A radiotherapy device collision avoidance system based on multiple detection modalities

Abstract

The application discloses a radiotherapy device collision avoidance system based on multiple detection modes, which comprises from top to bottom: an upper computer for displaying alarm information, a PLC controller connected to the upper computer upward and connected to a motion device downward, the motion device controlled by the PLC controller, and a collision detection unit for providing collision detection information for the PLC controller; the motion device comprises a rotating gantry, a treatment bed and other motion devices related to a treatment head; the treatment bed is a robot treatment bed, and a treatment bed envelope comprises an envelope of the treatment bed itself and an envelope of a patient on the treatment bed; the application combines a distance real-time collision detection method, a pressure real-time collision detection method and a space coordinate real-time collision detection method to form a complete collision avoidance system; through multiple detection methods, potential collision threats in the radiotherapy process are comprehensively protected, and the incidence of collision accidents is effectively reduced.

Description

Technical Field

[0001] This invention belongs to the field of radiotherapy, and more specifically, relates to a collision avoidance system for radiotherapy devices based on multiple detection methods. Background Technology

[0002] Radiotherapy equipment is currently primarily used for the radiotherapy of malignant tumors. A typical radiotherapy device includes: a gantry that rotates around a rotation axis, a treatment head mounted on the rotating gantry, an image-guided DR (radiotherapy) panel, and a treatment bed that precisely positions the patient under the treatment head for treatment. During actual radiotherapy, the rotation of the gantry, the swinging of the treatment head and DR panel, and the movement of the treatment bed all pose potential collision risks. Such collisions can injure patients and staff, and may also cause serious damage to the radiotherapy equipment.

[0003] The challenge in designing collision avoidance systems for radiotherapy devices lies in the fact that existing radiotherapy equipment designs often employ single-method collision detection. These single methods are often insufficient to address the complex and ever-changing environmental factors present in real-world scenarios. For example, 1. Collision detection using distance sensors requires sufficient space to deploy them at all potential collision points. Distributing sensors across all potentially hazardous areas would encroach on the space needed for the treatment bed or rotating gantry, which is clearly unacceptable. Therefore, distance sensors must be deployed in a limited number of areas without affecting their operation. This leaves most areas where installation is inconvenient as collision detection is impossible. 2. Some collisions cannot be accurately assessed using distance sensors. Not all collisions can be measured by distance alone. For instance, if a heavy metal plate suddenly falls on the treatment bed and patient, and the plate is relatively thin, the distance from the treatment bed + plate to the distance sensor is not significantly different from the distance from the treatment bed alone. In this case, a distance sensor alone is ineffective. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention proposes a collision avoidance system for radiotherapy devices based on multiple detection methods. The aim is to solve the problem that a single collision avoidance detection method often cannot meet the complex and ever-changing environmental factors in actual situations.

[0005] To solve its technical problem, the present invention adopts the following technical solution:

[0006] A collision avoidance system for radiotherapy devices based on multiple detection methods is characterized by the following: From top to bottom, the system includes: a host computer for displaying alarm information; a PLC controller connected to the host computer and to the motion equipment; the motion equipment controlled by the PLC controller; and a collision detection unit providing collision detection information to the PLC controller. The motion equipment includes a rotating gantry, a treatment bed, and other motion equipment related to the treatment head. The treatment bed is a robotic treatment bed, and its envelope includes both the treatment bed's own envelope and the patient's envelope on the treatment bed.

[0007] The collision detection unit includes: a distance comparison collision detection subunit based on the current distance to the treatment bed, a pressure comparison collision detection subunit based on the current pressure of the treatment bed, and a VxWorks spatial position comparison collision detection subunit based on the current spatial position of the treatment bed;

[0008] The distance comparison collision detection subunit includes detectors, laser distance sensors, multi-level warning zones divided around the treatment head, DR panel, and treatment bed, as well as a warning zone boundary calculation unit.

[0009] The pressure comparison collision detection subunit includes a pressure sensor installed on the treatment bed and a pressure comparison module for comparing whether the current pressure is normal or not.

[0010] The VxWorks spatial position comparison collision detection subunit includes programming software and 3D modeling software, a position comparison collision detection database, and a VxWorks-based position information acquisition and calculation module. The programming software and 3D modeling software save the 3D model to the position comparison collision detection database. The VxWorks-based position information acquisition and calculation module acquires 3D modeling information from the position comparison collision detection database and position information of relevant moving equipment, including the treatment bed and rotating frame. It then calculates these two types of information in VxWorks to obtain the collision prediction result for the next moment and reports it to the PLC. In dangerous situations, the PLC will quickly shut down or brake the relevant moving equipment. The 3D model is a model of the safe movement range of the treatment bed.

[0011] Furthermore, the multi-level warning zone of the distance comparison collision detection subunit includes dividing the treatment head into a warning zone and a danger zone based on the treatment head; dividing zone a and zone b centered on the treatment bed; the warning zone surrounds the danger zone; the shape of the warning zone is approximately trapezoidal, and the shape of the danger zone is approximately an envelope formed by the treatment head and the DR panel when deployed; zone a is the danger zone of the treatment bed, and zone b is the secondary danger zone of the treatment bed; when the distance between the treatment bed and the treatment head enters the warning zone of the treatment head, the system alarms; when the distance between a person or object and the treatment head enters the danger zone of the treatment head, the motion equipment of each radiotherapy facility stops moving; the motion equipment includes, but is not limited to, a rotating gantry, treatment bed, DR panel, and a range shifter on the treatment head; when a person or object is detected in zone b around the treatment bed, the system issues a warning, and the treatment bed decelerates; when a person or object is detected in zone a around the treatment bed, the system issues an audible and visual alarm, and the treatment bed brakes.

[0012] Furthermore, the warning zone boundary calculation unit of the distance comparison collision detection subunit calculates the envelope boundary of the danger zone of the treatment head based on the dimensions of the rotating gantry, treatment head, DR plate, and the maximum braking angle of the rotating gantry.

