A safety operation early warning system and method for an execution end

Abstract

The application discloses an execution end safety operation early warning system and method, and belongs to the technical field of engineering operation safety monitoring. Multi-source perception data of the surrounding environment of the execution end is collected by using a perception module, time stamp synchronization, spatial coordinate calibration and fusion processing are performed on the multi-source perception data, and a three-dimensional environment model of the execution end operation area is generated; based on the three-dimensional environment model, real-time operation conditions of the execution end and historical safety accident data, the risk level of an obstacle is determined and a dynamic safety distance threshold value is generated; a differentiated early warning signal is output according to the risk level of the obstacle and the dynamic safety distance threshold value, and a safety intervention action of the corresponding level is triggered.

Description

Technical Field

[0001] This invention belongs to the technical field of engineering operation safety monitoring, specifically relating to an execution-end safety operation early warning system and method. Background Technology

[0002] During the operation of engineering equipment (such as loaders, excavators, and elevators), the actuator (bucket, car, etc.) is the core component that directly participates in the operation or bears the load. Its operating area often contains various obstacles such as personnel, materials, and buildings. Moreover, the field of vision of the actuator is easily obstructed when it is operating, making it difficult for operators to fully perceive the surrounding risks.

[0003] Existing safety early warning technologies for engineering equipment are mostly designed for monitoring the surrounding environment during vehicle operation, focusing on the risk of vehicle collisions. They are not adapted to specific operational scenarios at the execution end (such as bucket lifting / lowering, elevator lifting, and excavator digging). On the one hand, sensor deployment is not aligned with the operational range of the execution end, resulting in blind spots in obstacle recognition (such as under the bucket or to the side of the elevator). On the other hand, early warning methods are mostly based on single audio-visual prompts, without designing differentiated visual alarms according to the operational conditions and obstacle risk levels at the execution end. Furthermore, when multiple obstacles are present simultaneously, the warning logic is confusing, which can easily lead to operators misjudging or ignoring risks, causing safety accidents such as collisions and crushing.

[0004] Therefore, there is an urgent need for a technical solution that focuses on the operational characteristics of the execution end, provides accurate perception, intuitive alarms, and collaborative early warning for multiple obstacles, to fill the gap in the field of execution end safety early warning in existing technologies. Summary of the Invention

[0005] To address the technical problems existing in the background art described above, the present invention provides an execution-end safety operation early warning system and method.

[0006] This invention is achieved through the following technical solution: an execution-end safety operation early warning system, comprising: Engineering equipment execution end; A sensing module is deployed in the working area of ​​the engineering equipment's execution end; the sensing module is used to collect multi-source sensing data of the surrounding environment of the execution end; The data fusion module is communicatively connected to the sensing module; the data fusion module is used to perform timestamp synchronization, spatial coordinate calibration and fusion processing on multi-source sensing data to generate a three-dimensional environment model of the execution end's working area. The risk assessment module is communicatively connected to the data fusion module; the risk assessment module determines the obstacle risk level and dynamic safety distance threshold based on the three-dimensional environment model, the real-time working conditions of the execution end, and preset safety rules. The graded early warning and execution control module is communicatively connected to the risk assessment module and the braking system of the engineering equipment. The graded early warning and execution control module is used to output differentiated early warning signals according to the risk level of the obstacle, and to trigger the corresponding level of safety intervention action when the risk level reaches a preset threshold.

[0007] In a further embodiment, the engineering equipment execution end includes at least: One or more of bucket, car, excavator bucket, and lifting device, and the actuator is equipped with an installation interface adapted to its working range for fixing the sensing module; The sensing module includes at least one of lidar, fisheye camera, infrared camera and ultrasonic radar.

[0008] The execution-end safety operation early warning method based on the early warning system described above includes the following steps: The system utilizes a sensing module to collect multi-source sensing data of the surrounding environment of the execution terminal. This multi-source sensing data is then time-stamped, spatially calibrated, and fused to generate a three-dimensional environmental model of the execution terminal's work area. The multi-source sensing data includes obstacle locations and distances. Based on a 3D environment model, real-time operating conditions of the execution end, and historical safety accident data, the risk level of obstacles is determined and dynamic safety distance thresholds are generated. Differentiated warning signals are output based on the obstacle risk level and dynamic safety distance threshold, and corresponding level of safety intervention actions are triggered.

