, and the coordinates of the site that has reached the temperature benchmark are removed; using the end coordinates of the previous round as the starting point, the remaining high-priority sites are re-optimized locally, and the second motion control trajectory is generated according to the principle of proximity concentration.
[0042] When the output round is 3, the motion trajectory planning data also includes the third round of supplementary fine-tuning trajectory planning: thermal imaging data is collected again, and only the coordinates of the supplementary measurement points in areas with insufficient temperature after the first two rounds of output are retained; the nearest neighbor greedy algorithm is used to directly plan from the end point coordinates of the previous round to the nearest supplementary measurement point coordinates to generate the third motion control trajectory.
[0043] The first, second, and third motion control trajectories mentioned above are sequentially connected in time series and do not return to the initial standby coordinates in the middle. Each round updates the control parameters and end effector attitude based on the output effect of the previous round.
[0044] Furthermore, the above scheme also includes the step of performing adaptive temperature closed-loop control for each target site: acquiring the temperature data of each target site and comparing it with the preset target temperature; if the temperature rises steadily to the preset target temperature, maintaining or fine-tuning the microwave output power to stabilize the temperature within the range of ±0.5℃ of the preset target temperature; if the temperature rise rate exceeds the preset safety threshold, immediately reducing the microwave output power to slow down the heating rate; if the temperature exceeds the preset safety upper limit, immediately cutting off the microwave radiation equipment output and issuing an alarm.
[0045] Intelligent Control System Examples This invention also provides an intelligent control system for a microwave radiation device, such as... Figure 2 As shown, the intelligent control system includes a multimodal sensing unit, a control unit, and a multi-axis execution unit.
[0046] A multimodal sensing unit (MSU) acquires spatial point cloud data and infrared thermal images of a target object, and performs data registration and fusion to construct a digital model containing geometric structure and thermal field distribution, thereby identifying the target region. This MSU can be composed of a high-precision industrial 3D depth camera and an infrared thermal imager. The depth camera acquires spatial 3D point cloud data of the target object, while the infrared thermal imager acquires real-time thermal images of the target object's surface temperature distribution.
[0047] The control unit, connected to the multimodal sensing unit, executes the intelligent control method for the microwave radiation device described above to generate corresponding motion control commands and power adjustment commands. Since this intelligent control method has been described in detail in the above embodiments, it will not be repeated here.
[0048] A multi-axis actuator, connected to a control unit, includes a multi-degree-of-freedom collaborative robotic arm and a microwave radiation generator mounted at the end of the robotic arm. It is configured to drive the robotic arm to various output positions and adjust the energy output of the microwave generator based on motion control commands and power adjustment commands generated by the control unit. For example, the multi-axis actuator can be a six-degree-of-freedom collaborative robotic arm with a microwave radiation generator (microwave radiator head) mounted on its end flange.
[0049] Application Examples This embodiment uses the physiotherapy process of a microwave therapy device as an example to illustrate the solution of the present invention. Multiple rounds of microwave irradiation are performed on an approximately 5cm × 5cm inflamed area (shown as a high-temperature area) on the posterior side of a patient's shoulder joint using a microwave radiation device. The path planning of the robotic arm of the microwave therapy device is then performed in conjunction with the solution of the present invention. For example... Figure 3 As shown, the patient's human tissue is detected, and a digital model of its geometry and thermal field distribution is generated. During the operation, the treatment head attached to the end of the robotic arm remains perpendicular to the surface of the target area.
[0050] Before starting the operation, the multimodal sensing unit first scans the target object (e.g., the patient). The depth camera acquires a 3D point cloud, the infrared thermal imager acquires the initial thermal field, and the control unit generates a digital model of the target area that integrates the geometric structure and thermal field distribution through spatial registration and point cloud fusion algorithms. It also automatically identifies the target area that needs energy projection, i.e., the user's diseased area.