[0013] Furthermore, the pressure comparison module of the pressure comparison collision detection subunit divides the pressure detection process into four time points: 0, t1, t2, and t3. From time 0 to time t1, this time period simulates the process of a patient lying on the treatment bed; from time t1 to time t2, this time period simulates the process of a patient lying steadily on the treatment bed receiving treatment. Although the curve fluctuates during this stage, it eventually tends to flatten out; from time t2 to time t3, this time period simulates the process of a sudden increase or decrease in pressure. The sudden change is a sudden increase or decrease in pressure within a short period of time. This process is defined as the pressure abnormal change stage. The device is very likely to have experienced a collision or other abnormal situation during this time period. When the pressure comparison unit detects such a sudden change, it will stop the operation of the radiotherapy device.

[0014] Furthermore, the VxWorks spatial position comparison collision detection subunit includes a 3D modeling module, a VxWorks spatial position coordinate acquisition module, and a real-time judgment module. The 3D modeling module models the safe movement zone of the treatment bed within the rotating frame using programming software and 3D modeling software. The established safe movement zone model is repeatedly adjusted in position and angle within the program, and then saved to the position comparison collision detection database. The VxWorks spatial position coordinate acquisition module acquires the current position and movement speed of the treatment bed and rotating frame in real time through the VxWorks system. The real-time judgment unit determines whether the treatment bed will collide with its surrounding equipment at the next moment based on the data obtained from the spatial position coordinate unit, and sends the judgment information to the PLC controller. If a collision is predicted, the PLC controller will brake the treatment bed in a timely manner.

[0015] Furthermore, the 3D modeling module includes a coordinate system establishment submodule, a treatment head angular position division submodule, a first-level loose mesh division submodule, and a second-level precise mesh division submodule. The treatment head angular position division submodule is used to divide the treatment head 360 degrees into several equal parts, each part corresponding to an angle of the treatment head. The first-level loose mesh division submodule, according to the current angle of the treatment head, defines the active area of ​​the treatment bed away from the collision object or the danger zone of the treatment head as multiple first-level loose mesh areas, and saves multiple first-level loose mesh areas to the database. The meshes of these multiple first-level loose mesh areas are relatively sparse and the mesh size is relatively large, with each loose mesh serving as a detection point. The second-level precise mesh division submodule, according to the current angle of the treatment head, defines the active area of ​​the treatment bed close to the collision object or the danger zone of the treatment head as multiple second-level precise mesh areas, and saves multiple second-level precise mesh areas to the database. The meshes of these multiple second-level precise mesh areas are relatively dense and the mesh size is relatively small, with each second-level precise mesh division submodule serving as a detection point. The collision object includes, but is not limited to, the rotating frame around the treatment bed, the treatment head, the DR plate, and the rolling floor.

[0016] Furthermore, the coordinate system establishment submodule establishes a three-dimensional rectangular coordinate system (X,Y,Z) with the center of the circular ring of the front cross section of the rotating frame as the origin. This coordinate system is the basis for all subsequent position calculations.

[0017] Furthermore, the VxWorks spatial coordinate acquisition module acquires the three-dimensional coordinates of the treatment bed, as well as the Euler angles (rX, rY, rZ) of the treatment bed and the moving speed and direction of the treatment bed; it also acquires the angle of the rotating frame and its rotation speed and direction.

[0018] Furthermore, the real-time judgment module determines whether the treatment bed and its surrounding equipment will collide at the next moment based on the current safe movement range model of the treatment bed, the three-dimensional coordinates of the treatment bed, the Euler angles (rX, rY, rZ) of the treatment bed, the movement speed and direction of the treatment bed, the angle of the rotating frame, and the rotation speed and direction. If it determines that a collision is about to occur, it reports to the PLC controller, and the PLC controller will brake the treatment bed in time.

[0019] Furthermore, the treatment head angular position division submodule is used to divide the treatment head 360 degrees into several equal parts, which include 72 parts, with each part spaced 5° apart; the multiple primary loose grid regions include 72 primary loose grid regions; the multiple secondary precise grid regions include 72 secondary precise grid regions.

[0020] Advantages and effects of the present invention

[0021] 1. This invention has a precise distance detection and response mechanism: it determines the specific location of the target to be detected in the radiotherapy equipment and monitors its surrounding area. The target position is detected in real time through the detection device, and the response is made according to the settings of safe distance, secondary danger distance and danger distance. It realizes precise monitoring of the radiotherapy equipment and its surrounding objects, and can take corresponding safety measures, such as slowing down or stopping operation, within different distance ranges to avoid collisions.

[0022] 2. This invention also features a unique pressure anomaly detection and shutdown mechanism: It uses a pressure detection device to monitor changes in the pressure and compression pressure on the target object in real time to determine if any anomalies exist; it can promptly shut down the radiotherapy equipment in case of pressure anomalies, preventing collisions or damage caused by pressure changes.

[0023] 3. This invention relates to a method for interacting with a proportional 3D model in 3D design simulation software using the C# programming language to construct a lookup table for collision-free safe positioning. The system is equipped with components for real-time monitoring of the 3D coordinates and movement speed of the treatment bed. The collision prevention algorithm runs on a control board running the VxWorks operating system. By matching real-time acquired data with a lookup table pre-stored in the database, the system can instantly assess whether there is a potential collision risk at the next moment's position of the treatment bed. Upon detecting a risk, the system automatically adjusts the speed of relevant equipment or implements emergency stop measures to ensure operational safety.

[0024] 4. This invention proposes a multi-layered, in-depth collision avoidance method based on existing collision detection techniques for radiotherapy equipment. This invention combines real-time distance collision detection, real-time pressure collision detection, and real-time spatial coordinate collision detection to form a complete collision avoidance system. Through multi-layered detection methods, it provides comprehensive protection against potential collision threats during radiotherapy, effectively reducing the incidence of collision accidents. Attached Figure Description

[0025] Figure 1 This is an illustration of the application effect of a radiotherapy device.

[0026] Figure 2 A schematic diagram of the six-dimensional motion of the robotic treatment bed;

[0027] Figure 3a This is a front view of the radiotherapy device along the x and y axes;

[0028] Figure 3b This is a side view of the radiotherapy device along the yz axis.

[0029] Figure 4a This is a schematic diagram of the multi-level warning zone of the distance comparison collision detection mechanism of the present invention;

[0030] Figure 4b This is a schematic diagram illustrating the calculation of the warning zone boundary of the distance comparison collision detection mechanism of the present invention.

