Robot out-of-bound behavior detection and alarm method and system

Abstract

The application relates to the technical field of autonomous navigation, and discloses a robot border-crossing behavior detection and alarm method and system, the method comprising the following steps: constructing a global constraint domain based on static boundary information, generating a dynamic permission domain along a task path according to robot motion parameters, expanding the dynamic permission domain outside the global constraint domain to obtain a sensing buffer zone, and binding a spatial coordinate and an expansion width. A boundary touch state identifier is determined according to the coincidence degree of the sensing buffer zone and the dynamic permission domain, a region-exceeding state identifier is generated by screening non-coverage grids of the global constraint domain through the dynamic permission domain, and finally, the speed vector is adjusted in combination with the two identifiers, and a hierarchical disposal instruction is output; the application can improve the efficiency of robot border-crossing behavior detection and alarm.

Description

Technical Field

[0001] This invention relates to the field of autonomous navigation technology, and in particular to a method and system for detecting and alarming robot boundary crossing behavior. Background Technology

[0002] In existing technologies, robot boundary violation detection and alarm methods typically rely on a single static safety boundary or a simple distance threshold for judgment. These methods treat the fixed boundary of the entire work environment as the sole constraint, failing to dynamically adjust the permissible area based on the robot's real-time movement. When the robot moves at high speeds or with significant acceleration, the static boundary cannot predict the risk of boundary violation in advance, causing the system to issue an alarm only after the robot has already crossed the boundary or is about to collide, resulting in severely insufficient warning time. Furthermore, the static boundary method treats the entire constraint domain as having the same safety level, failing to distinguish the gradual approach between the robot's current position and the boundary. When the robot is operating normally near the boundary, it is prone to triggering frequent false alarms, severely impacting the continuity and efficiency of the robot's work.

[0003] Furthermore, existing technologies lack a tiered handling mechanism for different degrees of boundary violation severity. Most solutions only set a single boundary violation threshold, and once the robot is detected to have crossed this threshold, an emergency stop is executed. This binary approach of stopping or freezing without decay cannot cope with the diverse needs of scenarios ranging from minor touch warnings to emergency braking for severe boundary violations. When the robot only slightly approaches the boundary but has not yet posed a substantial danger, existing technologies cannot output deceleration commands to smoothly reduce the risk, and can only adopt the same emergency stop treatment as for severe boundary violations, causing unnecessary production interruptions. At the same time, existing technologies mostly use continuous geometric operations for coverage detection within the global constraint domain, which involves large computational loads and poor real-time performance, making it difficult to achieve millisecond-level alarm responses in high-speed robot systems, resulting in alarm handling lagging behind the robot's actual boundary violation behavior. Summary of the Invention

[0004] This invention provides a method and system for detecting and alarming robot boundary crossing behavior, the main purpose of which is to solve the problem of low efficiency in detecting and alarming robot boundary crossing behavior.

[0005] To achieve the above objectives, the present invention provides a robot boundary crossing detection and alarm method, comprising: S1. Construct the global constraint domain of the target process based on the static boundary information of the robot's working scene during the target process; S2. Starting from the robot's current position in the target process, extend the robot's motion parameters along the path of the task to be executed to obtain the dynamic permission domain of the target process. S3. Using the current expansion width of the dynamic permission domain as the outward expansion width value, generate an outward expansion sensing band along the boundary line of the global constraint domain, and bind the spatial coordinate data of the area covered by the outward expansion sensing band with the current expansion width to obtain the sensing buffer of the target process. S4. Determine the boundary touch state identifier of the target process based on the overlap between the real-time spatial coordinates of the sensing buffer and the dynamic permission domain. S5. Based on the dynamic permission domain, filter the non-covered meshes in the inner domain mesh set of the global constraint domain, and encapsulate the inner domain number of the non-covered meshes as the region out-of-bounds status identifier of the target process; S6. Based on the indication type of the boundary touch status indicator and the area exceed status indicator, the current velocity vector of the dynamic permission domain is amplitude-adjusted to obtain the hierarchical handling instruction of the target process.

[0006] In a preferred embodiment, constructing the global constraint domain of the target process based on the static boundary information of the robot's working scenario during the target process includes: Extract the maximum permissible boundary coordinates of the robot's working scene during the target process to obtain an ordered set of geographic coordinates for the target process; Connect the coordinate points in the ordered geographic coordinate point set in sequence, and then perform a closed-loop splicing between the last coordinate point and the first coordinate point to obtain the closed boundary loop of the target process. The global constraint domain of the target process is obtained by filling the internal space enclosed by the closed boundary loop.

[0007] In a preferred embodiment, the step of extending the dynamic permission domain of the target process from the robot's current position along the path of the task to be executed based on the robot's motion parameters to obtain the dynamic permission domain of the target process includes: The target path points in the task path to be executed are connected sequentially to obtain the path curve of the target process, and the path segment starting from the current position of the robot along the path direction is used as the extended skeleton of the target process on the path curve. The dynamic broadening of the target process is calculated based on the motion parameters and the baseline broadening of the target process. Using the extended skeleton as the center line and the dynamic widening range as the lateral expansion radius, the space is expanded along both sides and the end point of the extended skeleton to obtain the dynamic permission domain of the target process.

[0008] In a preferred embodiment, the formula for calculating the dynamic widening amplitude includes: in, The dynamic widening range, The reference width expansion, It is a natural constant. This is the global scaling factor. For the speed term, the weighting coefficient is... This represents the robot's current speed. As the speed reference value, For speed sensitivity index, It is the hyperbolic tangent function. This represents the robot's current acceleration value. As the reference value for acceleration, For the acceleration term, For acceleration sensitivity index, It is the natural logarithm function. This is the directional consistency factor.

[0009] In a preferred embodiment, the step of spatially expanding along both sides and the endpoint of the extended skeleton, using the dynamic widening amplitude as the lateral expansion radius, with the extended skeleton as the centerline, to obtain the dynamic permission domain of the target process includes: Starting from the starting point of the extended skeleton, and moving against the direction of the extended skeleton, a semi-circular arc is extended with the dynamic widening range as the expansion radius to obtain the starting closed area of ​​the target process. Along both sides of the extended skeleton, with the dynamic widening amplitude as the vertical distance, side line segments parallel to the extended skeleton are generated; Centered on the endpoint of the extended skeleton, the dynamic widening amplitude is used as the radial distance, and an outer envelope is formed along the forward direction of the extended skeleton to obtain the endpoint extension region of the target process. Align the boundaries of the starting closed region, the side line segment, and the ending extended region to obtain the dynamic permission domain of the target process.

[0010] In a preferred embodiment, the step of generating an outward-expanding sensing band along the boundary line of the global constraint domain, using the current expansion width of the dynamic permission domain as the outward expansion width value, and binding the spatial coordinate data of the area covered by the outward-expanding sensing band with the current expansion width to obtain the sensing buffer of the target process, includes: The coordinate points on the boundary line of the global constraint domain are vertically offset outward from the global constraint domain, and the offset distance is the value of the outward expansion width, to obtain the outward expansion coordinate points of the target process; The outward coordinate points are connected sequentially to form an outward boundary line, and the strip-shaped region between the boundary line of the global constraint domain and the outward boundary line is taken as the outward sensing zone of the target process. The spatial coordinate data within the extended sensing band is associated with and stored with the extended width value to obtain the sensing buffer of the target process.

[0011] In a preferred embodiment, determining the boundary touch state identifier of the target process based on the overlap between the real-time spatial coordinates of the sensing buffer and the dynamic permission domain includes: By comparing the spatial coordinates of the sensing buffer and the dynamic permission domain, the coincident coordinates of the target process are obtained; The boundary touch status identifier of the target process is generated based on the address markers of the overlapping coordinate points.