[0009] In a further embodiment, the steps for generating the three-dimensional environment model are as follows: Employing a timestamp synchronization algorithm based on the system's unified clock. Align the timestamps of multi-source sensing data; Establish a reference coordinate system for the actuator, with the connection center point between the actuator and the robotic arm of the engineering vehicle as the origin. Defined as perpendicular to the axis of the engineering vehicle's robotic arm and passing through the origin. The virtual plane is the reference plane ; Determine the differentiated coordinate transformation matrix based on the shape and motion characteristics of the surface where the execution end is located. The original coordinates of each sensor are transformed to the reference coordinate system of the execution end to obtain the processed point cloud data. Processed toroidal image data and processed near-field distance data ; Processed point cloud data Processed toroidal image data Processed near-field range data Using infrared image data as input, through state equations and observation equations The target trajectory is predicted and corrected, various data are fused to complete the environmental information of the occluded area, and the fused data is output. That is, a three-dimensional environment model.

[0010] In a further embodiment, the real-time operating conditions of the actuator include: angle adjustment condition, linear movement condition, composite linkage condition, heavy-load operation condition, precise positioning condition, and space-constrained condition.

[0011] In a further embodiment, the process for determining the obstacle risk level is as follows: Based on the type of execution end and the operation scenario, a set of risk impact factors is constructed. ,in, Obstacle type The speed of the obstacle's movement. To improve the speed of execution, For workload, The dimensions of the workspace; Determine the weighting coefficients of each risk factor and construct a judgment matrix. The judgment matrix The expression is as follows: ,in, Risk impact factor Relative to risk impact factors The importance of , ; Based on the judgment matrix Determine the weighting coefficient for each risk factor. The weighting coefficient satisfy: ; Using the weighting coefficients A risk assessment value calculation model is constructed, and its expression is as follows: ,in, Standardized values ​​of each risk factor. This is the risk assessment value; Based on historical safety incident data, risk level thresholds are set; the risk assessment values ​​are... The risk level is compared with the risk level threshold to generate the corresponding obstacle risk level.

[0012] In a further embodiment, the process of generating the dynamic safety distance threshold is as follows: For the real-time operating conditions of the execution end, and in combination with the equipment operating characteristics and industry safety standards, basic safety distance benchmark values ​​are set respectively. ; A working condition correction coefficient is created based on the real-time working conditions of the execution terminal. , The working condition number is used to indicate the real-time operating conditions at the execution end; a risk level correction coefficient is introduced based on the number of obstacles and their corresponding risk levels. , The level number corresponding to the risk level. This represents the current number of obstacles. Based on the aforementioned basic safety distance benchmark value Based on this, a dynamic safety distance calculation model is constructed, which is expressed as follows: In the formula, This is a speed correction term.

[0013] In a further embodiment, the output process of the differentiated early warning signal is as follows: When the obstacle risk level is no risk, a safety status prompt signal is output; When the obstacle risk level is low, a mild warning signal is output; When the obstacle risk level is medium risk, a moderate warning signal is output; When the obstacle risk level is high, a severe warning signal is output, and the warning information is simultaneously synchronized to the cloud management module.

[0014] In a further embodiment, the risk level correction coefficient The value depends on the number of obstacles currently in the path, specifically: like The risk level adjustment coefficient The value is taken as the risk basis correction coefficient for the obstacle; like First, determine the highest risk level among all obstacles. Obstacles and obtain the corresponding risk base correction coefficient The risk level adjustment coefficient The calculation formula is: .

[0015] In a further embodiment, it also includes: If there are multiple obstacles, the warning signal will be output in the form corresponding to the highest risk level. At the same time, all obstacle icons will be displayed in descending order of risk level on the operation control interface, and the buzzer prompt intensity will be set according to the highest risk level. The warning signal is dynamically adjusted as the real-time distance between the actuator and the obstacle changes.

[0016] The beneficial effects of this invention are: This invention adds a sensing module at the engineering execution end to fit the working range of the bucket, car and other execution ends. Combined with a differentiated coordinate transformation matrix, it adapts to different execution end structures and dynamic operation characteristics such as curved surfaces and planes, solving the problem of blind spots in the identification of the surrounding area of ​​the execution end (such as under the bucket and on the side of the car) in traditional whole vehicle monitoring.

[0017] The 3D environment model completes the information of occluded areas by fusing multi-source data and combines it with the unified definition of the reference coordinate system of the execution end to ensure the consistency and accuracy of environmental perception under complex working conditions.