[0031] Figure 4c This is a schematic diagram of area a and area b of the treatment bed in the distance comparison collision detection mechanism of the present invention;

[0032] Figure 4d This is a logic flowchart of the distance comparison collision detection mechanism of the present invention.

[0033] Figure 5a This is a logic flowchart of the pressure comparison collision detection mechanism of the present invention.

[0034] Figure 5b This is a schematic diagram simulating four time points in the pressure comparison unit of the pressure comparison collision detection mechanism of the present invention.

[0035] Figure 6a First-person perspective of 3D modeling of the safe movement zone of the spatial position comparison collision detection mechanism of the present invention;

[0036] Figure 6b The second perspective of 3D modeling the safe movement range of the spatial position comparison collision detection mechanism of the present invention;

[0037] Figure 7 This is a schematic diagram illustrating the overlay effect of 3D modeling of the multi-level warning zone and safe movement zone of the present invention.

[0038] Figure 8 This is a logic flowchart of the spatial position comparison collision detection mechanism of the present invention.

[0039] Figure 9 This is a diagram illustrating the collision avoidance system architecture of a radiotherapy device based on multiple detection methods according to the present invention. Detailed Implementation

[0040] Design principle of the invention

[0041] 1. Innovation of this invention: The innovation lies in combining three collision detection methods. Each method leverages its strengths to compensate for the weaknesses of the others, supporting each other to achieve the combined effect. The combined effect is that it meets the needs of collision detection in the complex and ever-changing environmental factors of radiotherapy. Distance detection methods are only suitable for the area around the treatment head and DR panel, as sensors can be deployed around these areas. Although distance detection methods require additional space for sensor deployment, collision detection is relatively simple and direct, effectively saving computational resources. Pressure detection methods can compensate for the limitations of distance detection methods. Spatial position detection methods require a detection area that occupies 100% of the remaining space of the rotating gantry, or 100% of the remaining space outside the treatment head danger zone of the rotating gantry. The remaining space of the rotating gantry is the space within the rotating gantry after removing the treatment head, DR panel, rolling floor, and treatment bed. The remaining space outside the treatment head danger zone of the rotating gantry is the space within the rotating gantry after removing the treatment head, DR panel, rolling floor, treatment bed, and treatment head danger zone. Spatial position detection methods are sensorless collision detection methods. They do not rely on traditional detection equipment but instead use modeling and comparison with models, thus solving the collision detection problem without sensors. This is a major innovation of this invention.

[0042] 2. Design Challenges and Solutions of this Invention: The challenge lies in the excessive computational resources required for the spatial position detection method. With 72 models, each with a 1cm fine mesh, each model contains approximately hundreds of millions of data points (determined by the six-axis movement parameters of the treatment bed). If 100 data points can be calculated per second, calculating one model would take more than 11 days. To reduce computational burden and save time, firstly, models are pre-built. Modeling takes the longest time, so it's crucial to complete the modeling process in advance. Secondly, due to the insufficient response speed and flexibility of PLCs, and their relatively fixed hardware and software architecture, they cannot meet the requirements of the spatial position comparison and collision detection program. This invention employs a VxWorks system with a priority-driven preemptive scheduling strategy, a low-latency interrupt handling mechanism, and optimized inter-task communication to acquire real-time spatial position and motion data of the treatment bed and rotating gantry, along with pre-generated modeling data. Based on this data, it calculates the collision state for the next moment, ensuring the system can respond to external events within microseconds to milliseconds. Third, the detection grid is divided into a first-level loose grid and a second-level fine grid. The first-level loose grid is used as the detection point for areas far from the collision object, while the second-level fine grid is used for areas close to the collision object. For the same area size, the data volume of the second-level fine grid is approximately 1000 times that of the first-level loose grid (the first-level grid has a length, width, and height of 10 cm, while the second-level grid has 1 cm), and the first-level loose grid occupies more than 90% of the total grid space. This means that for the same area size, the time required to calculate the first-level loose grid is 1 / 1000th of the time required to calculate the second-level fine grid, thus reducing the total calculation time to less than 10.09% of the original calculation time. The reason for using a two-level fine mesh is that the treatment requires high precision. The movement of the treatment bed is often measured in increments of less than 1 cm, not 10 cm. Therefore, a two-level fine mesh is essential for determining the safe position of the treatment bed. Furthermore, the 72 models are independent of each other, allowing for separate calculations on multiple computers to save computation time. Assuming simultaneous calculations on 10 computers, the original 792 days required to calculate all 72 models can be reduced to approximately 8 days. This mesh generation method effectively reduces computational effort without sacrificing detection accuracy.

[0043] 3. Methods for handling overlapping areas: such as Figure 7 As shown, when the third spatial position detection method is used, VxWorks obtains the current spatial position of the treatment bed. When this spatial position enters the warning area or danger area of ​​the treatment head, the PLC processes the braking or warning of the equipment according to the distance detection method.

[0044] The invention will be further explained below with reference to the accompanying drawings:

[0045] Based on the above principles, this invention designs a radiotherapy device collision avoidance system based on multiple detection methods, such as... Figure 9 As shown, the system is characterized by the following features: from top to bottom, it includes: a host computer for displaying alarm information; a PLC controller connected to the host computer above and the motion equipment below; the motion equipment controlled by the PLC controller; and a collision detection unit that provides collision detection information to the PLC controller. The motion equipment includes a rotating frame, a treatment bed, and other motion equipment related to the treatment head. The treatment bed is a robotic treatment bed, and the treatment bed envelope includes the treatment bed's own envelope and the patient's envelope on the treatment bed.

[0046] like Figure 9 As shown, the collision detection unit includes: a distance comparison collision detection subunit based on the current distance to the treatment bed, a pressure comparison collision detection subunit based on the current pressure of the treatment bed, and a VxWorks spatial position comparison collision detection subunit based on the current spatial position of the treatment bed.

[0047] like Figure 9 As shown, the distance comparison collision detection subunit includes detectors, laser distance sensors, multi-level warning zones divided around the detectors and laser distance sensors, and a warning zone boundary calculation unit.