[0012] In a preferred embodiment, the step of filtering non-covered meshes in the inner domain mesh set within the global constraint domain based on the dynamic permission domain, and encapsulating the inner domain number of the non-covered meshes as a region overflow status identifier for the target process, includes: The internal space of the global constraint domain is divided into grids to obtain the internal grid set of the target process; Based on the real-time spatial coordinate data within the spatial range covered by the dynamic permission domain, the spatial range of the inner domain grid set is differentially divided to obtain the non-covered grid of the target process; The inner domain number of the non-covered grid is combined with the region overflow marker to obtain the region overflow status identifier of the target process.

[0013] In a preferred embodiment, the step of performing amplitude modulation processing on the current velocity vector of the dynamic permissioned domain based on the indication type of the boundary touch status indicator and the area exceedance status indicator to obtain the hierarchical handling instruction for the target process includes: Based on the coincident coordinate point data corresponding to the boundary touch state identifier, the magnitude of the current velocity vector in the dynamic permission domain is attenuated to obtain the deceleration command of the target process; Based on the out-of-area status identifier corresponding to the out-of-area number, the magnitude of the current velocity vector in the dynamic permission domain is set to zero to obtain the emergency stop command for the target process. The deceleration command and the emergency stop command are combined to obtain the graded handling command for the target process.

[0014] To address the above problems, the present invention also provides a robot boundary crossing detection and alarm system, the system comprising: The global constraint domain module constructs the global constraint domain of the target process based on the static boundary information of the robot's working scenario during the target process; The dynamic permission domain module starts from the robot's current position in the target process and extends it along the path of the robot's task to be executed based on the robot's motion parameters to obtain the dynamic permission domain of the target process. The sensing buffer module uses the current expansion width of the dynamic permission domain as the expansion width value, generates an outward expansion sensing band along the boundary line of the global constraint domain, and binds the spatial coordinate data of the area covered by the outward expansion sensing band with the current expansion width to obtain the sensing buffer of the target process. The boundary touch state identification module determines the boundary touch state identification of the target process based on the overlap of the real-time spatial coordinates of the sensing buffer and the dynamic permission domain. The region out-of-bounds status identification module filters the non-covered meshes in the inner domain mesh set of the global constraint domain based on the dynamic permission domain, and encapsulates the inner domain number of the non-covered meshes as the region out-of-bounds status identification of the target process;

[0015] The tiered handling instruction module performs amplitude modulation processing on the current velocity vector of the dynamic permissioned domain based on the indication types of the boundary touch status indicator and the area exceedance status indicator, thereby obtaining the tiered handling instruction for the target process. Compared with the prior art, the present invention has the following beneficial effects:

[0016] 1. This technology constructs a three-layer spatial structure consisting of a global constraint domain, a dynamic permitted domain, and a sensing buffer zone. Based on the robot's real-time motion parameters, it dynamically adjusts the width of the dynamic permitted domain, enabling the system to automatically expand or shrink the permitted area according to the robot's current speed and acceleration. This provides a greater safety margin during high-speed movement and avoids unnecessary spatial redundancy during low-speed movement, achieving adaptive and proactive early warning of boundary violation risks. Simultaneously, by comparing the spatial coordinates of the sensing buffer zone and the dynamic permitted domain point-by-point to generate coincident coordinate points, and using a grid differencing method to accurately extract the uncovered mesh within the global constraint domain, this technology can simultaneously detect boundary touch behavior and area overrun behavior. This ensures that violations ranging from slight approach to severe boundary violations are accurately captured, greatly improving the comprehensiveness of detection and positioning accuracy.

[0017] 2. This technology further performs modulus attenuation or zeroing processing on the current velocity vector of the dynamic permitted domain based on different indication types of boundary touch status and area exceedance status. It combines deceleration and emergency stop commands into hierarchical handling instructions, enabling the system to smoothly decelerate to maintain operational continuity when minor boundary touch is detected, and to immediately stop for absolute safety when area exceedance is detected, achieving a complete hierarchical response chain from warning to braking. Furthermore, by meshing the internal space of the global constraint domain and using numbered labels for differential operations, this technology transforms complex spatial geometric judgments into efficient mesh index comparisons, significantly improving the computational efficiency of real-time detection and response, ensuring that the system can still complete millisecond-level alarm handling even during high-speed robot movement. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a robot boundary crossing detection and alarm method according to an embodiment of the present invention.

[0019] Figure 2 This is a functional block diagram of a robot boundary crossing detection and alarm system provided in an embodiment of the present invention;

[0020] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0022] This application provides a method for detecting and alarming robot boundary crossing behavior. The executing entity of the robot boundary crossing behavior detection and alarm method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application embodiment: a server, a terminal, etc. In other words, the robot boundary crossing behavior detection and alarm method can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0023] Reference Figure 1 The diagram shown is a flowchart illustrating a robot boundary crossing detection and alarm method according to an embodiment of the present invention. In this embodiment, the robot boundary crossing detection and alarm method includes: In this embodiment of the invention, when constructing the global constraint domain of the target process based on the static boundary information of the robot's working scene during the target process, it is specifically used for: Extract the maximum permissible boundary coordinates of the robot's working scene during the target process to obtain an ordered set of geographic coordinates for the target process; Connect the coordinate points in the ordered geographic coordinate point set in sequence, and then perform a closed-loop splicing between the last coordinate point and the first coordinate point to obtain the closed boundary loop of the target process. The global constraint domain of the target process is obtained by filling the internal space enclosed by the closed boundary loop.

[0024] Specifically, all boundary points constituting physical fences or virtual walls are collected one by one from the static boundary information of the robot's working scene. These boundary points are arranged in clockwise or counterclockwise order according to their spatial location. After removing all non-extreme points, the coordinate point farthest from the center of the scene in each direction is retained.

[0025] Specifically, the first coordinate point in the ordered set of geographic coordinate points is taken as the starting point, and each subsequent coordinate point is taken out in sequence. The current point is connected to the next point using a straight line segment. When connecting, all intermediate points on the line segment are filled pixel by pixel according to the difference in Cartesian coordinates between the two points.

[0026] Specifically, the closed boundary ring is used as the boundary condition, and the scan line filling method is used to determine all integer coordinate points inside the ring. The specific method is to start from the bottom row of the smallest bounding rectangle of the ring, scan row by row to the right, record all intersection points in each row that intersect with the boundary ring, and sort these intersection points in ascending order of their horizontal coordinates.

[0027] Furthermore, these coordinate points are arranged clockwise from due north according to their actual geographical location, forming a sequence that is connected end to end, namely the ordered set of geographical coordinate points.

[0028] Furthermore, when connecting to the last coordinate point, the last coordinate point is connected to the first coordinate point of the beginning with the same straight line segment, so that the entire path forms a closed ring profile without any openings. This profile is the closed boundary loop.

[0029] Furthermore, all coordinate points between any two adjacent intersections are marked as internal points. This process is repeated until the top row is reached. All the coordinate points marked as internal points, together with the coordinate points on the boundary ring, constitute a complete two-dimensional region, which is the global constraint domain.

[0030] In summary, by extracting the maximum permissible boundary coordinates and arranging them in order, the physical boundary of the robot's working scene can be transformed into a discrete but ordered spatial coordinate sequence. This eliminates the ambiguity caused by the disordered acquisition of boundary points and provides an accurate and complete geometric basis for the subsequent construction of closed boundary loops. At the same time, the ordered set of coordinate points facilitates linear traversal and fast retrieval by the computer, improving the efficiency of boundary processing.