[0018] This invention constructs a quantitative risk assessment model based on the weighting of risk impact factors and historical accident data, achieving precise classification of risk into no risk, low risk, medium risk, and high risk, avoiding misjudgments or neglect caused by a single warning. Through multi-dimensional prompts such as lights, buzzers, interface icons, and pop-ups, combined with a display logic prioritizing the highest risk in multi-obstacle scenarios, it allows operators to quickly focus on core risk sources.

[0019] The intervention actions mentioned in this invention are strongly correlated with warning levels and real-time distance, progressing gradually from deceleration and direction locking to emergency braking. This avoids excessive intervention that could affect operational efficiency while ensuring mandatory protection in high-risk scenarios. The intervention logic is adapted to special working conditions such as heavy loads and space constraints. For example, it extends the braking buffer time in heavy load conditions and provides lateral fine-tuning guidance in space-constrained conditions, balancing safety and operational practicality. Attached Figure Description

[0020] Figure 1 This is an architecture diagram of the execution-end safety operation early warning system of Example 1.

[0021] Figure 2 This is a flowchart of the execution-end safety operation early warning method in Example 2. Detailed Implementation

[0022] The present invention will now be further described with reference to the accompanying drawings and embodiments.

[0023] Example 1 like Figure 1 This embodiment discloses an execution-end safety operation early warning system. The system focuses on the operation safety monitoring needs of engineering equipment execution ends, and is compatible with various engineering equipment such as loaders, excavators, elevators, and cranes. It prevents collision risks during the execution end operation process. The specific structure is as follows: The engineering equipment execution end described in this embodiment has flexible configuration of operating functions, including at least one or more of the following: bucket, car, excavator bucket, and lifting device. All execution ends in this embodiment are equipped with standardized installation interfaces adapted to their operating range, and the interfaces adopt a combination structure of bolt fixing and shock-absorbing pads for buffering.

[0024] Furthermore, the installation interface locations are optimized based on the distribution of blind spots in the execution end operation. For example, the bucket installation interfaces are deployed at the bottom, sides, and rear of the bucket body, while the car installation interfaces are deployed at the four top corners and the middle of the sides, achieving no blind spots in the operating area.

[0025] It also includes: a perception module deployed in the working area of ​​the engineering equipment execution end; the perception module is used to collect multi-source perception data of the surrounding environment of the execution end. The perception module described in this embodiment includes at least one of the following: lidar (such as 6-line or 32-line lidar, deployed at the top or rear end of the execution end, with a sampling frequency of 10Hz-20Hz), fisheye camera (used to collect panoramic image data of the surrounding environment, and obtain the contour features and category information of obstacles through image recognition technology), infrared camera (to identify living beings with a temperature higher than the ambient temperature (such as personnel), avoiding missed detection of obstacles due to insufficient light), and ultrasonic radar (deployed in the close-range working area of ​​the execution end (such as under the lifting device, at the edge of the bucket), used to compensate for the blind spots of lidar and camera in close-range detection).

[0026] The data fusion module is communicatively connected to the sensing module, such as via Ethernet or CAN bus protocols. The data fusion module is used to perform timestamp synchronization, spatial coordinate calibration, and fusion processing on multi-source sensing data to generate a three-dimensional environment model of the execution end's working area. The risk assessment module is communicatively connected to the data fusion module; the risk assessment module determines the obstacle risk level and dynamic safety distance threshold based on the three-dimensional environment model, the real-time working conditions of the execution end, and preset safety rules. The graded early warning and execution control module is communicatively connected to the risk assessment module and the braking system of the engineering equipment. The graded early warning and execution control module is used to output differentiated early warning signals according to the risk level of the obstacle, and to trigger the corresponding level of safety intervention action when the risk level reaches a preset threshold.

[0027] It also includes a cloud management module, which is communicatively connected to the perception module, risk assessment module and graded early warning module, for storing operation data, early warning records and equipment status information, and realizing remote monitoring and operation and maintenance management.

[0028] Example 2 Based on the execution-end safety operation early warning system disclosed in Embodiment 1, this embodiment discloses an execution-end safety operation early warning method, including the following steps: The system utilizes a sensing module to collect multi-source sensing data of the surrounding environment of the execution terminal. This multi-source sensing data is then time-stamped, spatially calibrated, and fused to generate a three-dimensional environmental model of the execution terminal's work area. The multi-source sensing data includes obstacle locations and distances. Based on a 3D environment model, real-time operating conditions of the execution end, and historical safety accident data, the risk level of obstacles is determined and dynamic safety distance thresholds are generated. Differentiated warning signals are output based on the obstacle risk level and dynamic safety distance threshold, and corresponding level of safety intervention actions are triggered.

