Chemical safety inspection robot control system and method based on sensor fusion

Abstract

The application relates to the technical field of machine control, and discloses a chemical safety inspection robot control system and method based on sensor fusion. The system comprises a data fusion module, a strategy construction module, a strategy analysis module, a real-time monitoring module, an optimization instruction module and an event determination module. The original data of a chemical environment are fused into standardized data, and the structural features are analyzed to identify risk factors. A safety situation map is constructed and mapped to a robot to determine a control strategy. When the robot starts inspection, the control strategy is analyzed to control the movement adjustment path action and the path safety inspection action of the robot. The movement adjustment path action and the path safety inspection action are coded into safety control instructions. The safety control instructions are optimized to obtain optimized control instructions. When the robot completes the inspection, an environment safety summary and an abnormal event record are obtained. The application can improve the accuracy of chemical safety inspection robot control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of machine control, and in particular to a chemical safety inspection robot control system and method based on sensor fusion. BACKGROUND

[0002] The prior art has significant deficiencies in the processing of robot environmental perception data. The multi-source heterogeneous chemical environment data is not systematically aligned in time and space and dimensionally unified, and only the original data or single-dimensional pre-processing is simply integrated, resulting in chaotic data format, missing temporal and spatial correlation, and inability to form standardized data. At the same time, the data structure features are not analyzed through multi-dimensional feature tensor synthesis, and only single-type features are extracted, making it difficult to fully capture the risk correlation information in the chemical scene, resulting in low precision and insufficient completeness of the environmental data perceived by the robot, and inability to provide a reliable scene awareness basis for subsequent inspection action control.

[0003] The prior art has prominent defects in the construction of control strategies and the optimization of instructions. The hierarchical safety situation map is not constructed based on chemical environmental risk factors, and the robot control strategy is only formulated based on fixed paths or simple environmental parameters, which cannot adapt to the complex and dynamically changing risk distribution in the chemical scene, resulting in lack of pertinence of the robot inspection path and insufficient coverage of key risk areas. The control instructions are not optimized based on real-time environmental change data, and only fixed instructions are used to drive the robot to perform inspection actions, which cannot dynamically adjust the robot's moving speed, inspection posture and other parameters, making it difficult to respond to sudden state changes in the chemical environment, and the safety record mechanism of action-environment association is not established, which cannot provide effective data support for subsequent robot inspection strategy iteration and safety control. SUMMARY

[0004] The present application provides a chemical safety inspection robot control system and method based on sensor fusion to solve the problems raised in the background art.

[0005] To achieve the above purpose, the chemical safety inspection robot control system based on sensor fusion provided by the present application is characterized in that the system comprises a data fusion module, a strategy construction module, a strategy analysis module, a real-time monitoring module, an optimization instruction module and an event determination module, wherein:

[0006] The data fusion module is used to fuse the original data of multi-source heterogeneous data in the chemical environment into standardized data, and analyze the structured features in the standardized data to identify the risk factors in the chemical environment;

[0007] The strategy construction module is configured to construct a benchmark based on the risk factors, reconstruct a security state graph of the chemical environment, obtain a security situation graph of the chemical environment, and map the security situation graph to the robot to determine a control strategy of the robot.

[0008] The strategy analysis module is configured to perform instruction analysis on the control strategy when the robot starts the inspection to control movement adjustment path actions and path security inspection actions of the robot.

[0009] The real-time monitoring module is configured to encode the movement adjustment path actions and the path security inspection actions into safety control instructions to control the robot to monitor real-time environmental change data of the chemical environment.

[0010] The optimization instruction module is configured to perform multi-level optimization on the safety control instructions based on the real-time environmental change data to obtain optimized control instructions of the robot.

[0011] The event determination module is configured to control the inspection behavior of the robot according to the optimized control instructions, and obtain an environmental security summary and an abnormal event record of the chemical environment when the robot completes the inspection.

[0012] In a preferred embodiment, when the data fusion module performs original data fusion of multi-source heterogeneous data in the chemical environment into standardized data, and analyzes the structured features in the standardized data to identify risk factors in the chemical environment, the data fusion module is specifically configured to:

[0013] Integrate environmental parameter data, equipment state data, and spatial perception data of the chemical environment collected by sensors into multi-source heterogeneous data of the chemical environment.

[0014] Align the multi-source heterogeneous data in time and space dimensions to obtain time-space aligned data of the chemical environment.

[0015] Unify the dimensions of the time-space aligned data to obtain preliminary standardized data of the chemical environment.