[0031] In summary, by connecting discrete coordinate points sequentially to form a closed loop, a continuous closed boundary is constructed. This closed boundary loop precisely defines the outer contour of the maximum range of robot movement, avoiding missed detection areas or misjudgments due to non-closed boundaries. At the same time, the closed loop structure clearly separates the internal space from the external space, providing clear closed curve constraints for subsequent full-domain filling operations.

[0032] In summary, by filling all spatial points within the closed boundary loop without omission, a complete and continuous global constraint domain is generated. This global constraint domain directly corresponds to the entire geographic space where the robot is permitted to move in the actual working scenario. This allows any subsequent determination of the spatial inclusion relationship between the robot's position and the global constraint domain to be completed through simple coordinate comparison, without the need to repeatedly calculate the boundary curve. At the same time, the global filling eliminates internal voids or gaps, ensuring the integrity and consistency of the constraints and providing reliable basic data for the differential operation between the dynamic permitted domain and the global constraint domain.

[0033] In this embodiment of the invention, the step of extending the dynamic permission domain of the target process along the path of the robot's task to be executed based on the robot's motion parameters, starting from the robot's current position in the target process, is specifically used for: The target path points in the task path to be executed are connected sequentially to obtain the path curve of the target process, and the path segment starting from the current position of the robot along the path direction is used as the extended skeleton of the target process on the path curve. The dynamic broadening of the target process is calculated based on the motion parameters and the baseline broadening of the target process. Using the extended skeleton as the center line and the dynamic widening range as the lateral expansion radius, the space is expanded along both sides and the end point of the extended skeleton to obtain the dynamic permission domain of the target process.

[0034] Specifically, each target path point in the task path to be executed is connected sequentially with straight line segments according to their order on the path, thereby forming a continuous spatial curve, which is named the path curve of the target process.

[0035] Specifically, firstly, the robot's current instantaneous velocity and instantaneous acceleration values ​​are obtained, along with a direction consistency factor reflecting the consistency between the robot's current motion direction and the extended skeleton direction. Then, the velocity contribution value and acceleration contribution value are calculated respectively. The velocity contribution value is a relative velocity ratio obtained by comparing the robot's current velocity value with a velocity reference value and then nonlinearly scaling it according to a velocity sensitivity index, and then multiplying this ratio by a velocity weighting coefficient. The acceleration contribution value is a relative acceleration ratio obtained by comparing the robot's current acceleration value with an acceleration reference value and then nonlinearly scaling it according to an acceleration sensitivity index, and then multiplying this ratio by an acceleration weighting coefficient.

[0036] Specifically, starting from the starting point of the extended skeleton, a semicircular arc is drawn against the direction of the extended skeleton's movement. The radius of this semicircular arc is equal to the dynamic widening amplitude. This semicircular arc connects with the starting point of the extended skeleton to form a semicircular closed area, which is named the starting closed area of ​​the target process. Then, on the left and right sides of the extended skeleton, two parallel lines extend outward in a direction perpendicular to the extended skeleton. The vertical distance between each parallel line and the extended skeleton is equal to the dynamic widening amplitude. These two parallel lines extend from the edge of the starting closed area to the end point of the extended skeleton. These two parallel line segments are named side line segments.

[0037] Furthermore, the robot's current position is located on the path curve, and a path segment is intercepted from the current position along the direction of the path. The endpoint of this path segment is the next target path point on the path curve after the current position, thus obtaining a directional line segment, which is named the extended skeleton of the target process.

[0038] Furthermore, the result of taking the natural logarithm of the relative velocity ratio needs to be weighted; then the contribution values ​​of the velocity term and the acceleration term are added together, multiplied by a global scaling factor, and then an exponential operation is performed with the natural constant as the base and the product result as the exponent to obtain an exponential scaling factor; finally, the baseline broadening amplitude is multiplied by the exponential scaling factor, and then multiplied by the directional consistency factor, and the result is the dynamic broadening amplitude of the target process.

[0039] Furthermore, taking the endpoint of the extended skeleton as the center point, a semi-circular envelope is extended outward along the forward direction of the extended skeleton, with the dynamic widening amplitude as the radial distance. This envelope connects with the end of the side line segment to form a complete endpoint region, which is named the endpoint extension region of the target process. Finally, the boundaries of the starting point closed region, the side line segment, and the endpoint extension region are sequentially aligned and spliced ​​to seamlessly connect these regions into a complete spatial region, which is the dynamic permissioned domain of the target process.

[0040] In summary, by connecting discrete target path points into a continuous path curve, the geometric direction and shape information of the task path can be completely preserved, avoiding directional ambiguity caused by jumps between path points. Then, a path segment starting from the robot's current position and moving in the forward direction is extracted from this path curve as an extended skeleton. This extended skeleton closely follows the robot's real-time position and the task path to be executed, ensuring that the dynamic permission domain is always constructed around the path segment where the robot is currently located. It will not include areas that the robot has already passed behind to reduce redundant calculations, nor will it miss areas that it is about to enter to ensure safe prediction.

[0041] In summary, the robot's real-time motion parameters, including current velocity and acceleration values, are nonlinearly fused with a pre-set baseline width range. This allows the calculated dynamic width range to adaptively adjust in real time according to the robot's motion state. When the robot moves at high speed or accelerates rapidly, the dynamic width range automatically expands, creating a greater safety margin in front of the robot and providing early warning of boundary crossing risks. When the robot moves at low speed or decelerates, the dynamic width range automatically shrinks, avoiding excessive occupation of workspace or triggering unnecessary alarms due to an excessively large permissible domain, thus achieving a dynamic balance between safety and work efficiency.

[0042] In summary, by using the extended skeleton as the reference centerline and the dynamic widening range as the lateral expansion radius to uniformly expand the space to both sides and the endpoint of the skeleton, a pipe-shaped permitted area with a defined width along the robot's forward path can be generated. This dynamic permitted area tightly encloses the path segment that the robot is about to traverse, and its width changes in real time with the robot's motion state. This structure ensures that any deviation from the path will first enter the boundary range of the dynamic permitted area, thus being detected by the system in a timely manner. At the same time, since the dynamic permitted area only covers a limited section of the path in front of the robot, rather than the entire working scene, the storage of spatial coordinate data and the computational load for real-time comparison are greatly reduced, improving the response speed of the detection and alarm system.

[0043] In this embodiment of the invention, the calculation formula for the dynamic widening amplitude is specifically used for: in, The dynamic widening range, The reference width expansion, It is a natural constant. This is the global scaling factor. For the speed term, the weighting coefficient is... This represents the robot's current speed. As the speed reference value, For speed sensitivity index, It is the hyperbolic tangent function. This represents the robot's current acceleration value. As the reference value for acceleration, For the acceleration term, For acceleration sensitivity index, It is the natural logarithm function. This is the directional consistency factor.

[0044] Specifically, the baseline expansion is preset by the system designer according to the safety distance requirements of the robot's working scenario and stored in the configuration file. The robot's current speed and current acceleration values ​​are collected in real time by the inertial measurement unit and encoder installed on the robot body and then processed by Kalman filtering. The speed and acceleration baseline values ​​are obtained by calibrating the maximum stable speed and maximum safe acceleration values ​​measured during standard motion testing of the robot. The speed and acceleration weight coefficients, as well as the speed sensitivity index and acceleration sensitivity index, are optimized and determined by performing multivariate regression analysis on the risk levels of exceeding the limits under different motion states in offline simulation experiments. The global scaling factor is manually set by the safety administrator according to the danger level of the working scenario. The direction consistency factor is calculated from the cosine of the angle between the robot's current motion direction and the extended skeleton direction. The natural constant, hyperbolic tangent function, and natural logarithm function are all predefined mathematical operations in the mathematical library.