[0051] In one application scenario, a 3D depth camera first scans the patient's body surface, generating a high-precision 3D point cloud. This point cloud forms a coordinate system with the treatment room as a reference. An infrared thermal imager simultaneously acquires surface temperature data—a heat map—sharing the field of view with the 3D camera. Through built-in calibration parameters, it can directly and automatically attach to the corresponding positions in the 3D point cloud, forming a skin surface model with temperature attributes. The patient's pre-stored CT / MRI images have their own coordinate system, completely different from the patient's current lying posture and position. Automatic registration is performed using the ICP algorithm. The algorithm extracts features: a large number of feature points are extracted from both the 3D point cloud and the CT / MRI-reconstructed bone surface model. Through continuous iteration, an optimal rotation matrix and translation vector are calculated, minimizing the average distance between the two point sets. Using this transformation matrix, the entire CT / MRI data (including all layers of bone, muscle, and blood vessels) is rigidly transformed into the coordinate system established by the 3D camera. Finally, the patient's real-time surface morphology (point cloud), real-time surface physiological state (thermograph), and high-resolution internal anatomical structure (CT / MRI) are all precisely aligned in the same three-dimensional space. A hot spot on the skin can accurately correspond to an inflamed area of a muscle or joint capsule at a specific depth under the skin.
[0052] The control unit extracts the surface area S of the target region. It reads the inherent physical parameters of the microwave radiation generator: the effective coverage diameter d of a single radiation point, and the overlap coefficient η (typically 0.1 ≤ η ≤ 0.3) set to ensure edge energy continuity. Based on the spatial geometric coverage formula, it calculates the total theoretical number of output points N required to achieve absolute full coverage of the target region.
[0053] Based on the total number N, the control unit generates N uniformly distributed discrete candidate output point coordinate sequences on the surface of the three-dimensional model of the target area, and calculates the spatial normal direction of each point to guide the posture alignment of the robotic arm end effector.
[0054] The system is configured with a standard output duration t for a single site. p (e.g., 15s), the average movement time t of the robotic arm between two adjacent points. m (e.g., 2s), and the system's maximum allowed total output time T. total (Safe operating window time). The control unit first calculates the predicted time T for single-round absolute full coverage. single .
[0055] Subsequently, the control unit will T single With T total The comparison is performed, and the total number of rounds P to be executed is adaptively determined and the site is assigned.
[0056] (1) Determine P = 1 (time is extremely tight).
[0057] When T single >T total This indicates that under the current constraints, the system cannot even complete a full coverage operation in one round. The control unit is forced to execute degradation control, determining the total number of rounds as P = 1, and according to the formula N1′ = T total +t m / t p +t m The number of computation sites in the first round is forcibly reduced by downsampling to remove some marginal sites, ensuring that the single-round reduction operation is completed within T. total Internal security has been completed.
[0058] (2) Determine whether P=2 or P=3 (lower-order rounds).
[0059] When T single <T total When the system determines that it can execute 2 or 3 rounds of progressive fine-grained planning, it will proceed according to the preset decay factor k. i The maximum allocation for subsequent rounds is distributed in descending order, where the maximum allocation for the second round, N2, is... k2N The maximum quota for the third round is N3 = k3N And satisfy 1>k2>k3>0.
[0060] (3) Determine if P is greater than 3 (multi-round progressive method).
[0061] When T single <T total Furthermore, without manually specifying a particular low-order round, the system determines that multi-round progressive fine-grained planning can be executed. This is done without exceeding T... total Given the premise of dividing the budget into P rounds, the control unit introduces a dynamic compression relaxation factor α (0 < α ≤ 1) and obtains the system's preset initial geometric decay factor k for each round. i (satisfying k1=1, 1>k2>k3>···>k) P >0).
[0062] The control unit establishes and solves the following time boundary equations:
[0063] Based on the solved relaxation factor, determine the number N of dynamic compression calculation points in the i-th round. i ′= N×k i ×α , where k iThe initial convergence factors for each round are i=1,2,…,P, and k1=1, where α is the relaxation factor.