[0016] Unify the data formats of different scales in the preliminary standardized data to obtain standardized data of the chemical environment.

[0017] Synthesize environmental parameter features, equipment state features, and spatial distribution features in the standardized data into a multi-dimensional feature vector of the standardized data.

[0018] Use the multi-dimensional feature vector as a risk assessment benchmark to identify risk factors in the chemical environment.

[0019] In a preferred implementation, the strategy construction module, when performing the mapping of the safety situation map to the robot and determining the control strategy of the robot, is specifically configured to:

[0020] constructing a risk level distribution map of the chemical environment, with high-risk factors in the risk factors as the upper layer of the map, medium-risk factors in the risk factors as the middle layer of the map, and low-risk factors in the risk factors as the bottom layer of the map;

[0021] performing spatial smoothing interpolation processing on the risk level distribution map to obtain a continuous risk surface of the risk level distribution map;

[0022] extracting an environmental safety boundary in the continuous risk surface to obtain a safety partition map of the chemical environment;

[0023] superimposing the safety partition map and the real-time environment of the chemical environment on the topological structure to obtain a safety situation map of the chemical environment.

[0024] In a preferred implementation, the strategy construction module, when performing the mapping of the safety situation map to the robot and determining the control strategy of the robot, is specifically configured to:

[0025] analyzing safety partition information in the safety situation map to determine key areas and ordinary areas to be inspected by the robot;

[0026] taking the key areas to be inspected as the priority inspection path of the robot and taking the ordinary areas as the secondary inspection path of the robot;

[0027] integrating the priority inspection path and the secondary inspection path into an adaptive inspection path of the robot;

[0028] determining a movement speed adjustment parameter and an inspection action sequence of the robot according to the adaptive inspection path;

[0029] reconstructing the movement speed adjustment parameter and the inspection action sequence to obtain the control strategy of the robot.

[0030] In a preferred implementation, the strategy construction module, when performing the mapping of the safety situation map to the robot and determining the control strategy of the robot, is specifically configured to:

[0031] when the robot starts inspection, analyzing environmental safety elements in the control strategy to obtain an inspection task sequence of the robot;

[0032] determine a task execution sequence of the robot according to the inspection task sequence;

[0033] connect key inspection positions and transition path point positions in the task execution sequence to obtain a path node sequence of the robot;

[0034] input the path node sequence to the robot to synchronously coordinate movement adjustment path actions and path safety inspection actions of the robot.

[0035] In a preferred embodiment, the real-time monitoring module, when performing encoding of the movement adjustment path actions and the path safety inspection actions into safety control instructions to control the robot to monitor real-time environmental change data of the chemical environment, is specifically configured to:

[0036] perform instruction mapping of a basic movement primitive in the movement adjustment path actions and a monitoring enhancement primitive in the path safety inspection actions to obtain a basic instruction set of the robot;

[0037] take environmental risk features of the chemical environment as a logical reference to eliminate logical inconsistency in the basic instruction set to obtain a composite control instruction of the robot;

[0038] verify effectiveness between instruction coordination in the composite control instruction to obtain a safety control instruction of the robot;

[0039] control the robot to monitor real-time environmental change data of the chemical environment according to the safety control instruction.

[0040] In a preferred embodiment, the optimization instruction module, when performing multi-level optimization of the safety control instruction based on the real-time environmental change data to obtain an optimized control instruction of the robot, is specifically configured to:

[0041] determine a state change trend of the chemical environment according to dynamic features in the real-time environmental change data;

[0042] perform optimization weight distribution on the safety control instruction based on the state change trend to obtain an optimization priority of the safety control instruction;

[0043] construct an instruction optimization sequence of the robot according to the optimization priority to obtain a preliminary optimized instruction of the robot;

[0044] verify data consistency of the preliminary optimized instruction to obtain the optimized control instruction of the robot.

[0045] In a preferred embodiment, a weight factor of the optimization weight distribution is calculated according to the following formula: ;

[0046] Equation, is the weight factor, is the risk sensitivity coefficient of the state change trend, is the stability control coefficient of the state change trend, is the timeliness decay coefficient of the state change trend, is the risk value change amount of the dynamic feature, is the unit time change amount corresponding to the dynamic feature, is the environmental parameter change amount of the real-time environmental change data, is the instantaneous time change amount corresponding to the real-time environmental change data, is the time decay factor, is the logarithmic function.