[0045] Furthermore, by performing nonlinear transformations on the robot's real-time velocity and acceleration values, weighting and combining them, multiplying by a global scaling factor and using it as the exponent of the natural constant, and finally multiplying by the directional consistency factor and the baseline widening amplitude, the dynamic widening amplitude at the current moment is calculated. This dynamic widening amplitude is used to determine the lateral expansion radius of the dynamic admissibility domain, so that the dynamic admissibility domain automatically widens during high-speed or rapid acceleration to provide early warning, and automatically narrows during low-speed or deceleration to improve work efficiency. At the same time, when the robot's movement direction deviates from the predetermined path, the directional consistency factor will reduce the dynamic widening amplitude to avoid ineffective buffer expansion.

[0046] In general, as the robot's current velocity increases relative to the velocity reference value, the velocity contribution increases exponentially, leading to a monotonically increasing dynamic width. As the robot's current acceleration increases relative to the acceleration reference value, the hyperbolic tangent function output approaches one from zero, and the acceleration contribution also increases, similarly increasing the dynamic width. When both the robot's velocity and acceleration increase simultaneously, the summation term within the exponent further increases, and the dynamic width expands exponentially. When both the robot's current velocity and acceleration are zero, the summation term within the exponent is zero, and the zero power of the natural constant equals one. At this point, the dynamic width equals the reference width multiplied by the orientation consistency factor. When the orientation consistency factor decreases from one to zero, the dynamic width shrinks proportionally to zero, indicating that no widening is needed when the motion direction completely deviates from the path.

[0047] In this embodiment of the invention, when the dynamic widening range is used as the lateral expansion radius and the extended skeleton is used as the center line to expand the space along both sides and the end point of the extended skeleton to obtain the dynamic permission domain of the target process, it is specifically used for: Starting from the starting point of the extended skeleton, and moving against the direction of the extended skeleton, a semi-circular arc is extended with the dynamic widening range as the expansion radius to obtain the starting closed area of ​​the target process. Along both sides of the extended skeleton, with the dynamic widening amplitude as the vertical distance, side line segments parallel to the extended skeleton are generated; Centered on the endpoint of the extended skeleton, the dynamic widening amplitude is used as the radial distance, and an outer envelope is formed along the forward direction of the extended skeleton to obtain the endpoint extension region of the target process. Align the boundaries of the starting closed region, the side line segment, and the ending extended region to obtain the dynamic permission domain of the target process.

[0048] Specifically, first, the spatial coordinates of the starting point of the extended skeleton are identified. This starting point is the foremost endpoint of the extended skeleton. Then, the tangent direction of the extended skeleton at the starting point is determined. This tangent direction is the forward direction of the extended skeleton. Next, the opposite direction of the forward direction is taken as the orientation of the semicircle. With the starting point as the center and the value of the dynamic widening amplitude as the radius, a semicircle is drawn in a plane perpendicular to the forward direction of the extended skeleton.

[0049] Specifically, each discrete point on the extended skeleton is traversed. For each discrete point, the normal vector perpendicular to the forward direction of the extended skeleton is calculated, resulting in two unit normal vectors in opposite directions, pointing to the left and right sides of the extended skeleton, respectively. Then, starting from the discrete point, the distance is moved along the direction of the left normal vector by a distance equal to the value of the dynamic widening amplitude, resulting in an offset point on the left. Similarly, the same distance is moved along the direction of the right normal vector to obtain an offset point on the right.

[0050] Specifically, the spatial coordinates of the endpoint of the extended skeleton are located. This endpoint is the last end point of the extended skeleton. Then, the tangent direction of the extended skeleton at the endpoint is determined, which is the forward direction of the extended skeleton. Next, with the endpoint as the center and the value of the dynamic widening amplitude as the radius, a complete semicircular arc is drawn in a plane perpendicular to the forward direction of the extended skeleton. The arc of this semicircular arc starts from the left vertical direction of the forward direction, passes around the front of the endpoint, and extends to the right vertical direction of the forward direction, forming a semicircular arc with an opening facing the inside of the extended skeleton.

[0051] Specifically, first, identify the two intersection points between the semi-circular arc boundary of the starting closed region and the tangent at the starting point of the extended skeleton. These two intersection points are located on the left and right sides of the extended skeleton, respectively. Then, find the starting point of the left side line segment. This starting point coincides with the left intersection point of the starting closed region in space. Connect the left side line segment to this left intersection point to ensure that the starting point of the left side line segment completely overlaps with the left boundary endpoint of the starting closed region.

[0052] Furthermore, the arc of the semicircular arc starts from the left vertical direction of the forward direction, goes around the back of the starting point, and extends to the right vertical direction of the forward direction, thus forming a semicircular arc. The semicircular arc is connected to the tangent segment at the starting point to close the semicircular plane area, which is the starting closed area of ​​the target process.

[0053] Furthermore, after performing the above offset operation on all discrete points, all offset points on the left side are connected sequentially to form a smooth curve according to their order on the extended skeleton. This curve is the left side line segment. Similarly, all offset points on the right side are connected sequentially to form a smooth curve. This curve is the right side line segment. Both side line segments are parallel to the extended skeleton, and the vertical distance between them and the extended skeleton is equal to the dynamic widening amplitude at every point.

[0054] Furthermore, the two endpoints of the semicircular arc are connected to the endpoint position with straight line segments, or the semicircular arc is directly enclosed by the tangent segment at the endpoint position. The resulting semicircular planar region is the endpoint extension region of the target process, which is located entirely outside the direction of the extension skeleton's forward movement.

[0055] Furthermore, the starting point of the right side line segment is completely overlapped with the right boundary endpoint of the starting point closed area. Then, the ending point of the left side line segment is found and overlapped with the left boundary endpoint of the ending point extension area. Similarly, the ending point of the right side line segment is overlapped with the right boundary endpoint of the ending point extension area. Finally, the semi-circular boundary of the starting point closed area, the left side line segment, the right side line segment, and the semi-circular boundary of the ending point extension area are connected end to end in sequence to form a closed continuous boundary line. The spatial area enclosed by this boundary line is the dynamic permissioned domain of the target process.

[0056] In summary, by enclosing the starting point of the dynamic licensed domain in a semi-circular arc shape, the enclosed area at the starting point can smoothly cover the semi-circular space behind the robot's current position. This avoids the distortion of the licensed domain or incomplete spatial coverage caused by the use of right angles or acute angles at the starting point, ensuring that the robot remains within the effective monitoring range of the dynamic licensed domain when it retreats or turns near the starting point of the path. At the same time, the semi-circular arc structure eliminates the adverse effects of cusp stress concentration on subsequent boundary alignment calculations.

[0057] In summary, the two side segments are completely parallel to the extended skeleton and the vertical distance between them is always equal to the dynamic widening range, thus constructing a strip-shaped lateral safety channel of uniform width. This allows any lateral deviation of the robot during its journey along the path to be accurately captured by the side segments. At the same time, the parallel structure ensures that the distance from each point on the side segment to the extended skeleton is equal, simplifying the complexity of distance calculation during subsequent overlap comparison.

[0058] In summary, the endpoint extension region adopts a semi-circular envelope structure to cover the entire space in front of the endpoint of the extension skeleton, so that the dynamic permission domain forms a smooth arc-shaped closed boundary at the end of the path, avoiding the detection blind spot caused by the sudden truncation of the endpoint, ensuring that the robot still has a complete forward permission range when approaching the endpoint of the path. At the same time, the semi-circular outer envelope can adapt to various motion postures of the robot entering the endpoint region from different angles.