[0029] To address the issues of low accuracy in environmental modeling caused by asynchronous multi-source sensing data, inconsistent coordinate references among different types of execution terminals, and lack of environmental information in occluded areas in existing technologies, the steps for generating a 3D environment model described in this embodiment are as follows: Employing a timestamp synchronization algorithm based on the system's unified clock. The timestamps of the multi-source sensing data are aligned. In this embodiment, the multi-source sensing data includes obstacle positions and distances. The specific time requirements for alignment are as follows: satisfy , , , To ensure consistency in time; among which, The timestamp of the cloud data for LiDAR points. Timestamps for the panoramic image data from the fisheye camera. This is a timestamp for the low-light image data from the infrared camera. This is the timestamp for the near-field range data of the ultrasonic radar.

[0030] Establish a reference coordinate system for the actuator, with the connection center point between the actuator and the robotic arm of the engineering vehicle as the origin. Define a virtual plane perpendicular to the axis of the engineering vehicle's robotic arm and passing through the origin O as the reference plane. ; It should be noted that due to the diversity of the execution terminals and the different working conditions, the shape and motion state of the surface on which the execution terminal is located vary significantly: for example, the bucket has an arc-shaped structure and is accompanied by dynamic movements such as pitching and digging during operation, which can easily create blind spots; the car has a planar structure and is mainly vertically lifting during operation, with a relatively fixed motion trajectory; the bucket and lifting device are composite shapes and have both rotational and translational movements during operation, making the coordinate reference easily deviate; based on the above differences, the shape (arc / planar / composite shape) and motion characteristics and motion trajectory information of the surface on which the execution terminal is located are used.

[0031] Therefore, the origin described in this embodiment The origin is uniformly defined as the connection center point between the actuator and the robotic arm of the engineering vehicle, i.e., the mounting hinge point of the actuator. Examples include the hinge center between the bucket and the boom, the connection center between the car and the lifting guide rail, the hinge center between the bucket and the stick, and the connection center between the lifting device and the boom. The displacement is synchronized with the movement of the robotic arm to ensure a strong correlation between the coordinate reference and the motion trajectory of the actuator.

[0032] The reference plane Defined as perpendicular to the axis of the engineering vehicle's robotic arm and passing through the origin. The virtual plane, such as the boom corresponding to the bucket, has a reference plane perpendicular to the boom axis and passing through the hinge center; the lifting guide rail corresponds to the car, with the reference plane perpendicular to the guide rail axis and passing through the connection center. This embodiment uses the origin... and reference plane A unified definition is provided to simplify the coordinate conversion logic of different types of execution ends and improve the consistency and accuracy of data fusion under complex working conditions.

[0033] Furthermore, considering the significant differences in the shape and motion state of the surface where the actuator is located (e.g., the bottom of the bucket is an arc surface, and the surface of the bucket is a composite curved surface), and the different installation positions of different sensing sensors (which can be deployed in different areas such as the bottom / side of the bucket and the surface of the bucket), there are significant differences in the coordinate conversion between different types of actuators and different positions of the same actuator.

[0034] Therefore, this embodiment uses a differentiated coordinate transformation matrix. The original coordinates of each sensor are transformed to the reference coordinate system, where, , For rotation matrix, It is a translation vector.

[0035] The differential coordinate transformation matrix Depending on the shape of the surface where the execution terminal is located, further examples will be provided below: If the sensing sensor is installed on a curved surface, such as the bottom surface of a bucket or the curved surface of a digging bucket, then the rotation matrix... The expression is: ;in, , , These are the bucket pitch, roll, and yaw angles, which are collected in real time by the inertial measurement unit. Translation vector The expression is: ,in, , and These are the sensing sensors relative to the geometric center of the bucket. axis, shaft and Axis translation amount, This is the transpose of the matrix.

[0036] If the sensing sensor is mounted on a flat surface, such as the surface of the car, then the rotation matrix... The expression is: ; Corresponding translation vector Still: .