[0047] In a preferred embodiment, the event determination module executes the optimization control instruction to control the inspection behavior of the robot, and when the robot completes the inspection, obtains the environmental safety summary and the abnormal event record of the chemical environment, and is specifically used for:

[0048] driving the robot to complete the scheduled inspection task with the optimization control instruction to obtain the execution trajectory of the robot;

[0049] spatiotemporally matching the execution trajectory with the real-time environmental change data to obtain the action-environment association combination of the robot;

[0050] evolving the action-environment association combination to obtain the environmental safety situation distribution of the chemical environment;

[0051] performing multi-dimensional feature condensation on the environmental safety situation distribution to obtain the environmental safety summary and the abnormal event record of the chemical environment.

[0052] To solve the above problems, the application also provides a chemical safety inspection robot control method based on sensor fusion, which comprises the following steps:

[0053] S1. Fusion of raw data fusion of multi-source heterogeneous data in a chemical environment into standardized data, and analysis of the structured features in the standardized data to identify risk factors in the chemical environment;

[0054] S2. Taking the risk factors as a construction reference, reconstructing the safety state of the chemical environment into a safety situation map, and mapping the safety situation map to a robot to determine the control strategy of the robot;

[0055] S3. When the robot starts the inspection, the control strategy is parsed for instruction to control the movement adjustment path action and path safety inspection action of the robot;

[0056] S4. The movement adjustment path action and the path safety inspection action are encoded as safety control instructions to control the robot to monitor real-time environment change data of the chemical environment;

[0057] S5. Based on the real-time environment change data, the safety control instructions are optimized in multiple levels to obtain optimized control instructions of the robot;

[0058] S6. According to the optimized control instructions, the inspection behavior of the robot is controlled, and when the robot completes the inspection, an environment safety summary and an abnormal event record of the chemical environment are obtained

[0059] Compared with the prior art, the present application has the following beneficial effects:

[0060] 1. The present application optimizes the robot perception ability through multi-source sensor data fusion, and improves the basic precision of inspection control. The environmental parameters, equipment states and spatial perception data of the chemical environment are integrated, and the time and space are aligned, the dimensions are unified and the format is standardized, to generate structured standardized data; the data is synthesized in multiple dimensions to accurately identify risk factors, providing comprehensive and accurate environmental perception basis for the robot, ensuring that the robot can accurately obtain the inspection scene information, and laying a reliable perception foundation for subsequent action control.

[0061] 2. The present application improves the accuracy and adaptability of the robot inspection action through dynamic strategy construction and instruction optimization. A safety situation map is constructed based on risk factors, and a control strategy containing priority inspection path, movement speed parameter and inspection action sequence is generated; the strategy is parsed and encoded into safety control instructions, and the optimized instructions are obtained through weight distribution and sequence optimization combined with real-time environment data to drive the robot to coordinate the movement adjustment and safety inspection action; after the inspection is completed, an environment safety summary and an abnormal record are generated, which not only ensures that the robot inspection action meets the safety needs of the chemical scene, but also realizes the efficient combination of inspection task and safety control. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1 The system architecture diagram of the chemical safety inspection robot control system based on sensor fusion provided by an embodiment of the present application is shown;

[0063] Figure 2 The flowchart of the chemical safety inspection robot control method based on sensor fusion provided by an embodiment of the present application is shown.

[0064] The objectives, technical solutions and advantages of the embodiments of the present application will be further described with reference to the drawings. DETAILED DESCRIPTION

[0065] In order to make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments belong to only a part of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by those of ordinary skill in the art without creative effort should belong to the range of the protection of the present application.

[0066] The terms used in the embodiments of the present application are only for the purpose of describing particular embodiments and are not intended to limit the present application. The singular forms "a", "said" and "the" used in the embodiments of the present application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. "Plural" generally includes at least two.

[0067] Depending on the context, the word "if" or "if" as used herein can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting". Similarly, depending on the context, the phrase "if determined" or "if detecting (a stated condition or event)" can be interpreted as "when determined" or "in response to determining" or "when detecting (a stated condition or event)" or "in response to detecting (a stated condition or event)".

[0068] In addition, the sequence of steps in each of the following method embodiments is only an example, and is not strictly limited.