[0048] like Figure 9 As shown, the pressure comparison collision detection subunit includes a pressure sensor installed on the treatment bed and a pressure comparison module for comparing whether the current pressure is normal or not.

[0049] like Figure 9 As shown, the VxWorks spatial position comparison collision detection subunit includes programming software and 3D modeling software, a position comparison collision detection database, and a VxWorks-based position information acquisition and calculation module. The programming software and 3D modeling software save the 3D modeling model to the position comparison collision detection database. The VxWorks-based position information acquisition and calculation module acquires 3D modeling information from the position comparison collision detection database and position information of relevant moving equipment, including the treatment bed and rotating frame. It then calculates these two types of information in VxWorks to obtain the collision prediction result for the next moment and reports it to the PLC. In dangerous situations, the PLC will quickly shut down or brake the relevant moving equipment. The 3D model is a model of the safe movement range of the treatment bed.

[0050] like Figure 4a , 4bAs shown in Figure 4c, the multi-level warning zone of the distance comparison collision detection subunit includes a treatment head warning zone and a treatment head danger zone based on the treatment head; zones a and b centered on the treatment bed; the treatment head warning zone surrounds the treatment head danger zone; the shape of the treatment head warning zone is approximately trapezoidal, and the shape of the treatment head danger zone is approximately an envelope formed by the treatment head and the DR panel when deployed; zone a is the treatment bed danger zone, and zone b is the treatment bed secondary danger zone; when the distance between the treatment bed and the treatment head enters the treatment head warning zone, the system alarms; when the distance between a person or object and the treatment head enters the treatment head danger zone, the motion equipment of each radiotherapy facility stops moving; the motion equipment includes, but is not limited to, a rotating gantry, treatment bed, DR panel, and a range shifter on the treatment head; when a person or object is detected in zone b around the treatment bed, the system issues a warning, and the treatment bed decelerates; when a person or object is detected in zone a around the treatment bed, the system issues an audible and visual alarm, and the treatment bed brakes.

[0051] like Figure 4a , 4b As shown in 4c, the warning zone boundary calculation unit of the distance comparison collision detection subunit calculates the envelope boundary of the danger zone of the treatment head based on the dimensions of the rotating gantry, treatment head, DR plate, and the maximum braking angle of the rotating gantry.

[0052] like Figure 5b As shown, the pressure comparison module of the pressure comparison collision detection subunit divides the pressure detection process into four time points: 0, t1, t2, and t3. From time 0 to time t1, this time period simulates the process of a patient lying on the treatment bed. From time t1 to time t2, this time period simulates the process of a patient lying steadily on the treatment bed receiving treatment. Although the curve fluctuates during this stage, it eventually flattens out. From time t2 to time t3, this time period simulates the process of a sudden increase or decrease in pressure. The sudden change is a sudden increase or decrease in pressure within a short period of time. This process is defined as the pressure abnormal change stage. The device is very likely to have experienced a collision or other abnormal situation during this time period. When the pressure comparison unit detects such a sudden change, it will stop the operation of the radiotherapy device.

[0053] The VxWorks spatial position comparison collision detection subunit includes a 3D modeling module, a VxWorks spatial position coordinate acquisition module, and a real-time judgment module. The 3D modeling module uses programming software and 3D modeling software to model the safe movement zone of the treatment bed within the rotating frame, and repeatedly adjusts the position and angle of the established safe movement zone model in the program, saving the adjusted model to the position comparison collision detection database. The VxWorks spatial position coordinate acquisition module uses the VxWorks system to obtain the current position and movement speed of the treatment bed and the rotating frame in real time. The real-time judgment unit determines whether the treatment bed will collide with its surrounding equipment at the next moment based on the data obtained by the spatial position coordinate unit, and sends the judgment information to the PLC controller. If a collision is determined to occur, the PLC controller will brake the treatment bed in time.

[0054] like Figure 6a , 6b As shown, the 3D modeling module includes a coordinate system establishment submodule, a treatment head angular position division submodule, a first-level loose mesh division submodule, and a second-level precise mesh division submodule. The treatment head angular position division submodule divides the treatment head 360 degrees into several equal parts, each corresponding to an angle of the treatment head. The first-level loose mesh division submodule, according to the current angle of the treatment head, defines the active area of ​​the treatment bed away from the collision object or the danger zone of the treatment head as multiple first-level loose mesh regions, and saves these multiple first-level loose mesh regions to the database. The meshes of these multiple first-level loose mesh regions are relatively sparse and the mesh size is relatively large, with each loose mesh serving as a detection point. The second-level precise mesh division submodule, according to the current angle of the treatment head, defines the active area of ​​the treatment bed close to the collision object or the danger zone of the treatment head as multiple second-level precise mesh regions, and saves these multiple second-level precise mesh regions to the database. The meshes of these multiple second-level precise mesh regions are relatively dense and the mesh size is relatively small, with each second-level precise mesh division submodule serving as a detection point. The collision object includes, but is not limited to, the rotating frame around the treatment bed, the treatment head, the DR plate, and the rolling floor.

[0055] like Figure 6a , 6b As shown, the coordinate system establishment submodule establishes a three-dimensional rectangular coordinate system (X,Y,Z) with the center of the circular ring of the front cross section of the rotating frame as the origin. This coordinate system is the basis for all subsequent position calculations.

[0056] like Figure 6a , 6bAs shown, the VxWorks spatial coordinate acquisition module acquires the three-dimensional coordinates of the treatment bed, as well as the Euler angles (rX, rY, rZ) of the treatment bed and the moving speed and direction of the treatment bed; it also acquires the angle of the rotating frame and its rotation speed and direction.

[0057] like Figure 6a , 6b As shown, the real-time judgment module determines whether the treatment bed and its surrounding equipment will collide at the next moment based on the current safe movement range model of the treatment bed, the three-dimensional coordinates of the treatment bed, the Euler angles (rX, rY, rZ) of the treatment bed, the movement speed and direction of the treatment bed, the angle of the rotating frame, and the rotation speed and direction. If it determines that a collision is about to occur, it reports to the PLC controller, and the PLC controller will brake the treatment bed in time.