[0037] Based on the aforementioned differential coordinate transformation matrix The processed point cloud data were obtained respectively. Processed toroidal image data and processed near-field distance data ; Furthermore, the point cloud data acquired directly is This refers to the point cloud data before processing. For example, the currently collected point cloud data is... ,in, For point clouds It is represented in homogeneous coordinate form as a 4×1 dimensional column vector. , Point clouds The three-dimensional coordinates.

[0038] Then, the corresponding differential coordinate transformation matrix is ​​determined based on the surface where the lidar where the point cloud is obtained. Processed point cloud data : .

[0039] For the processed panoramic image data The acquisition methods are as follows: Four sets of panoramic image data output by the fisheye camera For example, The acquired image includes several pixels and their corresponding... , The number of groups of images in the panorama view. For pixels.

[0040] Distortion correction and intrinsic / extrinsic parameter calibration are performed on the fisheye camera to obtain the camera's intrinsic parameter matrix and distortion coefficients. The pixel coordinates are then calculated using the distortion correction formula. Convert to normalized coordinates in camera coordinate system : ,in, This is the camera intrinsic parameter matrix. The distortion coefficient is... This is a distortion correction function. It takes pixel coordinates, distortion coefficients, and intrinsic parameter matrix as input and outputs normalized coordinates.

[0041] Based on this, select the appropriate differential coordinate transformation matrix. Processed ring view image data The calculation method is as follows: ;in, , and This refers to the three-dimensional coordinates of a pixel after transformation to the reference coordinate system of the execution end, i.e. .

[0042] Correspondingly, the raw near-field distance data is obtained by using ultrasonic radar to measure the straight-line distance between the obstacle and the probe. ,in, These represent the straight-line distance between the obstacle and the probe measured by the ultrasonic radar, which is then directly mapped to three-dimensional coordinates in the radar's own coordinate system. Based on the selected differential coordinate transformation matrix The formula for transforming the coordinates in the radar coordinate system to the reference coordinate system of the actuator is as follows: ; Based on the transformed The processed near-field range data is obtained. , : .

[0043] Processed point cloud data Processed toroidal image data Processed near-field range data Using infrared image data as input, through state equations and observation equations The target trajectory is predicted and corrected, various data are fused to complete the environmental information of the occluded area, and the fused data is output. That is, a three-dimensional environment model.

[0044] Processed point cloud data Processed toroidal image data Processed near-field range data Using infrared image data as input, through state equations and observation equations The target trajectory is predicted and corrected, various data are fused to complete the environmental information of the occluded area, and the fused data is output. That is, a three-dimensional environment model.

[0045] Furthermore, the state equation described in this embodiment The expression is as follows: ,in, For the first The system state vector at any given time. for The system state prediction vector at time step, This is the state transition matrix, which describes the evolution of the state over time. This is the control matrix, which reflects the influence of control inputs on the system state; For the first Control the input vector at all times; For the first The system process noise at any given time follows a mean of 0 and a variance of . The Gaussian distribution.

[0046] Observation equations The expression is as follows: In the formula, For the observation matrix, for The observation vector at time; For the first The noise observed at all times follows a mean of 0 and a variance of 1. The Gaussian distribution.

[0047] Therefore, the merged data The expression form is: .

[0048] In a further embodiment, the real-time operating conditions of the execution end include angle adjustment, linear movement, composite linkage, heavy-load operation, precise positioning, and space-constrained operation. These six types of conditions directly address the core operating scenarios of the execution end, providing concise and clear scenario support for 3D environment model adaptation and risk assessment, as detailed below: Angle adjustment mode: Adjustment of the attitude of the execution end in pitch, roll, and yaw direction (such as bucket digging angle adaptation and lifting tool level calibration); Linear movement operation: Vertical lifting, horizontal translation, and inclined propulsion of the actuator (such as car lifting, bucket extension and retraction, and bucket slope propulsion). Composite linkage working conditions: coordinated angle adjustment and linear movement (such as bucket rotation and extension, lifting and translation of spreader, bucket pitching and lifting); Heavy-duty operation conditions: Operations under full load or high load on the actuator (such as full-load lifting of the car, heavy-load excavation of the bucket, and lifting of the lifting equipment under rated load). Precise positioning conditions: The execution end work point is precisely aligned or the landing point is precisely stopped (such as the positioning of the lifting device docking and the positioning of the excavator bucket work point). Space-constrained work conditions: Operations in confined spaces or gaps between obstacles (such as underground bucket excavation, indoor hoisting, and bucket excavation next to walls).