[0069] In fact, the server device deployed by the chemical safety inspection robot control system based on sensor fusion can be composed of one or more devices. The chemical safety inspection robot control system based on sensor fusion can be implemented as a business instance, a virtual machine, or a hardware device. For example, the chemical safety inspection robot control system based on sensor fusion can be implemented as a business instance deployed on one or more devices in a cloud node. In short, the chemical safety inspection robot control system based on sensor fusion can be understood as a software deployed on a cloud node, which is used to provide a chemical safety inspection robot control system based on sensor fusion for each user terminal. Alternatively, the chemical safety inspection robot control system based on sensor fusion can also be implemented as a virtual machine deployed on one or more devices in a cloud node. The virtual machine has application software installed for managing each user terminal. Alternatively, the chemical safety inspection robot control system based on sensor fusion can also be implemented as a server composed of a plurality of same or different types of hardware devices, and one or more hardware devices are provided to provide a chemical safety inspection robot control system based on sensor fusion for each user terminal.

[0070] In an implementation form, the chemical safety inspection robot control system based on sensor fusion and the user terminal are mutually adapted. That is, the chemical safety inspection robot control system based on sensor fusion is an application installed on a cloud service platform, and the user terminal is a client that establishes a communication connection with the application; or the chemical safety inspection robot control system based on sensor fusion is implemented as a website, and the user terminal is implemented as a webpage; or the chemical safety inspection robot control system based on sensor fusion is implemented as a cloud service platform, and the user terminal is implemented as an applet in an instant messaging application.

[0071] As shown in Figure 1 FIG. 1 is a system architecture diagram of a chemical safety inspection robot control system based on sensor fusion according to an embodiment of the present application.

[0072] The chemical safety inspection robot control system based on sensor fusion 100 can be provided in a cloud server, and in an implementation form, can be one or more service devices, or can be an application installed on a cloud (such as a server of a mobile service operator, a server cluster, etc.), or can be developed as a website. According to the functions implemented, the chemical safety inspection robot control system based on sensor fusion 100 can include a data fusion module 101, a policy construction module 102, a policy analysis module 103, a real-time monitoring module 104, an optimization instruction module 105, and an event determination module 106. The modules of the present application can also be referred to as units, which refer to a series of computer program segments that can be executed by an electronic device processor and can complete a fixed function, and are stored in the memory of the electronic device.

[0073] In the embodiment of the present application, each of the above modules can be independently implemented and called by other modules in the chemical safety inspection robot control system based on sensor fusion. The calling here can be understood as that a module can be connected to multiple modules of another type and provide corresponding services for the connected multiple modules. The chemical safety inspection robot control system based on sensor fusion provided in the embodiment of the present application can adjust the application scope of the chemical safety inspection robot control system based on sensor fusion by increasing modules and directly calling without modifying program codes, realize cluster horizontal expansion, and achieve the purpose of quickly and flexibly expanding the chemical safety inspection robot control system based on sensor fusion. In actual application, the above modules can be arranged in the same device or different devices, or in a virtual device, such as a service instance in a cloud server.

[0074] The following will describe the components and specific work flow of the chemical safety inspection robot control system based on sensor fusion with reference to specific embodiments.

[0075] The data fusion module 101 is configured to fuse the original data of multi-source heterogeneous data in the chemical environment into standardized data, and analyze the structured features in the standardized data to identify the risk factors in the chemical environment.

[0076] In the embodiment of the present application, when the data fusion module fuses the original data of multi-source heterogeneous data in the chemical environment into standardized data, and analyzes the structured features in the standardized data to identify the risk factors in the chemical environment, it is specifically configured to:

[0077] Integrate the environmental parameter data, equipment state data and spatial perception data of the chemical environment collected by the sensor into the multi-source heterogeneous data of the chemical environment.

[0078] Align the multi-source heterogeneous data in time and space dimensions to obtain the time-space aligned data of the chemical environment.

[0079] Unify the dimensions of the time-space aligned data to obtain the preliminary standardized data of the chemical environment.

[0080] Unify the data formats of different scales in the preliminary standardized data to obtain the standardized data of the chemical environment.

[0081] Synthesize the environmental parameter features, equipment state features and spatial distribution features in the standardized data into a multi-dimensional feature vector of the standardized data.

[0082] The multi-dimensional feature vector is taken as a risk assessment benchmark to identify risk factors in the chemical environment.

[0083] Through various sensors deployed in the chemical environment, environmental parameter data reflecting environmental quality, equipment state data reflecting equipment operating conditions, and spatial perception data describing spatial layout and object position relationships are collected respectively. The environmental parameter data covers gas concentration, temperature, humidity, pressure, and other safety-related indicators of the chemical environment. The equipment state data includes equipment operating power, vibration amplitude, temperature, and fault warning signals. The spatial perception data includes equipment distribution position, pipeline orientation, and regional division. The three types of data are integrated in chronological order to form multi-source heterogeneous data of the chemical environment covering environmental, equipment, and spatial dimensions.