[0058] like Figure 6a , 6b As shown, the treatment head angular position division submodule is used to divide the treatment head 360 degrees into several equal parts, which include 72 parts, with each part spaced 5° apart; the multiple primary loose grid regions include 72 primary loose grid regions; the multiple secondary precise grid regions include 72 secondary precise grid regions.

[0059] Example 1: Distance-based collision detection based on the distance between the current treatment bed and the treatment head

[0060] (1) Key components of radiotherapy facilities requiring collision detection: such as Figure 1 As shown, the treatment head, treatment bed, and rotating gantry are the main environmental components during the treatment process. For example, the target being tested is most likely to collide with components of the radiotherapy equipment, such as the treatment head or rotating gantry, which need to move during radiotherapy. The collision risk mentioned in the steps can be considered as the collision risk between the target being tested and components of the radiotherapy equipment, such as the rotating gantry or treatment head. To facilitate the definition of distance detection for collisions during treatment, we divide the components containing collision risk into two types.

[0061] (2) Two detection standards: The first distance comparison collision detection uses the treatment bed as the detection benchmark to determine the collision-prone locations around the treatment bed and define the treatment bed danger zone and treatment bed secondary danger zone. The second distance comparison collision detection uses the treatment head as the detection benchmark. During the treatment process, the treatment head will move and adjust the treatment distance as the rotating gantry rotates. While finding the optimal treatment position, it is necessary to prevent collisions with the treatment bed, the patient on the treatment bed, and other moving parts. We define the treatment head warning zone and treatment head danger zone based on this.

[0062] (3) Six-axis treatment bed, such as Figure 2As shown, the treatment bed has six independent axes of rotation, giving it six degrees of freedom. These axes allow for precise positioning of the robotic treatment bed within specified motion constraints, including translation (along all three axes X, Y, Z) and rotation (around all three axes Rx, Ry, Rz).

[0063] (4) Division of dangerous and secondary dangerous areas of a six-axis treatment bed: Taking the bed board and rotating frame of the treatment bed as the target to be detected as an example, the dangerous distance detection is based on the treatment bed. Figure 4c As shown, Figure 4c The images show the treatment bed in both its unfolded and closed states. Because the base is fixed to the ground, the upward space around the base as a column is immovable. The shaft fixed to the base will not collide within this range, so we do not include this area in the collision distance detection range. This method can also be used to define shafts fixed in other rotational directions. Figure 2 We can clearly see that in the horizontal direction, shaft 1 rotates along direction J1, shaft 2 rotates along direction J2, causing shaft 3 of the treatment bed to rotate along direction J6. Each shaft that can rotate independently horizontally is susceptible to collision. Therefore... Figure 4c The collision detection range shown for a single axis actually applies to every axis during detection. For example, Figure 4c Using J1 as the center axis of a collision detection distance for the treatment bed, ranges a and b are defined.

[0064] like Figure 4c As shown, the dark shaded area 'a' is the danger zone of the treatment bed, and the light shaded area 'b' is the secondary danger zone. The danger zone of the treatment bed is the robot body surrounding axis J1. If a person or object appears within this range, the treatment bed will brake rapidly. The secondary danger zone of the treatment bed is a 15cm radius around the robot body along axis J1, ensuring at least a safe anti-pinch space between the machine and any object. Both the danger zone and the secondary danger zone are defined with the edge of the axis being detected at that moment as the critical location. We define the edge of area b as the threshold of the secondary danger zone envelope; and the edge of area a as the threshold of the danger zone envelope. When the target falls outside areas a and b, it is considered to be in a safe zone, and the radiotherapy-related motion equipment operates normally. When the target falls in area b, it is considered to be at a secondary danger distance, and the radiotherapy-related motion equipment will decelerate. When the target falls in area a, it is considered to be at a danger distance, and the radiotherapy-related motion equipment will immediately stop operating.

[0065] (5) Division of dangerous and warning areas in the treatment head: such as Figure 4aAs shown, when using the treatment head as the distance detection reference, the type of sensor used to detect the target and its coverage area should be clearly defined first. These sensors should have high precision and high sensitivity to ensure accurate capture of the real-time position information of the target. For example, a safety-related laser distance sensor may be used. Figure 4a The range shown is the detection range of the sensor, which is divided into detection zones based on the probability of a collision. When the treatment head detects a person or object in the warning zone, it will issue a warning to alert staff to the potential collision risk; when it detects a person or object in the danger zone, it will issue an audible and visual alarm and apply emergency braking to the currently moving equipment.

[0066] (6) Calculation of the danger zone of the treatment head: such as Figure 4b As shown, ① Setting parameters: Let the maximum braking angle of the rotating frame be θ. max The maximum braking angle is the angle by which the rotating frame rotates from receiving the braking command to stopping. Let θ0 be the angle from the lowest point of the DR plate to the horizontal line of the axis of the rotating frame; let L1 be the distance from the lowest point of the DR plate to the axis of the rotating frame; let L2 be the horizontal component of the distance from the lowest point of the DR plate to the axis of the rotating frame; let L... max Let L be the vertical distance from the lowest point of the DR plate to the envelope of the danger zone. ② Solve the triangle to calculate L. max All four parameters above are known. Based on this, the first step is to calculate the length L of the side opposite θ0. θ0 The second step is to calculate θ0 - θ max Length L of opposite side θ0-θmax The third step is to calculate L. max =L θ0 -L θ0-θmax Fourth step, considering the safety threshold, the calculated L... max Multiply by a safety factor of 1.5 to get L. ref The fifth step will be L ref The distance from the danger zone to the treatment head and DR panel is used as the threshold of the envelope. The curve surrounding the treatment head and DR panel is then used as the envelope of the danger zone.

[0067] (7) PLC Controller: This invention uses a safety-grade PLC as the core of the judgment module. This PLC is specifically designed for safety-critical areas, possessing powerful logic processing capabilities and high reliability, playing a crucial role in the collision avoidance system for radiotherapy equipment. The PLC program adheres to safety standards, receives distance detection data in real time, and makes rapid and accurate judgments based on preset safety logic. Once a collision risk is detected, a safety response is triggered. The processing module consists of relays, photoelectric converters, etc., and responds quickly after receiving PLC instructions, controlling the lower-level devices. When the PLC issues a critical instruction, the processing module immediately activates the corresponding components to achieve precise control of the radiotherapy equipment, ensuring the system responds quickly to potential collision risks and protecting the safety of personnel and equipment.