[0049] Based on the above analysis of the scenarios and the execution end, the obstacle risk level determination process in this embodiment is as follows: A set of risk influencing factors is constructed based on the type of execution device (bucket / cab / excavator / lifting device) and the operation scenario (open outdoor / confined indoor / underground operation, etc.). ,in, The obstacle type (personnel / materials / buildings / others) is quantified by hazard priority as 1.0 / 0.6 / 0.3 / 0.1. The speed of the obstacle (unit: m / s). The execution speed (unit: m / s) is the speed at which the device operates. This refers to the workload (quantified as a percentage of rated load, ranging from 0 to 1.0). The working space size is quantified as the ratio of the actual working space to the maximum working range of the execution end, ranging from 0 to 1.0. Determine the weighting coefficients of each risk factor and construct a judgment matrix. , ,in, Risk impact factor Relative to risk impact factors The importance is determined by a 1-9 scale (1 indicates equal importance, 3 indicates slightly important, 5 indicates significant importance, 7 indicates strong importance, 9 indicates extreme importance, and the reciprocal indicates the opposite importance). , ; Calculate the judgment matrix Product of each row of elements And calculate the geometric mean. ; for the geometric mean Normalization is performed to obtain the weight coefficients. The weighting coefficients are satisfied satisfy: ; Calculate the judgment matrix Maximum eigenvalue ,in For matrix with vector The k-th element of the product; vector , This is the matrix transpose of the matrix.

[0050] Based on weighting coefficients Construct a risk assessment value calculation model ,in Standardize the values ​​of each factor (mapping the raw data to the 0-1.0 range); combine historical accident statistics to set risk level thresholds; and assign the risk assessment values ​​to... The risk level is compared with the risk level threshold to generate the corresponding obstacle risk level.

[0051] Specifically: Obstacles pose no risk Low risk of obstacles Risks in obstacles High risk of obstacles This leads to the formation of risk level determination rules.

[0052] Considering the uncontrollability of obstacles and the inherent state of the actuator, a fixed basic safety distance cannot adapt to the diverse operating conditions of the actuator: existing technologies do not differentiate the safety requirements of different operating conditions such as angle adjustment and heavy-load operation, and a uniform threshold leads to inconvenience in space-constrained conditions or insufficient protection under heavy-load conditions. Therefore, the generation process of the dynamic safety distance threshold in this embodiment is as follows: For the real-time operating conditions of the execution end, and in combination with the equipment operating characteristics and industry safety standards, basic safety distance benchmark values ​​are set respectively. Based on the six working conditions mentioned above, the basic safety distance benchmark value is... The specific value can be: the baseline safety distance value for the angle adjustment working condition. The baseline safety distance for linear movement conditions The baseline safety distance for combined linkage operating conditions Basic safety distance benchmark for heavy-load operation conditions The baseline safety distance for precise positioning conditions Basic safety distance benchmark for space-constrained working conditions .

[0053] A working condition correction coefficient is created based on the real-time working conditions of the execution terminal. , This refers to the working condition number for the real-time operation of the execution end; in this embodiment, working condition correction coefficients are set for angle adjustment, linear movement, composite linkage, heavy-load operation, precise positioning, and space-constrained operation. for , , , , and The specific values ​​can be: , , , , and .

[0054] Meanwhile, existing technologies do not consider the cumulative effect of the number of obstacles and risk levels: when multiple obstacles coexist, the safe distance is determined only based on a single obstacle, ignoring the actual scenario where high-risk obstacles dominate and multiple sources of risk are superimposed, which can easily lead to collision hazards.

[0055] To address the aforementioned technical issues, a risk level correction coefficient is introduced based on the number of obstacles and their corresponding risk levels. , The risk level is assigned a level number (four sub-levels are given above, therefore...). The value can be 1 to 4. This represents the current number of obstacles. Based on the aforementioned basic safety distance benchmark value Based on this, a dynamic safety distance calculation model is constructed, which is expressed as follows: In the formula, For the speed correction term, further, ,in, To obtain the real-time operation speed of the execution end based on the 3D environment model, The maximum speed of movement for all obstacle species. This is the calculated dynamic safety distance.

[0056] Furthermore, risk level adjustment coefficient The value depends on the number of obstacles currently in the path, specifically: like The risk level adjustment coefficient The value is taken as the risk basis correction coefficient for the obstacle, that is, the obstacle is risk-free. Low risk of obstacles Risks in obstacles High risk of obstacles .