[0084] Based on a unified time reference and spatial coordinate system, the multi-source heterogeneous data is processed in time and space dimensions. In the time dimension, non-synchronous data collected by different sensors is adjusted to the same time node at a preset fixed time interval, ensuring the corresponding matching of various data at the same time node. In the spatial dimension, all data are associated with corresponding physical space positions based on the unified spatial coordinates of the chemical environment, such as binding the gas concentration data of a certain area with the equipment state data within that area, so that the data can accurately correspond to specific spatial ranges. Ultimately, the time-space aligned data of the chemical environment is obtained.

[0085] The dimensions of various indicators in the time-space aligned data are sorted out, and a unified conversion standard is developed for indicators of different dimensions. For physical quantity indicators, the standard units commonly used in the industry are converted, such as converting different pressure units to standard atmospheric pressure and converting different temperature units to Celsius. For non-physical quantity indicators, a unified description specification and quantization rule are defined to convert them into standardized numerical forms, ensuring that all data are consistent in dimensions and eliminating the impact of dimensional differences on subsequent data processing. Preliminary standardized data of the chemical environment is obtained.

[0086] The scale-related attributes of various data in the preliminary standardized data, such as storage format, data precision, and coding method, are checked, and a unified format specification is developed for data of different scales. The storage structure of the data is unified, using the same file format and data organization method. The precision of the data is standardized, with different precision data being selected or completed to ensure consistent data precision. The coding rules of the data are unified, converting data of different coding methods to a preset standard coding, so that all types of data are completely unified in format. Ultimately, the standardized data of the chemical environment is obtained.

[0087] The environmental parameter features, equipment state features and spatial distribution features in the standardized data are extracted, wherein the environmental parameter features focus on the value change trend and range of indicators such as gas concentration and temperature, the equipment state features pay attention to the stability and abnormal signals of equipment operation parameters, and the spatial distribution features emphasize the position correlation and regional distribution law of spatial elements such as equipment and pipelines. In a tensor synthesis manner, the three types of features are dimensionally fused according to their internal logical correlation, and the dispersed single-dimensional features are integrated into overall features that can comprehensively reflect the multiple aspects of the chemical environment, forming a multi-dimensional feature vector of the standardized data.

[0088] The multi-dimensional feature vector is used as a risk assessment benchmark, combined with the safety operation standards of the chemical industry, historical risk cases and risk identification rules, to analyze each feature in the feature vector one by one. It is determined whether the environmental parameter features exceed the safety threshold, whether the equipment state features have abnormal fluctuations, and whether the spatial distribution features have safety hidden danger correlations, such as abnormal gas concentration features in a certain area and corresponding equipment state features showing faults, and at the same time, the spatial distribution features show that the area is adjacent to a personnel-intensive area, then it is determined as a risk factor. By comprehensively analyzing the abnormal conditions and mutual correlations of various features, the risk factors in the chemical environment are accurately identified.

[0089] The beneficial effect is that the environmental parameter data, equipment state data and spatial perception data of the chemical environment collected by the sensor are integrated into multi-source heterogeneous data of the chemical environment, which can comprehensively collect the core data dimensions required for chemical safety monitoring. The environmental parameter data reflects the basic safety conditions of the chemical environment such as gas concentration, temperature and humidity, the equipment state data reflects the running stability of the core production equipment, and the spatial perception data clearly shows the spatial layout correlation of equipment, pipelines and other elements. After the three are integrated according to the time sequence correlation, a complete data set covering the environment, equipment and space is formed, avoiding the one-sidedness of risk identification caused by the lack of single-dimensional data, and providing comprehensive raw data support for subsequent data standardization processing.

[0090] The multi-source heterogeneous data is aligned in time and space to obtain the time-space aligned data of the chemical environment, which can eliminate the misalignment interference of the data in time and space dimensions. In the time dimension, the non-synchronous collection data of different sensors is adjusted to a unified time node to ensure that the environment, equipment and space data at the same time are accurately corresponding; in the spatial dimension, based on the unified coordinate system of the chemical environment, various types of data are bound to specific physical locations, such as the correlation of gas concentration data in a certain area with the operation data of equipment in that area, avoiding the distortion of data correlation caused by time-space misalignment, and laying a time-space consistent data foundation for dimension unification.