[0059] In summary, by precisely aligning the semicircular boundary of the starting closed area with the starting endpoints of the side line segments on both sides, and then precisely aligning the ending endpoints of the side line segments with the semicircular boundary of the ending extended area, a seamless, non-overlapping, and continuously closed overall spatial area is finally spliced ​​together. The shape of this dynamic permission domain completely covers the entire permission space from the semicircle behind the robot's current position to both sides of the path in front and then to the semicircle in front of the end of the path, ensuring that there will be no missed detections or misjudgments due to boundary misalignment when comparing the overlap with the sensing buffer in the future.

[0060] In this embodiment of the invention, when generating an outward-expanding sensing band along the boundary line of the global constraint domain using the current expansion width of the dynamic permission domain as the outward expansion width value, and binding the spatial coordinate data of the area covered by the outward-expanding sensing band with the current expansion width to obtain the sensing buffer of the target process, it is specifically used for: The coordinate points on the boundary line of the global constraint domain are vertically offset outward from the global constraint domain, and the offset distance is the value of the outward expansion width, to obtain the outward expansion coordinate points of the target process; The outward coordinate points are connected sequentially to form an outward boundary line, and the strip-shaped region between the boundary line of the global constraint domain and the outward boundary line is taken as the outward sensing zone of the target process. The spatial coordinate data within the extended sensing band is associated with and stored with the extended width value to obtain the sensing buffer of the target process.

[0061] Specifically, all boundary coordinate points are extracted from the boundary line of the global constraint domain at uniform intervals to form an ordered sequence of boundary coordinate points. For each boundary coordinate point in the sequence, the tangent direction of the boundary line at that point is first calculated. Then, the normal direction perpendicular to the tangent and pointing to the outside of the global constraint domain is calculated based on the tangent direction. Next, the boundary coordinate point is moved a straight line distance along the outer normal direction. The value of this distance is equal to the outward expansion width value. The new position reached after the movement is the outward expansion coordinate point corresponding to the boundary coordinate point.

[0062] Specifically, the first outward coordinate point in the sequence of outward coordinate points is taken as the starting point. The first outward coordinate point is connected to the second outward coordinate point with a straight line segment. Then the second is connected to the third, and so on, until the second to last outward coordinate point is connected to the last outward coordinate point. Then the last outward coordinate point is connected to the first outward coordinate point with a straight line segment to close the loop, thus forming a complete closed curve. This curve is named the outward boundary line. The outward boundary line is located outside the boundary line of the global constraint domain, and the vertical distance between the two is equal to the outward width value everywhere.

[0063] Specifically, the process iterates through each spatial coordinate point covered by the extended sensing band, obtains the three-dimensional coordinate values ​​of each spatial coordinate point, and then collects all these spatial coordinate data to form a spatial coordinate dataset. At the same time, the specific value of the currently used extended width is obtained, and the extended width value is used as the attribute parameter of the extended sensing band. Then, a one-to-one binding relationship is established between the spatial coordinate dataset and the extended width value. That is, in this binding relationship, the extended width value is used as the index key, the spatial coordinate dataset is used as the index value, and the two are stored in the same record.

[0064] Furthermore, the same outward expansion operation is performed on each point in the boundary coordinate point sequence to obtain a set of outward expansion coordinate points with the same number as the boundary coordinate points. These outward expansion coordinate points are arranged in the same order as the original boundary coordinate points to form an outward expansion coordinate point sequence. Each outward expansion coordinate point in this sequence corresponds one-to-one with a coordinate point on the original boundary line.

[0065] Furthermore, all spatial regions between the boundary line of the global constraint domain and the outward expansion boundary line are extracted. This region is a ring-shaped strip region along the original boundary line, with its inner boundary being the boundary line of the global constraint domain and its outer boundary being the outward expansion boundary line. Each spatial point within this strip region satisfies that the shortest distance to the boundary line of the global constraint domain is less than or equal to the outward expansion width value. This strip region is the outward expansion sensing zone of the target process.

[0066] Furthermore, the binding relationship is finally written into the system's buffer storage structure. This storage structure can be a lookup table or a set of key-value pairs, so that any subsequent operation that needs to obtain the sensing buffer information can quickly retrieve the corresponding spatial coordinate data through the outer width value. After the above association storage is completed, the entire storage structure and its internal data set together constitute the sensing buffer of the target process.

[0067] In summary, by vertically offsetting the boundary line of the global constraint domain outward by a precise expansion width value, a new coordinate point is generated on the outside for each original boundary coordinate point. This translates the static global constraint domain boundary outward as a whole, providing a set of offset points with accurate positions and uniform spacing for the subsequent construction of the expanded sensing zone. This avoids the computational complexity and boundary deformation problems caused by directly expanding the internal space of the global constraint domain. At the same time, the vertical offset ensures that the vertical distance between the expanded boundary line and the original boundary line is equal everywhere, keeping the width of the sensing zone constant.

[0068] In summary, by connecting the outward coordinate points sequentially to form a closed outward boundary line, and together with the original boundary line of the global constraint domain, a strip-shaped region with a constant width and consistent direction is defined. This outward sensing strip precisely covers a buffer space adjacent to the outside of the global constraint domain, so that any movement behavior approaching the boundary of the global constraint domain from the dynamic permission domain will first enter this sensing strip, thus providing the system with a space area for early warning. At the same time, the strip structure avoids the huge storage overhead caused by the expansion of the entire space.

[0069] In summary, by binding and storing the coordinate data of each spatial point within the extended sensing zone with the extended width value used to generate the sensing zone, the system can directly retrieve the corresponding extended width value based on the overlapping coordinates when comparing the overlap between the dynamic permission domain and the sensing buffer, thereby determining the robot's proximity to the boundary of the global constraint domain. At the same time, the associated storage structure supports the coexistence of multiple sensing buffers corresponding to different extended width values, which facilitates the implementation of hierarchical early warning strategies.

[0070] In this embodiment of the invention, when determining the boundary touch state identifier of the target process based on the overlap of the real-time spatial coordinates of the sensing buffer and the dynamic permission domain, it is specifically used for: By comparing the spatial coordinates of the sensing buffer and the dynamic permission domain, the coincident coordinates of the target process are obtained; The boundary touch status identifier of the target process is generated based on the address markers of the overlapping coordinate points.

[0071] Specifically, all spatial coordinate data contained in the sensing buffer are extracted to form a first set of coordinate points, and all spatial coordinate data covered by the dynamic permission domain are extracted to form a second set of coordinate points. Then, each coordinate point in the first set of coordinate points is compared with each coordinate point in the second set of coordinate points one by one. During the comparison, it is determined whether the three-dimensional spatial coordinate values ​​of the two coordinate points are completely equal.

[0072] Specifically, each coincident coordinate point in the sequence of coincident coordinate points is traversed to obtain the address mark of the coincident coordinate point in the system spatial coordinate system. The address mark contains the spatial grid number to which the point belongs or its unique position code in the entire working scene. Then, the address marks of all coincident coordinate points are arranged in the order of their appearance in the sequence of coincident coordinate points to form an address mark list.

[0073] Furthermore, for all points with completely equal coordinate values, they are marked as coincident coordinate points. The spatial locations of all the marked coincident coordinate points are collected to form a sequence of coincident coordinate points. Each point in this sequence is simultaneously located within the spatial range of the sensing buffer and the spatial range of the dynamic permission domain. This sequence is the coincident coordinate point of the target process.

[0074] Furthermore, the address tag list is encapsulated into a data packet, and a boundary touch type field is added to the header of the data packet. The value of this field is fixed to a boundary touch status code, indicating that the currently detected state belongs to the boundary touch category. Finally, this complete data packet is output as the boundary touch status identifier of the target process. This boundary touch status identifier can be directly read and parsed by the subsequent hierarchical handling instruction module.