[0068] Example 2: Pressure Comparison Collision Detection Based on Current Treatment Bed Pressure Comparison

[0069] The purpose of this invention is to provide a pressure detection device and its method of use that can effectively detect and prevent collision accidents involving radiotherapy equipment. The specific detection principle is as follows: Figure 5a As shown.

[0070] (1) Pressure Contrast Collision Detection Threshold and Mechanism: When a patient makes a slight turn or adjusts their posture on the treatment bed, the pressure generated changes accordingly. To accurately capture these dynamic changes, we focus not only on the real-time pressure value, but also on the new pressure generated at the moment of movement and the average value. The differences between them, and the relationship between this difference and The ratio between the two values. This ratio, a key indicator for assessing whether the action has caused abnormal force, is set within a range of ±10%. As long as the ratio remains within this range, the system considers the force state normal and treatment can continue. However, once the ratio exceeds the preset safety range, it is considered an abnormal force, and the treatment bed will immediately activate its safety mechanism and automatically stop operating to avoid potential harm to the patient.

[0071] Considering the possibility of external impacts or disturbances to the treatment bed during use, the system also possesses a keen monitoring capability. These external forces often manifest as rapid pressure changes, distinctly different from the pressure fluctuations (irregular, small, and short-lived) caused by normal turning over or adjusting posture. When the system detects a change in the pressure curve... In cases of significant mutations, it is quickly identified as a sign of external impact (or a patient turning over or adjusting their posture dramatically, which is usually unsafe during treatment), and emergency measures are taken immediately, including stopping all exercise equipment, to ensure the patient's absolute safety.

[0072] (2) Simulation of the pressure curve during treatment: Figure 5b The process of detecting the pressure on a treatment bed from normal to abnormal was simulated.

[0073] ① Time period 0~t1: This time period simulates the process of a patient lying on the treatment bed. Initially, the pressure is zero. When the patient lies on the treatment bed, a pressure generated by body weight distribution is immediately generated. This pressure fluctuates and increases as the patient fully lies down. After the patient is fully lying down, the average pressure monitored by the pressure detection module is marked as... This value has become an important benchmark for assessing the stress experienced by an individual. Until an average pressure value is formed. That is, before time t1, the pressure changes significantly within a short period of time. Therefore, the exercise device will not be allowed to move during the 0-t1 time period to prevent injury to the patient if the exercise device moves when the patient is not fully lying down.

[0074] ②T1~t2 time period: This time period simulates the process of the patient lying stably on the treatment bed and receiving treatment. Although the pressure fluctuates, the overall fluctuation is not significant. It can be assumed that no collision occurred during this period, and the sports equipment can operate normally during this time.

[0075] ③The time period from t2 to t3: During this period, we can clearly see a sudden change in pressure. We define this sudden increase or decrease in pressure within a short period as an abnormal pressure mutation. The equipment is highly likely to have experienced a collision or other abnormal situation (such as significant patient movement) during this time. When the pressure sensor detects such a mutation, it will stop the radiotherapy equipment from operating.

[0076] (3) Pressure Detection Module: The detection process for the above-mentioned simulated situation is completed by the pressure detection module. The pressure detection module is responsible for collecting the pressure values ​​of the target to be detected (such as the treatment bed or patient) in real time, paying particular attention to the squeezing pressure that may be generated between the equipment. The pressure detection device installed in the key parts continuously monitors the pressure status of the above-mentioned target, and once an abnormal pressure change (such as a sudden increase in pressure) is detected, the subsequent processing procedure is immediately triggered.

[0077] (4) PLC Controller: In the pressure detection device of this invention, the judgment module is integrated into a safety-type PLC, which is responsible for receiving pressure data in real time and performing rapid analysis and judgment according to a preset range. Once an abnormal pressure is detected, the PLC immediately activates the emergency braking program to quickly stop the equipment to avoid a collision accident. If the pressure is normal, it will not affect the operation of the radiotherapy equipment. The processing module is a key part of the collision avoidance system, containing actuators such as relays, used to receive emergency braking commands from the PLC and respond quickly. When the PLC determines that there is a risk of collision, the processing module quickly cuts off or adjusts the power of the equipment through the actuators to achieve immediate stop or deceleration. The execution process is highly reliable and fast, significantly reducing the possibility of collision accidents.

[0078] Example 3: VxWorks Spatial Position Comparison Collision Detection Based on Current Treatment Bed Spatial Position Comparison

[0079] The present invention further provides a real-time position comparison collision detection device for radiotherapy equipment. The device monitors the precise position of radiotherapy equipment (such as treatment bed, patient on treatment bed, treatment head, rotating gantry, etc.) in real time, and analyzes the real-time data using a pre-established real-time model to detect the current motion trajectory of the equipment and identify potential collision risks.

[0080] (1) Classification of spatial position comparison collision detection mechanisms: such as Figure 9 As shown, the spatial position comparison collision detection mechanism consists of a 3D modeling unit, a spatial position coordinate acquisition unit, and a real-time judgment unit. The 3D modeling unit is responsible for establishing a model of the safe movement zone of the treatment bed within the rotating frame and adjusting the model's position and angle in the program. The spatial position coordinate acquisition unit acquires the position and speed of the treatment bed and the rotating frame in real time. The real-time judgment unit determines whether the treatment bed will collide with its surrounding equipment in the next moment based on the acquired data; if a collision is imminent, it brakes the treatment bed in a timely manner.