[0091] The dimensional uniformization of the spatio-temporal alignment data obtains the preliminary standardized data of the chemical environment, which can eliminate the unit difference obstacles of different types of data. For physical quantity data, different units are uniformly converted according to the industry standard, such as converting various pressure units into standard atmospheric pressure and temperature units into Celsius; for non-physical quantity data, they are converted into standardized numerical form through preset quantization rules, ensuring that all data remain consistent in the measurement scale, avoiding feature analysis errors caused by dimensional confusion, and improving the comparability of data and the accuracy of subsequent processing.

[0092] The standardized data of the chemical environment can be obtained by uniformly preliminary standardized data of different scales, which can realize the structured integration of data. The storage format, precision requirement, coding method and other scale attributes of the data are combed, and the same file structure, data precision and coding rules are uniformly adopted to structurally reconstruct different formats of data, such as completing numerical data of different precisions to uniform precision and converting text data of different encodings to standard encoding, which eliminates the difficulty of data integration caused by format difference, forms standardized data that can be directly used for feature analysis, and provides high-quality data carrier for structured feature extraction.

[0093] The multi-dimensional feature vector of the standardized data can be obtained by tensor synthesis of the environmental parameter features, equipment state features and spatial distribution features in the standardized data, which can realize the deep fusion of multi-dimensional features. The numerical change trend of environmental parameters, the stability features of equipment state, the position correlation law of spatial distribution and other core information are extracted, and the dispersed single-dimensional features are integrated into overall feature vectors with dimensional attributes through tensor synthesis, avoiding the risk of missing risk correlation caused by single feature analysis, and making the feature vector fully reflect the comprehensive safety state of the chemical environment, providing comprehensive feature support for risk identification.

[0094] The multi-dimensional feature vector can be used as a risk assessment benchmark to identify risk factors in the chemical environment, which can improve the accuracy and comprehensiveness of risk identification. Combined with the safety standards of the chemical industry, historical risk cases and risk identification rules, the features in the feature vector are analyzed cooperatively to determine whether the environmental parameters exceed the safety threshold, whether the equipment state has abnormal fluctuations, and whether the spatial layout has safety hidden danger correlation, such as abnormal gas concentration features in a certain area and corresponding equipment state features showing faults, and spatial distribution features showing that the area is adjacent to a personnel-intensive area, which accurately determines the risk factors. This comprehensive analysis method avoids the one-sidedness of single feature evaluation, ensures that there is no omission and misjudgment in risk factor identification, and provides reliable risk basis for subsequent control strategy construction.

[0095] The policy construction module 102 is configured to construct a safety state graph of the chemical environment by taking the risk factors as a construction benchmark, and map the safety state graph to the robot to determine a control strategy of the robot.

[0096] In the embodiment of the present application, when the policy construction module is used to construct a safety state graph of the chemical environment by taking the risk factors as a construction benchmark, the policy construction module is specifically configured to:

[0097] construct a risk level distribution graph of the chemical environment by taking a high-risk factor in the risk factors as an upper layer of the graph, taking a medium-risk factor in the risk factors as a middle layer of the graph, and taking a low-risk factor in the risk factors as a bottom layer of the graph.

[0098] perform spatial smoothing interpolation processing on the risk level distribution graph to obtain a continuous risk surface of the risk level distribution graph.

[0099] extract an environmental safety boundary in the continuous risk surface to obtain a safety partition graph of the chemical environment.

[0100] superimpose the safety partition graph and a real-time environment of the chemical environment on a topological structure to obtain the safety state graph of the chemical environment.

[0101] When the policy construction module is used to map the safety state graph to the robot to determine the control strategy of the robot, the policy construction module is specifically configured to:

[0102] analyze safety partition information in the safety state graph to determine a key area to be inspected and a common area of the robot.

[0103] take the key area to be inspected as a priority inspection path of the robot, and take the common area as a secondary inspection path of the robot.

[0104] integrate the priority inspection path and the secondary inspection path into a self-adaptive inspection path of the robot.

[0105] determine a movement speed adjustment parameter and an inspection action sequence of the robot according to the self-adaptive inspection path.

[0106] reconstruct a strategy of the movement speed adjustment parameter and the inspection action sequence to obtain the control strategy of the robot.