[0075] In summary, by comparing all spatial coordinates within the sensing buffer with all spatial coordinates within the dynamic permissioned domain point by point, the system can accurately extract overlapping coordinate points that belong to both regions. This allows the system to directly locate the specific spatial position where the dynamic permissioned domain intrudes into the sensing buffer, avoiding positioning errors caused by using approximate bounding boxes or distance estimation methods. At the same time, the extraction of overlapping coordinate points provides a quantitative basis for subsequent determination of the severity of boundary touch; the more overlapping points, the higher the risk of boundary crossing.

[0076] In summary, by encapsulating the address markers of overlapping coordinate points into boundary touch status identifiers, the system can quickly determine the specific spatial location and range of the touch without repeatedly comparing the original coordinate data. At the same time, the identifier is output in a standardized format, which facilitates direct parsing by the subsequent graded handling instruction module to execute response operations such as deceleration, thereby shortening the response delay from detection to handling and improving the real-time performance of the alarm system.

[0077] In this embodiment of the invention, when filtering the non-covered meshes of the inner domain mesh set in the global constraint domain based on the dynamic permission domain, and encapsulating the inner domain number of the non-covered meshes as the region overflow status identifier of the target process, it is specifically used for: The internal space of the global constraint domain is divided into grids to obtain the internal grid set of the target process; Based on the real-time spatial coordinate data within the spatial range covered by the dynamic permission domain, the spatial range of the inner domain grid set is differentially divided to obtain the non-covered grid of the target process; The inner domain number of the non-covered grid is combined with the region overflow marker to obtain the region overflow status identifier of the target process.

[0078] Specifically, the three-dimensional space range enclosed by the global constraint domain is determined, and the minimum and maximum values ​​of the space in the three coordinate axes are obtained. Then, the entire space range is divided into equally spaced grid cells in the three coordinate axes according to the preset grid side length. Each grid cell is a cube voxel of the same size. Starting from the minimum coordinate value, grid lines are generated sequentially along each coordinate axis. The intersection of the grid lines is the grid vertex, and the spatial region between adjacent grid lines is a grid cell. All spatial points inside the global constraint domain are traversed.

[0079] Specifically, the real-time spatial coordinate data of the dynamic licensed domain is obtained. This real-time spatial coordinate data includes the coordinate values ​​of all spatial points occupied by the dynamic licensed domain at the current moment. Then, each grid cell in the inner domain grid set is traversed. For each grid cell, all coordinate points of the cubic space range covered by the grid cell are obtained. The set of spatial coordinate points of the grid cell is compared with the set of spatial coordinate points of the dynamic licensed domain. If any spatial coordinate point in the grid cell belongs to the dynamic licensed domain, the grid cell is determined to be an covered grid.

[0080] Specifically, the inner domain number corresponding to each non-covered grid is taken out from the list of non-covered grids, and these inner domain numbers are arranged in sequence to form an inner domain number sequence. Then, a region overrun marker with a fixed format is generated. The region overrun marker is a predefined string constant used to indicate that the current state belongs to the situation where the robot has overrun the permitted area. The inner domain number sequence and the region overrun marker are concatenated and combined.

[0081] Furthermore, for each grid cell, it is determined whether the center point of the grid cell falls within the internal space enclosed by the boundary line of the global constraint domain. If the center point is inside, the grid cell is marked as belonging to the inner domain grid. If the center point is on the boundary or outside, the grid cell is discarded. All grid cells marked as inner domain grids are collected to form a grid cell set. Each grid cell in this set is assigned a unique inner domain number, which is sequentially incremented according to the generation order of the grid cells. This complete grid cell set is the inner domain grid set of the target process.

[0082] Furthermore, if none of the spatial coordinate points in the grid cell belong to the dynamic permission domain, the grid cell is determined to be an uncovered grid. After traversal, all grid cells determined to be uncovered grids are selected from the inner domain grid set, and these uncovered grids are arranged in ascending order of their inner domain numbers to form an uncovered grid list. This list is the uncovered grid of the target process differentiated from the inner domain grid set.

[0083] Furthermore, the combination method involves first writing the region overflow marker in the data header, then sequentially writing the inner domain numbers of all non-covered grids, with each inner domain number separated by a delimiter, and finally encapsulating the combined complete data structure into a data packet. The format of this data packet allows subsequent modules to directly parse out the region overflow marker and the corresponding list of non-covered grid numbers from it. This complete data packet is the region overflow status identifier of the target process.

[0084] In summary, by discretizing the continuous and smooth internal space of the global constraint domain into a series of regularly arranged grid cells and assigning each grid cell a unique internal domain number, the spatial inclusion relationship judgment, which originally required complex geometric calculations, is transformed into a simple grid marker comparison, significantly reducing the spatial computational complexity of computer processing. At the same time, the gridded data structure facilitates storage and fast retrieval using bitmaps or sparse matrices, providing an efficient indexing foundation for subsequent real-time differential operations between the dynamic permission domain and the global constraint domain, ensuring that the system can complete the coverage detection of a large spatial area within milliseconds.

[0085] In summary, by performing a grid-by-grid differential comparison between the real-time spatial coordinate data of the dynamic permission domain and the spatial range of each grid cell in the inner domain grid set, the system accurately filters out those grid cells that are not covered by the dynamic permission domain at all. These uncovered grids directly correspond to the blank areas outside the robot's current permission range within the global constraint domain, enabling the system to quantitatively assess the spatial volume and distribution location of the robot outside the permission range. This avoids the approximation errors caused by using distance field or boundary projection methods. At the same time, the differential operation utilizes the discrete characteristics of gridding, simplifying complex spatial geometric intersection operations into a set subtraction operation of grid markers, which significantly improves detection efficiency.

[0086] In summary, by concatenating and encapsulating the internal domain numbers of all uncovered grids with a fixed string type region overflow marker in a unified format, a compact and standardized status identifier data packet is generated. This identifier carries both qualitative information about whether a region overflow has occurred and specific quantitative information about the overflow location. This allows the subsequent hierarchical handling instruction module to directly parse which grid regions are not covered without re-accessing the original spatial data, thereby quickly determining the severity of the overflow and triggering the corresponding emergency stop instruction. At the same time, this combined format facilitates efficient transmission and storage between different modules of the system, reducing data transmission bandwidth and parsing latency.

[0087] In this embodiment of the invention, when performing amplitude modulation processing on the current velocity vector of the dynamic permission domain based on the indication types of the boundary touch state indicator and the area exceedance state indicator to obtain the hierarchical handling instruction for the target process, it is specifically used for: Based on the coincident coordinate point data corresponding to the boundary touch state identifier, the magnitude of the current velocity vector in the dynamic permission domain is attenuated to obtain the deceleration command of the target process; Based on the out-of-area status identifier corresponding to the out-of-area number, the magnitude of the current velocity vector in the dynamic permission domain is set to zero to obtain the emergency stop command for the target process. The deceleration command and the emergency stop command are combined to obtain the graded handling command for the target process.

[0088] Specifically, address markers of all overlapping coordinate points are extracted from the boundary touch state identifier. Based on these address markers, the specific spatial location where the sensing buffer and the dynamic permissioned domain overlap is located. Then, the velocity vector of the dynamic permissioned domain at the current moment is obtained. This velocity vector contains direction information and magnitude information. The magnitude represents the magnitude of the robot's current movement speed. Next, an attenuation coefficient is determined based on the number of overlapping coordinate points and the distance of each overlapping coordinate point from the center line of the dynamic permissioned domain. The more overlapping coordinate points or the closer the overlapping points are to the center line, the smaller the value of the attenuation coefficient. The magnitude of the current velocity vector of the dynamic permissioned domain is multiplied by the attenuation coefficient to obtain a new magnitude value, which is less than the original magnitude value.