[0064] Specifically, the control unit uses a bisection method to inversely solve for the value of the maximum relaxation factor α that satisfies the inequality. For example, if the set number of rounds is high (e.g., P=5) or the time is relatively tight, causing the time constraint equation to exceed the limit when α=1, then an under-relaxation factor less than 1 is obtained (e.g., α=0.65). In this case, the number of dynamically compressed pre-allocated sites in the i-th round will be limited to N. i ′= N×k i ×α This compresses the high-level, multi-round requirements into a safe timeline.
[0065] After clarifying the site quotas (N) for each round i or N i After ′), the control unit initiates cascaded motion trajectory generation: First round (full coverage trajectory generation): The control unit takes the coordinates of all N uniformly distributed candidate points as input and uses the Global Traveling Salesman (TSP) optimization algorithm to calculate a cyclic control trajectory with the shortest total path. The control unit drives the robotic arm carrying the microwave generator to move sequentially to each coordinate point along this trajectory, and stops at each point for t seconds. p The basic energy is projected over time. After the operation is completed, the control unit records the endpoint coordinates of the robotic arm in memory.
[0066] The second and subsequent rounds (focusing on reinforcement and remediation): Trajectory planning in subsequent rounds will no longer return to the initial standby point, but will instead start directly from the endpoint coordinates of the previous round. Before entering the next round, the infrared thermal imager will collect the latest heat distribution data of the target area in real time. The control unit will read the real-time actual temperature T around each location. real and the preset standard target temperature T safe Compare: If T real ≥T safe This indicates that the medium energy corresponding to the site has reached the standard, and the control unit removes the site coordinates from the candidate set in the current round.
[0067] If T real <T safe This indicates that there is a heat gap in the area, which is a cold spot. The control unit retains this location and processes it according to the degree of temperature deficiency (T). real T safe Sort in descending order using the ')' option.
[0068] The control unit allocates site quota N according to the current round. i′, extract the N that is at the top of the list and most in need of additional energy. i Given the coordinates of 'cold points, a local path optimization algorithm is invoked to generate the motion control trajectory for the current round. The trajectories of each round are connected end-to-end on the time axis. As round i approaches P, the unmet target areas achieve dynamic convergence under thermal field feedback.
[0069] Furthermore, the solution of this invention also includes a timeout protection mechanism and a temperature control power closed-loop strategy. To cope with the surge in movement time caused by sudden trajectory disturbances such as obstacle avoidance by the robotic arm, a high-priority dynamic timing compression mechanism runs internally within the control unit. After each point is executed, the system re-predicts the remaining total time T for all subsequent rounds in real time. sum Once T is discovered sum >T total Immediately follow the procedure of "canceling the last round of planning → proportionally reducing the remaining positions in the current round → shortening the single-point output time t by microseconds". p The cascading priority is used to compress trajectory data in real time online, establishing a proactive safety defense.
[0070] Meanwhile, when the robotic arm stops at any predetermined output position to project microwaves, the control unit activates adaptive closed-loop temperature control: continuously compares the real-time temperature rise fed back by the infrared thermal imager; if the temperature rise rate dT / dt exceeds the safety threshold, immediately reduce the microwave output power signal; if the absolute temperature exceeds the absolute safety limit, the control unit immediately outputs an emergency stop signal, cuts off the robotic arm drive current and completely cuts off the energy output of the microwave generator, ensuring the absolute safety of the entire automated operation process.
[0071] In addition, to reduce the accumulated error of multi-point movement of the robotic arm, one or more of the following methods can be used: 1. Real-time position comparison and micro-compensation correction are performed using 3D vision at each treatment point; 2. Global absolute coordinates are established based on the patient's skeletal landmarks, and all treatment points are calculated based on these global coordinates, independent of relative positions; 3. The kinematic parameters of the robotic arm are calibrated in real time to eliminate mechanical backlash and temperature drift errors; 4. Long-distance movement adopts segmented paths, and visual correction is performed after each segment to control the accumulated error within the allowable range. Furthermore, changes in body position are monitored in real-time using 3D vision, and the path is automatically replanned when the deviation exceeds a threshold.