[0107] Firstly, the classification basis of risk factors is determined. Combined with the chemical environmental safety standards and historical risk disposal cases, the risk factors that may cause major safety accidents and serious harm to personnel and equipment are defined as high-risk factors, the risk factors that may lead to general safety problems and affect the normal production but have limited harm are defined as medium-risk factors, and the risk factors that only cause slight impact, can be quickly disposed and have no serious consequences are defined as low-risk factors. Taking the physical space layout of the chemical environment as the basic framework of the graph, the high-risk factors are arranged in the upper layer of the graph to highlight their core risk warning position, the medium-risk factors are placed in the middle layer of the graph, and the low-risk factors are arranged in the bottom layer of the graph. Through hierarchical division, the severity and priority of different risks are clearly presented, and the risk level distribution graph of the chemical environment is constructed.

[0108] The discrete risk data points in the risk level distribution graph are supplemented and improved by using spatial smoothing interpolation. Based on the existing risk data points in the risk level distribution graph, the risk levels of the data blank areas are filled by reasonably inferring the risk levels of the adjacent areas, the risk levels of the adjacent areas, the environmental parameters, the equipment distribution and other related information, so that the originally dispersed risk points are connected to form continuous risk distribution areas. In the processing process, the transition of the risk levels of the adjacent areas is ensured to be natural, and abrupt level mutations are avoided. Finally, the continuous risk surface of the risk level distribution graph is obtained, and the spatial distribution and change trend of the risk in the chemical environment are intuitively presented.

[0109] Based on the continuous risk surface of the risk level distribution, the environmental safety boundary determination standard is set, that is, the critical line between different risk level areas is the environmental safety boundary. By analyzing the change law of the risk level in the continuous risk surface, the boundary positions between the high-risk area and the medium-risk area, and the medium-risk area and the low-risk area are identified. These boundary positions form the lines of the environmental safety boundary. According to these boundaries, the chemical environment is divided into different safety areas, each area corresponds to a clear risk level, and then the safety zoning graph of the chemical environment is obtained, and the safety status of each area is clearly defined.

[0110] The real-time environmental data of the chemical environment is retrieved, including real-time environmental parameters, equipment operating status, personnel activity positions and other information, and the real-time environmental topology reflecting the actual situation of the current chemical environment is constructed. The safety zoning graph and the real-time environmental topology are accurately superimposed according to the spatial coordinates, so that each safety area in the safety zoning graph can accurately correspond to the specific physical area, equipment and personnel activity range in the real-time environment. In the superposition process, the spatial proportion and position relationship of the two are kept consistent, ensuring that the boundary of the safety zoning is consistent with the geographical features and equipment layout in the real-time environment. Finally, the safety situation map of the chemical environment is obtained, which comprehensively and timely presents the safety status and risk distribution of the chemical environment.

[0111] The safety partition information in the safety situation diagram is deeply analyzed, and the risk level, area range, risk factor type and distribution density corresponding to each partition are extracted. According to the chemical environmental safety inspection standard, the high-risk level area, the area with dense risk factors and the area that may cause serious consequences, and the area containing key equipment and core production link are determined as the key areas to be inspected by the robot; the low-risk level area, the area with scattered risk factors and the area with slight harm, and the non-core production auxiliary area are determined as the ordinary area of the robot, and the inspection priority and focus of the two types of areas are clear.

[0112] Based on the spatial layout of the chemical environment, combined with the distribution position and area size of the key areas to be inspected, the priority inspection path that can fully cover all key areas with the shortest path and the least time consumption is planned. The planning of the priority inspection path needs to avoid obstacles in the environment to ensure smooth movement of the robot, while ensuring sufficient inspection time in the key area to allow detailed investigation of risk hidden dangers; for ordinary areas, the secondary inspection path connecting each key area and covering all ordinary areas is planned, which shortens the moving distance as much as possible under the premise of ensuring complete coverage, avoids occupying too many inspection resources, and clarifies the sequence and path of the robot inspection.

[0113] The spatial correlation and time connection of the priority inspection path and the secondary inspection path are sorted out, and the two paths are integrated according to the principle of optimal inspection efficiency. In the integration process, the core position of the priority inspection path is ensured, and the secondary inspection path is supplemented to form a seamless connection with the priority inspection path, avoiding path repetition or omission. According to the distribution of key areas and ordinary areas, the path order is reasonably adjusted, so that the robot can enter the adjacent ordinary area for inspection after completing the inspection of a key area, and then go to the next key area, forming a coherent and efficient self-adaptive inspection path of the robot, taking into account the comprehensiveness and efficiency of the inspection.