[0089] Specifically, an out-of-bounds number list is parsed from the region out-of-bounds status identifier. This list contains the inner domain numbers of all non-covered meshes. Each out-of-bounds number represents a mesh cell in the global constraint domain that is not covered by the dynamic permission domain. Then, it is checked whether the out-of-bounds number list is empty. If the list is not empty, it is determined that a region out-of-bounds behavior has occurred. At this time, the velocity vector of the dynamic permission domain at the current moment is obtained, and the magnitude of the velocity vector is directly set to zero. At the same time, the direction of the velocity vector remains unchanged or the direction is also set to zero vector, thus completing the zeroing process of the velocity vector magnitude.

[0090] Specifically, the deceleration command and the emergency stop command generated at the current moment are obtained. First, it is checked whether the emergency stop command exists. If the emergency stop command exists, it is treated as the highest priority handling command, while the deceleration command is ignored or temporarily stored because the priority of emergency stop is higher than that of deceleration. Then, the emergency stop command is encapsulated into the main command slot of a hierarchical handling command package, and the command level field is set to the highest level. If the emergency stop command does not exist but the deceleration command exists, the deceleration command is encapsulated into the secondary command slot of the hierarchical handling command package, and the command level field is set to the second highest level.

[0091] Furthermore, the direction of the current velocity vector in the dynamic permission domain remains unchanged, and only its magnitude is replaced with this new magnitude value to complete the attenuation process of the velocity vector magnitude. Finally, the attenuated velocity vector is encapsulated into a control command, which is named the deceleration command of the target process. This deceleration command is sent to the robot's motion control system so that the robot continues to move at the attenuated velocity.

[0092] Furthermore, the zeroed velocity vector is encapsulated into a control command, which is named the emergency stop command for the target process. After the emergency stop command is sent to the robot's motion control system, the robot immediately stops all movement, the speed drops to zero, and the robot will not execute any new motion commands until the emergency stop command is cleared.

[0093] Furthermore, if two instructions exist simultaneously and the emergency stop instruction takes precedence, the graded handling instruction package contains only the emergency stop instruction. Finally, the encapsulated graded handling instruction package is output as the graded handling instruction for the target process. This graded handling instruction contains an instruction type identifier and corresponding speed vector control parameters, which are used by the robot control system to perform the corresponding speed adjustment operations according to the instruction level.

[0094] In summary, by directly utilizing the overlapping coordinate point data carried in the boundary touch status identifier, the specific location and range of the dynamic permission domain intruding into the sensing buffer are accurately located. The magnitude of the velocity vector is attenuated based on the degree of overlap rather than completely stopped, allowing the robot to smoothly decelerate instead of abruptly stop when it slightly touches the sensing buffer, avoiding operation interruption or mechanical impact caused by over-response. At the same time, the attenuation process retains the directional information of the velocity vector, ensuring that the robot can still move along the task path but at a safer speed. This achieves a balance between boundary warning and operation continuity, reducing time loss caused by unnecessary downtime.

[0095] In summary, by parsing the list of out-of-bounds numbers in the region out-of-bounds status identifier, it can be directly determined that there are mesh cells in the global constraint domain that are not covered by the dynamic permission domain. That is, the robot has exceeded the permitted safe area. At this time, setting the magnitude of the velocity vector to zero can trigger the most urgent braking response, ensuring that the robot immediately stops all movement and avoids entering the dangerous area outside the global constraint domain or colliding with obstacles due to continued movement. At the same time, the zeroing process has the highest priority and is not affected by the number of overlapping coordinate points or the attenuation coefficient, ensuring that the system executes the most severe handling measures in the case of such a serious violation as region out-of-bounds.

[0096] In summary, by combining deceleration and emergency stop commands into a hierarchical response command package based on priority, the robot control system can automatically select and execute the corresponding speed adjustment operation according to the level field in the command package, without the need for external logic to determine conflicts between different commands. When both boundary touch and area violation exist simultaneously, the hierarchical response command package prioritizes the emergency stop command to ensure the fastest response in the most dangerous situation. When only boundary touch exists, the deceleration command is executed, realizing a complete hierarchical response chain from warning to emergency braking. This combined structure reduces the system's switching delay between multiple commands and improves the overall reliability and timeliness of alarm handling.

[0097] Compared with the prior art, the present invention has the following beneficial effects:

[0098] 1. This technology constructs a three-layer spatial structure consisting of a global constraint domain, a dynamic permitted domain, and a sensing buffer zone. Based on the robot's real-time motion parameters, it dynamically adjusts the width of the dynamic permitted domain, enabling the system to automatically expand or shrink the permitted area according to the robot's current speed and acceleration. This provides a greater safety margin during high-speed movement and avoids unnecessary spatial redundancy during low-speed movement, achieving adaptive and proactive early warning of boundary violation risks. Simultaneously, by comparing the spatial coordinates of the sensing buffer zone and the dynamic permitted domain point-by-point to generate coincident coordinate points, and using a grid differencing method to accurately extract the uncovered mesh within the global constraint domain, this technology can simultaneously detect boundary touch behavior and area overrun behavior. This ensures that violations ranging from slight approach to severe boundary violations are accurately captured, greatly improving the comprehensiveness of detection and positioning accuracy.

[0099] 2. This technology further performs modulus attenuation or zeroing processing on the current velocity vector of the dynamic permitted domain based on different indication types of boundary touch status and area exceedance status. It combines deceleration and emergency stop commands into hierarchical handling instructions, enabling the system to smoothly decelerate to maintain operational continuity when minor boundary touch is detected, and to immediately stop for absolute safety when area exceedance is detected, achieving a complete hierarchical response chain from warning to braking. Furthermore, by meshing the internal space of the global constraint domain and using numbered labels for differential operations, this technology transforms complex spatial geometric judgments into efficient mesh index comparisons, significantly improving the computational efficiency of real-time detection and response, ensuring that the system can still complete millisecond-level alarm handling even during high-speed robot movement.

[0100] like Figure 2 The diagram shown is a functional block diagram of a robot boundary crossing detection and alarm system provided in an embodiment of the present invention.

[0101] The robot boundary crossing detection and alarm system 100 of this invention can be installed in an electronic device. Depending on the functions implemented, the robot boundary crossing detection and alarm system 100 may include a global constraint domain module 101, a dynamic permission domain module 102, a sensing buffer module 103, a boundary touch status indicator module 104, a region exceedance status indicator module 105, and a graded handling instruction module 106. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0102] In this embodiment, the functions of each module / unit are as follows: The global constraint domain module constructs the global constraint domain of the target process based on the static boundary information of the robot's working scenario during the target process; The dynamic permission domain module starts from the robot's current position in the target process and extends it along the path of the robot's task to be executed based on the robot's motion parameters to obtain the dynamic permission domain of the target process. The sensing buffer module uses the current expansion width of the dynamic permission domain as the expansion width value, generates an outward expansion sensing band along the boundary line of the global constraint domain, and binds the spatial coordinate data of the area covered by the outward expansion sensing band with the current expansion width to obtain the sensing buffer of the target process. The boundary touch state identification module determines the boundary touch state identification of the target process based on the overlap of the real-time spatial coordinates of the sensing buffer and the dynamic permission domain. The region out-of-bounds status identification module filters the non-covered meshes in the inner domain mesh set of the global constraint domain based on the dynamic permission domain, and encapsulates the inner domain number of the non-covered meshes as the region out-of-bounds status identification of the target process; The graded handling instruction module performs amplitude adjustment processing on the current velocity vector of the dynamic permission domain according to the indication type of the boundary touch status indicator and the area exceed status indicator, so as to obtain the graded handling instruction for the target process.