[0057] like First, determine the highest risk level among all obstacles. Obstacles and obtain the corresponding risk base correction coefficient The risk level adjustment coefficient The calculation formula is: .

[0058] By integrating the highest risk level and the number of obstacles using a risk level correction coefficient, the dominant role of high-risk obstacles such as personnel is highlighted, while also adjusting for quantity discrepancies (e.g., by integrating the highest risk level and the number of obstacles using a risk level correction coefficient, the dominant role of high-risk obstacles such as personnel is highlighted, while also adjusting for quantity discrepancies). The logarithmic correction term at time (time) prevents the superposition of multiple source risks and improves the security of complex scenarios.

[0059] Based on the above level settings and threshold determination, the output process of the differentiated early warning signal is as follows: When the obstacle risk level is no risk, a safety status prompt signal is output: such as a green static sign is displayed on the operation control interface, with no sound prompt, only textual feedback on the safety status of the surrounding environment; When the obstacle risk level is low, a mild warning signal is output: a yellow indicator light flashes at a frequency of 1Hz, accompanied by a 1Hz low-frequency buzzer (volume ≤60dB), and the operation control interface displays a yellow dot mark to indicate the obstacle location, without pop-up interference; When the obstacle risk level is medium risk, a medium warning signal is output: the orange indicator light flashes rapidly at a frequency of 2Hz, accompanied by a 2Hz medium frequency buzzer (volume range of 70~75dB), the operation control interface displays an orange ring mark to circle the obstacle area, and at the same time a simplified warning prompt box pops up (including the real-time distance between the obstacle and the execution end). When the obstacle risk level is high, a severe warning signal is output: the red indicator light stays on, accompanied by a 4Hz high-frequency buzzer (volume range of 80~85dB, which automatically drops to 75dB after 3 seconds to avoid noise interference), the operation interface highlights the obstacle and collision risk area with a red highlight, and a detailed warning pop-up window appears (including obstacle type, real-time distance, and suggested avoidance direction), while the warning information is synchronized to the cloud management module.

[0060] Based on this, it also includes: if there are multiple obstacles, the warning signal is output in the form corresponding to the highest risk level, and all obstacle icons are displayed in descending order of risk level on the operation control interface (e.g., high-risk obstacle icons are highlighted at the top), and the buzzer prompt intensity is set according to the highest risk level. When the real-time distance between the actuator and the obstacle changes, the warning signal is dynamically adjusted: if the actual distance is greater than the corresponding dynamic safe distance, the warning level is gradually reduced until the buzzer stops (the status indicator is retained); otherwise, on the basis of the corresponding risk level warning, a vibration prompt on the control handle (the vibration frequency is the same as the buzzer frequency) is added to enhance the warning perception.

Claims

1. A safety operation early warning system for the execution end, characterized in that, include: Engineering equipment execution end; The sensing module is deployed in the working area of ​​the engineering equipment. The sensing module is used to collect multi-source sensing data of the surrounding environment of the execution terminal; The data fusion module is communicatively connected to the sensing module; the data fusion module is used to perform timestamp synchronization, spatial coordinate calibration and fusion processing on multi-source sensing data to generate a three-dimensional environment model of the execution end work area. The risk assessment module is communicatively connected to the data fusion module; The risk assessment module determines the obstacle risk level and dynamic safety distance threshold based on a three-dimensional environment model, real-time operating conditions of the execution end, and preset safety rules. The graded early warning and execution control module is communicatively connected to the risk assessment module and the braking system of the engineering equipment. The graded early warning and execution control module is used to output differentiated early warning signals according to the risk level of the obstacle, and to trigger the corresponding level of safety intervention action when the risk level reaches a preset threshold.

2. The execution-end safety operation early warning system according to claim 1, characterized in that, The engineering equipment execution end includes at least: One or more of bucket, car, excavator bucket, and lifting device, and the actuator is equipped with an installation interface adapted to its working range for fixing the sensing module; The sensing module includes at least one of lidar, fisheye camera, infrared camera and ultrasonic radar.

3. A method for early warning of safe operation at the execution end of an early warning system based on any one of claims 1 to 2, characterized in that, Includes the following steps: The system utilizes a sensing module to collect multi-source sensing data of the surrounding environment of the execution terminal. This multi-source sensing data is then time-stamped, spatially calibrated, and fused to generate a three-dimensional environmental model of the execution terminal's work area. The multi-source sensing data includes obstacle locations and distances. Based on a 3D environment model, real-time operating conditions of the execution end, and historical safety accident data, the risk level of obstacles is determined and dynamic safety distance thresholds are generated. Differentiated warning signals are output based on the obstacle risk level and dynamic safety distance threshold, and corresponding level of safety intervention actions are triggered.