[0072] Furthermore, this embodiment also includes intelligent temperature feedback control, encompassing the following three scenarios: Scenario A (Normal Heating): The temperature rises steadily to the preset value, and the algorithm maintains or fine-tunes the power to keep the temperature constant. Scenario B (Rapid Heating): The rate of temperature increase exceeds a threshold, and the algorithm immediately reduces the microwave power to prevent overheating. Scenario C (Abnormal or Exceeding Temperature Limits): The temperature at the monitoring point exceeds the safe upper limit, the system triggers an emergency stop, suspends microwave output, and issues an alarm.
[0073] Through the above-described solution of the present invention, the planning rounds can be adaptively adjusted, output sites can be dynamically allocated, and temperature closed-loop control can be achieved under strict time constraints, effectively improving the targeting accuracy of energy projection. Simultaneously, by introducing a dynamic relaxation factor and boundary integral equations, a high-order general-purpose iterative algorithm that supports any P rounds (P>3) without modifying the underlying software is provided. The control unit solves for the relaxation factor through a one-dimensional search, adaptively mapping and compressing the high-order round requirements into a finite time axis, achieving a high-precision control effect of gradual approximation and infinite convergence. Through infrared feedback, a mechanism is used to completely eliminate sites that meet the criteria and sort and extract those that do not meet the criteria in descending order of their deficiency. This allows electromagnetic energy and movement time to be concentrated at high density on the coldest points (core target areas) that most require additional energy, greatly improving the thermal field uniformity and overall processing quality of the system.
[0074] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.
Claims
1. An intelligent control method for a microwave radiation device, characterized in that, The method is executed by the control unit and includes the following steps: S1: Acquire spatial point cloud data and infrared thermal map of the target object, and perform data registration and fusion to construct a digital model containing geometric structure and thermal field distribution and identify the target area; S2: Calculate the total number of theoretical output sites required for full coverage based on the spatial geometric information of the target area, and calculate the total duration of a single round of full coverage based on the standard output duration of a single site and the average movement time between adjacent sites. S3: Compare the total duration of the single-round full coverage with the preset total output time limit to determine the output round executed by the microwave radiation device, and allocate the corresponding number of calculation points and output time to each output round so that the predicted total output time does not exceed the total output time limit. S4: Based on the assigned output rounds, generate multi-round progressive motion trajectory planning data in sequence; wherein, the trajectory planning of the next round takes the output end coordinates of the previous round as the starting point, and dynamically selects the target position of the next round based on the real-time temperature feedback data after the output of the previous round, forming hierarchical control data.
2. The intelligent control method according to claim 1, characterized in that, In step S2, the calculation of the total number of theoretical output sites required for full coverage includes: Extract the surface area S of the target region; Obtain the effective coverage diameter d and overlap coefficient η of a single radiation point; Calculate the total number of theoretical output sites required for full coverage. The formula for calculating the total number of theoretical output sites is: N = S / (d(1 η)) 2 .
3. The intelligent control method according to claim 1, characterized in that, Step S3, which involves assigning the corresponding number of computation points to each output round, specifically includes: Calculate the single-round full coverage time T single =N×t p +(N-1)×t m Where N is the total number of output sites, t p The duration of a single standard output, t m This represents the average travel time. If the output round is 1 and T single >T total When, according to the formula N1′= T total +t m / t p +t m Redetermine the number of output sites, T total This represents the maximum total output time. If the output round is 2 or 3, the preset attenuation factor k will be applied. i The maximum allocation for subsequent rounds is distributed in descending order, where the maximum allocation for the second round, N2, is... k2N The maximum quota for the third round is N3 = k3N And satisfy 1>k2>k3>0.