[0103] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0104] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0105] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0106] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0107] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0108] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for detecting and alarming robot boundary crossing behavior, characterized in that, The method includes: S1. Construct the global constraint domain of the target process based on the static boundary information of the robot's working scene during the target process; S2. Starting from the robot's current position in the target process, extend the robot's motion parameters along the path of the task to be executed to obtain the dynamic permission domain of the target process. S3. Using the current expansion width of the dynamic permission domain as the outward expansion width value, generate an outward expansion sensing band along the boundary line of the global constraint domain, and bind the spatial coordinate data of the area covered by the outward expansion sensing band with the current expansion width to obtain the sensing buffer of the target process. S4. Determine the boundary touch state identifier of the target process based on the overlap between the real-time spatial coordinates of the sensing buffer and the dynamic permission domain. S5. Based on the dynamic permission domain, filter the non-covered meshes in the inner domain mesh set of the global constraint domain, and encapsulate the inner domain number of the non-covered meshes as the region out-of-bounds status identifier of the target process; S6. Based on the indication type of the boundary touch status indicator and the area exceed status indicator, the current velocity vector of the dynamic permission domain is amplitude-adjusted to obtain the hierarchical handling instruction of the target process.

2. The robot boundary crossing detection and alarm method as described in claim 1, characterized in that, The construction of the global constraint domain for the target process based on the static boundary information of the robot's working scenario includes: Extract the maximum permissible boundary coordinates of the robot's working scene during the target process to obtain an ordered set of geographic coordinates for the target process; Connect the coordinate points in the ordered geographic coordinate point set in sequence, and then perform a closed-loop splicing between the last coordinate point and the first coordinate point to obtain the closed boundary loop of the target process. The global constraint domain of the target process is obtained by filling the internal space enclosed by the closed boundary loop.

3. The robot boundary crossing detection and alarm method as described in claim 1, characterized in that, The step of extending the dynamic permission domain of the target process from the robot's current position along the path of the task to be executed based on the robot's motion parameters to obtain the dynamic permission domain of the target process includes: The target path points in the task path to be executed are connected sequentially to obtain the path curve of the target process, and the path segment starting from the current position of the robot along the path direction is used as the extended skeleton of the target process on the path curve. The dynamic broadening of the target process is calculated based on the motion parameters and the baseline broadening of the target process. Using the extended skeleton as the center line and the dynamic widening range as the lateral expansion radius, the space is expanded along both sides and the end point of the extended skeleton to obtain the dynamic permission domain of the target process.

4. The robot boundary crossing detection and alarm method as described in claim 3, characterized in that, The formula for calculating the dynamic widening range includes: in, The dynamic widening range, The reference width expansion, It is a natural constant. This is the global scaling factor. For the speed term, the weighting coefficient is... This represents the robot's current speed. As the speed reference value, For speed sensitivity index, It is the hyperbolic tangent function. This represents the robot's current acceleration value. As the reference value for acceleration, For the acceleration term, For acceleration sensitivity index, It is the natural logarithm function. This is the directional consistency factor.

5. The robot boundary crossing detection and alarm method as described in claim 3, characterized in that, The process of spatially expanding along both sides and the endpoint of the extended skeleton, using the dynamic widening range as the lateral expansion radius, with the extended skeleton as the centerline, to obtain the dynamic permission domain of the target process includes: Starting from the starting point of the extended skeleton, and moving against the direction of the extended skeleton, a semi-circular arc is extended with the dynamic widening range as the expansion radius to obtain the starting closed area of ​​the target process. Along both sides of the extended skeleton, with the dynamic widening amplitude as the vertical distance, side line segments parallel to the extended skeleton are generated; Centered on the endpoint of the extended skeleton, the dynamic widening amplitude is used as the radial distance, and an outer envelope is formed along the forward direction of the extended skeleton to obtain the endpoint extension region of the target process. Align the boundaries of the starting closed region, the side line segment, and the ending extended region to obtain the dynamic permission domain of the target process.

6. The robot boundary crossing detection and alarm method as described in claim 1, characterized in that, The step of generating an outward-expanding sensing band along the boundary line of the global constraint domain, using the current expansion width of the dynamic permission domain as the outward expansion width value, and binding the spatial coordinate data of the area covered by the outward-expanding sensing band with the current expansion width, to obtain the sensing buffer of the target process, includes: The coordinate points on the boundary line of the global constraint domain are vertically offset outward from the global constraint domain, and the offset distance is the value of the outward expansion width, to obtain the outward expansion coordinate points of the target process; The outward coordinate points are connected sequentially to form an outward boundary line, and the strip-shaped region between the boundary line of the global constraint domain and the outward boundary line is taken as the outward sensing zone of the target process. The spatial coordinate data within the extended sensing band is associated with and stored with the extended width value to obtain the sensing buffer of the target process.

7. The robot boundary crossing detection and alarm method as described in claim 1, characterized in that, The step of determining the boundary touch state identifier of the target process based on the overlap of the real-time spatial coordinates of the sensing buffer and the dynamic permission domain includes: By comparing the spatial coordinates of the sensing buffer and the dynamic permission domain, the coincident coordinates of the target process are obtained; The boundary touch status identifier of the target process is generated based on the address markers of the overlapping coordinate points.

8. The robot boundary crossing detection and alarm method as described in claim 1, characterized in that, The step of filtering non-covered meshes in the inner domain mesh set of the global constraint domain based on the dynamic permission domain, and encapsulating the inner domain number of the non-covered meshes as the region overflow status identifier of the target process, includes: The internal space of the global constraint domain is divided into grids to obtain the internal grid set of the target process; Based on the real-time spatial coordinate data within the spatial range covered by the dynamic permission domain, the spatial range of the inner domain grid set is differentially divided to obtain the non-covered grid of the target process; The inner domain number of the non-covered grid is combined with the region overflow marker to obtain the region overflow status identifier of the target process.

9. The robot boundary crossing detection and alarm method as described in claim 1, characterized in that, The step of performing amplitude modulation processing on the current velocity vector of the dynamic permissioned domain based on the indication types of the boundary touch status indicator and the area exceedance status indicator to obtain the hierarchical handling instructions for the target process includes: Based on the coincident coordinate point data corresponding to the boundary touch state identifier, the magnitude of the current velocity vector in the dynamic permission domain is attenuated to obtain the deceleration command of the target process; Based on the out-of-area status identifier corresponding to the out-of-area number, the magnitude of the current velocity vector in the dynamic permission domain is set to zero to obtain the emergency stop command for the target process. The deceleration command and the emergency stop command are combined to obtain the graded handling command for the target process.

10. A robot boundary crossing detection and alarm system, used to implement the robot boundary crossing detection and alarm method according to any one of claims 1-9, characterized in that, The system includes: The global constraint domain module constructs the global constraint domain of the target process based on the static boundary information of the robot's working scenario during the target process; The dynamic permission domain module starts from the robot's current position in the target process and extends it along the path of the robot's task to be executed based on the robot's motion parameters to obtain the dynamic permission domain of the target process. The sensing buffer module uses the current expansion width of the dynamic permission domain as the expansion width value, generates an outward expansion sensing band along the boundary line of the global constraint domain, and binds the spatial coordinate data of the area covered by the outward expansion sensing band with the current expansion width to obtain the sensing buffer of the target process. The boundary touch state identification module determines the boundary touch state identification of the target process based on the overlap of the real-time spatial coordinates of the sensing buffer and the dynamic permission domain. The region out-of-bounds status identification module filters the non-covered meshes in the inner domain mesh set of the global constraint domain based on the dynamic permission domain, and encapsulates the inner domain number of the non-covered meshes as the region out-of-bounds status identification of the target process; The graded handling instruction module performs amplitude adjustment processing on the current velocity vector of the dynamic permission domain according to the indication type of the boundary touch status indicator and the area exceed status indicator, so as to obtain the graded handling instruction for the target process.

Images