4. The method for early warning of safe operation at the execution end according to claim 3, characterized in that, The steps for generating the three-dimensional environment model are as follows: Employing a timestamp synchronization algorithm based on the system's unified clock. Align the timestamps of multi-source sensing data; Establish a reference coordinate system for the actuator, with the connection center point between the actuator and the robotic arm of the engineering vehicle as the origin. Defined as perpendicular to the axis of the engineering vehicle's robotic arm and passing through the origin. The virtual plane is the reference plane ; Determine the differentiated coordinate transformation matrix based on the shape and motion characteristics of the surface where the execution end is located. The original coordinates of each sensor are transformed to the reference coordinate system of the execution end to obtain the processed point cloud data. Processed toroidal image data and processed near-field distance data ; Processed point cloud data Processed toroidal image data Processed near-field range data Using infrared image data as input, through state equations and observation equations The target trajectory is predicted and corrected, various data are fused to complete the environmental information of the occluded area, and the fused data is output. That is, a three-dimensional environment model.

5. The method for early warning of safe operation at the execution end according to claim 3, characterized in that, The real-time operating conditions of the actuator include: angle adjustment, linear movement, compound linkage, heavy-load operation, precise positioning, and space-constrained operation.

6. The method for early warning of safe operation at the execution end according to claim 3, characterized in that, The process for determining the risk level of the obstacle is as follows: Based on the type of execution end and the operation scenario, a set of risk impact factors is constructed. ,in, Obstacle type The speed of the obstacle's movement. To improve the speed of execution, For workload, The dimensions of the workspace; Determine the weighting coefficients of each risk factor and construct a judgment matrix. The judgment matrix The expression is as follows: ,in, Risk impact factor Relative to risk impact factors The importance of , ; Based on the judgment matrix Determine the weighting coefficient for each risk factor. The weighting coefficient satisfy: ; Using the weighting coefficients A risk assessment value calculation model is constructed, and its expression is as follows: ,in, Standardized values ​​of each risk factor. This is the risk assessment value; Based on historical safety incident data, risk level thresholds are set; the risk assessment values ​​are... The risk level is compared with the risk level threshold to generate the corresponding obstacle risk level.

7. The method for early warning of safe operation at the execution end according to claim 3, characterized in that, The process for generating the dynamic safety distance threshold is as follows: For the real-time operating conditions of the execution end, and in combination with the equipment operating characteristics and industry safety standards, basic safety distance benchmark values ​​are set respectively. ; A working condition correction coefficient is created based on the real-time working conditions of the execution terminal. , The condition number is the real-time operating condition number of the execution end; A risk level correction coefficient is introduced based on the number of obstacles and their corresponding risk levels. , The level number corresponding to the risk level. This represents the current number of obstacles. Based on the aforementioned basic safety distance benchmark value Based on this, a dynamic safety distance calculation model is constructed, which is expressed as follows: In the formula, This is a speed correction term.

8. The method for early warning of safe operation at the execution end according to claim 3, characterized in that, The output process of the differentiated early warning signal is as follows: When the obstacle risk level is no risk, a safety status prompt signal is output; When the obstacle risk level is low, a mild warning signal is output; When the obstacle risk level is medium risk, a moderate warning signal is output; When the obstacle risk level is high, a severe warning signal is output, and the warning information is simultaneously synchronized to the cloud management module.

9. The method for early warning of safe operation at the execution end according to claim 7, characterized in that, The risk level correction coefficient The value depends on the number of obstacles currently in the path, specifically: like The risk level adjustment coefficient The value is taken as the risk basis correction coefficient for the obstacle; like First, determine the highest risk level among all obstacles. Obstacles and obtain the corresponding risk base correction coefficient The risk level adjustment coefficient The calculation formula is: 。 10. The method for early warning of safe operation at the execution end according to claim 8, characterized in that, Also includes: If there are multiple obstacles, the warning signal will be output in the form corresponding to the highest risk level. At the same time, all obstacle icons will be displayed in descending order of risk level on the operation control interface, and the buzzer prompt intensity will be set according to the highest risk level. The warning signal is dynamically adjusted as the real-time distance between the actuator and the obstacle changes.

Images