[0081] (2) Establishment of spatial location comparison model:

[0082] ① 3D Model Establishment: A program written in C# is linked with the 3D modeling software. Each component of the model is named, ensuring that the program name matches the modeling component. The 3D modeling unit includes a coordinate system establishment unit, a treatment head angular position division unit, a first-level loose mesh division unit, and a second-level precise mesh division unit. The treatment head angular position division unit divides the treatment head 360 degrees into 72 equal parts, each part corresponding to a treatment head angle (e.g., ...). Figure 6a and Figure 6bThe diagram shows the mesh division at treatment head angles of 0° and 90°, respectively. To reduce the amount of data stored in the database and improve search speed, the first-level loose mesh division unit divides the active area of ​​the treatment bed away from the collision object into multiple first-level loose mesh regions based on the treatment head angle, and saves them to the database. The mesh is relatively sparse, and the amount of data is relatively small. The second-level precise mesh division unit divides the active area of ​​the treatment bed close to the collision object into multiple second-level precise mesh regions based on the treatment head angle, and saves them to the database. The mesh is relatively dense, and the amount of data is relatively large. The collision objects include the rotating gantry, treatment head, DR plate, and rolling floor. A three-dimensional rectangular coordinate system (x, y, z) is established with the center of the annulus on the front cross-section of the rotating gantry as the origin, serving as the basis for subsequent position calculations.

[0083] ②C# control of 3D model movement: such as Figure 2 As shown, the six independent rotation axes controlling the movement of the treatment bed give it six degrees of freedom, allowing for precise positioning of translation and rotation within specified motion constraints. A real-time collision detection database generation program written in C# is used to simulate the spatial position of the treatment bed under different postures and treatment head positions. The spatial coordinate range of the six-axis treatment bed is determined by the length of its robotic arm: the x-direction range is (-1500, 1500), the y-direction range is (-1200, 200), and the z-direction range is (-1500, 1300), all in mm, with units of 100 mm in a first-level loose grid and 10 mm in a second-level fine grid. The rotation angles (Euler angles) of the treatment bed range as follows: yaw angle ψ is -95° to +95°, pitch angle θ is -5° to +5°, and roll angle φ is -5° to +5°, with units of 5°. The rotating frame has a rotation range of -185° to +185°, with each 5° increment serving as a simulation step, simulating a total of 72 different positions.

[0084] ③ Data Recording and Storage: Records eight-dimensional information (including rotation angle of the rotating gantry, treatment bed roll, treatment bed pitch, treatment bed yaw, treatment bed x, treatment bed y, treatment bed z, and whether a collision occurred) of the treatment bed in different postures and treatment head positions. The total data volume can reach hundreds of millions of rows. A high-performance NoSQL database is used to store the recorded data to ensure query speed.

[0085] (3) VxWorks obtains spatial coordinates and performs collision detection:

[0086] ① Obtaining Spatial Coordinates: After the database is established, communication between the various components of the radiotherapy equipment and the PowerPC architecture hardware board equipped with the VxWorks system is established via HTTP and Snap7 protocols to monitor the moving parts of the radiotherapy equipment in real time. This determines the current position of the treatment head, the posture of the treatment table, and performs positional collision detection to ensure optimal real-time performance and low latency.

[0087] ② Collision Detection Program: The real-time position comparison collision detection unit's program runs on a hardware board equipped with the VxWorks system. The VxWorks system sends instructions to the treatment bed system via the HTTP protocol, reads the treatment bed's spatial coordinates and velocity information, and acquires the angle, direction, and speed of the rotating frame via the Snap7 protocol. The collision detection program calculates the next position based on the treatment bed's current position and movement speed, compares it with a pre-organized lookup table, and if a collision is likely to occur, immediately transmits an electrical signal to the PLC to brake the relevant motion equipment.

[0088] (4) Collision prediction logic for the treatment bed: The movement of the treatment bed involves only a single coordinate change in one direction. Knowing the direction and speed of the movement, the position, orientation, and collision situation at the next moment can be predicted. For example... Figure 8 As shown, if the treatment bed is currently moving in the x-direction at a speed of 2 centimeters per second and its current position is known, a database search indicates that if it continues at the current speed, it will collide with the equipment in the next moment. In practical applications, by real-time detection of the treatment bed's current position and motion state, the processing module will immediately brake the relevant moving equipment when a collision is predicted in the next moment.

[0089] (5) PLC Controller: The processing module of this invention adopts a safety-functional PLC. The safety-functional PLC integrates a dual-redundant power supply module, a dedicated safety CPU, and a dedicated safety I / O module in its overall architecture. Its main responsibility is to implement safety interlock control for the subsystem. This PLC can monitor the operation status of the subsystem in real time and connect to external devices and the VxWorks system via digital hardwired signals. The PLC undertakes the safety interlock control of the entire radiotherapy equipment. If it determines that a collision is about to occur based on the current position and motion information, it will immediately activate the preset emergency braking system for the equipment receiving the braking command, rapidly decelerating until it comes to a complete stop, thereby avoiding a collision. After braking is completed, the system will automatically perform a reset operation, check the status of all equipment, and ensure that everything returns to normal before continuing the treatment process.

Claims

1. A collision avoidance system for radiotherapy devices based on multiple detection methods, characterized in that: The system, from top to bottom, includes: a host computer for displaying alarm information; a PLC controller connected to the host computer above and the motion equipment below; the motion equipment controlled by the PLC controller; and a collision detection unit that provides collision detection information to the PLC controller. The motion equipment includes a rotating frame, a treatment bed, and other motion equipment related to the treatment head. The treatment bed is a robotic treatment bed, and the envelope of the treatment bed includes the envelope of the treatment bed itself and the envelope of the patient on the treatment bed. The collision detection unit includes: a distance comparison collision detection subunit based on the current distance to the treatment bed, a pressure comparison collision detection subunit based on the current pressure of the treatment bed, and a VxWorks spatial position comparison collision detection subunit based on the current spatial position of the treatment bed; The distance comparison collision detection subunit includes detectors, laser distance sensors, multi-level warning zones divided around the treatment head, DR panel, and treatment bed, as well as a warning zone boundary calculation unit. The pressure comparison collision detection subunit includes a pressure sensor installed on the treatment bed and a pressure comparison module for comparing whether the current pressure is normal or not. The VxWorks spatial position comparison collision detection subunit includes programming software and 3D modeling software, a position comparison collision detection database, and a VxWorks-based position information acquisition and calculation module. The programming software and 3D modeling software save the 3D model to the position comparison collision detection database. The VxWorks-based position information acquisition and calculation module acquires 3D modeling information from the position comparison collision detection database and position information of relevant moving equipment, including the treatment bed and rotating frame. It then calculates these two types of information in VxWorks to obtain the collision prediction result for the next moment and reports it to the PLC. In dangerous situations, the PLC will quickly shut down or brake the relevant moving equipment. The 3D model is a model of the safe movement range of the treatment bed.