4. The intelligent control method according to claim 3, characterized in that, It also includes a timeout dynamic compression step: Calculate the estimated time for each round separately, and when the calculated total time T sum Exceeding the total output time limit T total At the same time, data is dynamically compressed according to the priority of canceling the third round, reducing the number of sites in the second round, and shortening the single-point output time, so as to ensure that the control sequence is completed within a safe time.
5. The intelligent control method according to claim 3, characterized in that, The motion trajectory planning data specifically includes the first round of full-coverage trajectory planning: Generate coordinates of all candidate output sites that are uniformly distributed based on the 3D model of the target region; The shortest loop path is calculated using the global traveling salesman optimization algorithm, and the first motion control trajectory is generated by sequentially traversing all candidate output points. Record the endpoint coordinates of the first motion control trajectory as the starting point for the next round of trajectory planning.
6. The intelligent control method according to claim 5, characterized in that, The motion trajectory planning data also includes a second round of focused trajectory planning: Based on the real-time temperature feedback data after the first round of output, the target area or core target area whose temperature has not reached the preset benchmark is automatically selected as a high-priority site, and the coordinates of the sites whose temperature has reached the benchmark are removed. Starting from the endpoint coordinates of the previous round, the remaining high-priority sites are re-optimized locally, and a second motion control trajectory is generated according to the principle of proximity.
7. The intelligent control method according to claim 6, characterized in that, The motion trajectory planning data also includes a third round of error correction and fine-tuning trajectory planning: Thermal imaging data was collected again, and only the coordinates of the supplementary measurement points in areas where the temperature was still insufficient after the first two rounds of output were retained. The nearest neighbor greedy algorithm is used to directly plan the coordinates of the nearest supplementary measurement point from the end point coordinates of the previous round to generate the third motion control trajectory; The first, second, and third motion control trajectories are sequentially connected in time and do not return to the initial standby coordinates in the middle. Each round updates the control parameters and end effector attitude based on the output effect of the previous round.
8. The intelligent control method according to claim 3, characterized in that, If the output round is greater than 3, the steps for dynamically calculating the number of calculation points allocated to subsequent rounds include: Construct the time boundary equations and solve in reverse for the maximum allowable value of the relaxation factor. The time boundary equations are as follows: Based on the solved relaxation factor, determine the number N of dynamic compression calculation points in the i-th round. i ′= N×k i ×α , where k i The initial convergence factors for each round are i=1,2,…,P, and k1=1, where α is the relaxation factor.
9. The intelligent control method according to claim 1, characterized in that, It also includes the step of performing adaptive temperature closed-loop control for each target site: Acquire temperature data for each target location and compare it with the preset target temperature; If the temperature rises steadily to the preset target temperature, maintain or fine-tune the microwave output power to keep the temperature stable within the preset target temperature range of ±0.5℃. If the rate of temperature rise exceeds the preset safety threshold, immediately reduce the microwave output power to slow down the heating rate. If the temperature exceeds the preset safety limit, immediately cut off the output of the microwave radiation equipment and issue an alarm.
10. An intelligent control system for a microwave radiation device, characterized in that, include: A multimodal sensing unit is used to acquire spatial point cloud data and infrared thermal images of the target object, and to register and fuse the two data to construct a digital model containing geometric structure and thermal field distribution and identify the target area; A control unit, connected to the multimodal sensing unit, is used to execute the intelligent control method of the microwave radiation device as described in any one of claims 1-9, so as to generate corresponding motion control commands and power adjustment commands; A multi-axis actuator, connected to the control unit, includes a multi-degree-of-freedom cooperating robotic arm and a microwave radiation generator mounted at the end of the robotic arm. It is configured to drive the robotic arm to move to each output position and adjust the energy output of the microwave generator according to motion control commands and power adjustment commands generated by the control unit.