2. The collision avoidance system for radiotherapy devices based on multiple detection methods according to claim 1, characterized in that: The multi-level warning zone of the distance comparison collision detection subunit includes a treatment head warning zone and a treatment head danger zone based on the treatment head; zones a and b centered on the treatment bed; the treatment head warning zone surrounds the treatment head danger zone; the shape of the treatment head warning zone is approximately trapezoidal, and the shape of the treatment head danger zone is approximately an envelope formed by the treatment head and the DR panel when deployed; zone a is the treatment bed danger zone, and zone b is the treatment bed secondary danger zone; when the distance between the treatment bed and the treatment head enters the treatment head warning zone, the system alarms; when the distance between a person or object and the treatment head enters the treatment head danger zone, the motion equipment of each radiotherapy facility stops moving; the motion equipment includes, but is not limited to, rotating gantry, treatment bed, DR panel, and range shifter on the treatment head; when a person or object is detected in zone b around the treatment bed, the system issues a warning, and the treatment bed decelerates; when a person or object is detected in zone a around the treatment bed, the system issues an audible and visual alarm, and the treatment bed brakes.

3. The collision avoidance system for radiotherapy devices based on multiple detection methods according to claim 1, characterized in that: The warning zone boundary calculation unit of the distance comparison collision detection subunit calculates the envelope boundary of the danger zone of the treatment head based on the dimensions of the rotating gantry, treatment head, DR plate, and the maximum braking angle of the rotating gantry.

4. The collision avoidance system for radiotherapy devices based on multiple detection methods according to claim 1, characterized in that: The pressure comparison module of the pressure comparison collision detection subunit divides the pressure detection process into four time points: 0, t1, t2, and t3. From time 0 to time t1, this time period simulates the process of a patient lying on the treatment bed. From time t1 to time t2, this time period simulates the process of a patient lying steadily on the treatment bed receiving treatment. Although the curve fluctuates during this stage, it eventually flattens out. From time t2 to time t3, this time period simulates the process of a sudden increase or decrease in pressure. A sudden change is a sudden increase or decrease in pressure within a short period of time. This process is defined as the pressure abnormal change stage. The equipment is very likely to have experienced a collision or other abnormal situation during this time period. When the pressure comparison unit detects such a change, it will stop the operation of the radiotherapy equipment.

5. The collision avoidance system for radiotherapy devices based on multiple detection methods according to claim 1, characterized in that: The VxWorks spatial position comparison collision detection subunit includes a 3D modeling module, a VxWorks spatial position coordinate acquisition module, and a real-time judgment module. The 3D modeling module uses programming software and 3D modeling software to model the safe movement zone of the treatment bed within the rotating frame, and repeatedly adjusts the position and angle of the established safe movement zone model in the program, saving the adjusted model to the position comparison collision detection database. The VxWorks spatial position coordinate acquisition module uses the VxWorks system to obtain the current position and movement speed of the treatment bed and the rotating frame in real time. The real-time judgment module determines whether the treatment bed will collide with its surrounding equipment at the next moment based on the data obtained by the spatial position coordinate unit, and sends the judgment information to the PLC controller. If a collision is determined to occur, the PLC controller will brake the treatment bed in time.

6. The collision avoidance system for radiotherapy devices based on multiple detection methods according to claim 5, characterized in that: The 3D modeling module includes a coordinate system establishment submodule, a treatment head angular position division submodule, a first-level loose mesh division submodule, and a second-level precise mesh division submodule. The treatment head angular position division submodule divides the treatment head 360 degrees into several equal parts, each corresponding to an angle of the treatment head. The first-level loose mesh division submodule, based on the current angle of the treatment head, defines the active areas of the treatment bed away from collision objects or the danger zone of the treatment head as multiple first-level loose mesh regions, and saves these regions to the database. These multiple first-level loose mesh regions have relatively sparse meshes and relatively large mesh sizes, with each loose mesh serving as a detection point. The second-level precise mesh division submodule, based on the current angle of the treatment head, defines the active areas of the treatment bed close to collision objects or the danger zone of the treatment head as multiple second-level precise mesh regions, and saves these multiple second-level precise mesh regions to the database. These multiple second-level precise mesh regions have relatively dense meshes and relatively small mesh sizes, with each second-level precise mesh division submodule serving as a detection point. The collision objects include, but are not limited to, rotating frames around the treatment bed, treatment heads, DR flat panels, and rolling floors.

7. The collision avoidance system for radiotherapy devices based on multiple detection methods according to claim 6, characterized in that: The coordinate system establishment submodule establishes a three-dimensional rectangular coordinate system (X,Y,Z) with the center of the circular ring of the forward cross section of the rotating frame as the origin. This coordinate system is the basis for all subsequent position calculations.

8. The collision avoidance system for radiotherapy devices based on multiple detection methods according to claim 6, characterized in that: The VxWorks spatial coordinate acquisition module acquires the three-dimensional coordinates of the treatment bed, as well as the Euler angles (rX, rY, rZ) of the treatment bed and the moving speed and direction of the treatment bed; it also acquires the angle of the rotating frame and its rotation speed and direction.

9. The collision avoidance system for radiotherapy devices based on multiple detection methods according to claim 5, characterized in that: The real-time judgment module determines whether the treatment bed and its surrounding equipment will collide at the next moment based on the current safe movement range model of the treatment bed, the three-dimensional coordinates of the treatment bed, the Euler angles (rX, rY, rZ) of the treatment bed, the movement speed and direction of the treatment bed, the angle of the rotating frame, and the rotation speed and direction. If it determines that a collision is about to occur, it reports to the PLC controller, and the PLC controller will brake the treatment bed in time.

10. A radiotherapy device collision avoidance system based on multiple detection methods according to claim 6, characterized in that: The treatment head angular position division submodule is used to divide the treatment head 360 degrees into several equal parts, which include 72 parts, with each part spaced 5° apart; the multiple primary loose grid regions include 72 primary loose grid regions; the multiple secondary precise grid regions include 72 secondary precise grid regions.

Images