[0114] According to the environmental characteristics, regional risk level and inspection requirements of different sections in the self-adaptive inspection path, the moving speed adjustment parameters of the robot are determined. In the high-risk area of the priority inspection path, a slower moving speed is set to ensure that the robot has enough time to collect environmental data in detail and investigate risk hidden dangers; in the ordinary area of the secondary inspection path, a relatively faster moving speed is set to improve the overall inspection efficiency; for special sections such as turns and climbs in the path, the moving speed is adjusted to adapt to the stability of movement. At the same time, combined with the inspection focus of each area, the corresponding inspection action sequence is developed, such as performing device state close-range detection, environmental parameter multi-point collection, risk factor accurate positioning in key areas, and performing routine environmental monitoring, equipment appearance inspection and other actions in ordinary areas, to clarify the specific inspection operation of the robot in different positions.

[0115] The determined mobile speed adjustment parameter and the patrol action sequence are systematically reconstructed, and the speed parameter and the action sequence are organically combined according to the control logic and the execution specification of the robot. It is clear that the robot adjusts the speed at each section of the adaptive patrol path, adjusts to what speed, and the specific patrol action to be performed at the speed, ensuring that the speed adjustment and action execution are coordinated and consistent. At the same time, the safety operation rules of the robot are integrated, such as speed emergency adjustment and action emergency switching when encountering sudden risks, etc. The reconstructed strategy not only meets the requirements of the patrol task, but also guarantees the safe operation of the robot itself, and finally forms a complete and feasible control strategy of the robot, guiding the robot to accurately and efficiently complete the chemical environment patrol work.

[0116] The beneficial effect is that the high, medium and low risk factors in the risk factors correspond to the top, middle and bottom layers of the risk level distribution map of the chemical environment on the map, which can intuitively present the severity and spatial level relationship of different risks. By clearly defining the top warning position of high-risk factors, the middle transition role of medium-risk factors, and the bottom basic attributes of low-risk factors, the staff can quickly identify the core risk sources and secondary risk points, avoid the deviation of safety control focus caused by risk level confusion, provide clear level basis for subsequent risk area division, and ensure that risk control resources can be preferentially invested in high-risk areas.

[0117] The spatial smoothing interpolation processing of the risk level distribution map obtains a continuous risk surface, which can eliminate the distribution blind area caused by discrete risk data points. Based on the existing risk data points, referring to the risk level, environmental parameters and equipment distribution characteristics of the adjacent area, the risk level of the data blank area is reasonably inferred and filled, so that the originally isolated risk points are connected to form a continuous and naturally transitioned risk distribution area, avoiding unreasonable risk level mutations, and intuitively presenting the gradual change rule of risks from high to low in the chemical environment, providing a complete risk distribution carrier for precise extraction of the environmental safety boundary.

[0118] The extraction of the environmental safety boundary in the continuous risk surface obtains the safety partition map of the chemical environment, which can realize the precise division of the risk area. By analyzing the critical position of different risk level areas in the continuous risk surface, the boundary line between high risk and medium risk, and medium risk and low risk is determined as the environmental safety boundary, and the chemical environment is divided into multiple independent areas with clear safety levels according to these boundaries, each area corresponds to a unique risk level and control requirement, avoiding safety control loopholes caused by fuzzy area division, and making the safety partition directly guide the subsequent patrol path planning.

[0119] The safety partition map is superimposed on the real-time environment of the chemical environment to obtain a safety situation map of the chemical environment, which can realize the deep combination of risk distribution and actual environment. The real-time environment topology structure includes dynamic information such as current environmental parameters, equipment operating state, and personnel activity position, which is accurately superimposed through spatial coordinates, so that each risk area in the safety partition map can correspond to the specific physical location, equipment, and personnel activity range in the real-time environment. For example, the high-risk partition is bound to the real-time fault state of a core device, so that the safety situation map can not only present the risk distribution, but also reflect the real-time environment dynamics, providing a comprehensive and practical safety state basis for subsequent robot control strategy formulation, ensuring that the strategy can accurately adapt to the real-time safety needs of the chemical environment.

[0120] Analyzing the safety partition information in the safety situation map to determine the key areas and ordinary areas to be inspected by the robot can make the robot inspection target more accurate. By extracting the risk level, area range, and risk factor density in the safety partition information, the area with high risk, dense risk factors, and containing key equipment is determined as the key area to be inspected, and the auxiliary area with low risk and scattered risk factors is determined as the ordinary area, avoiding the core risk missed inspection caused by the fuzzy priority of the inspection area, and ensuring that the inspection resources are preferentially allocated to the high-value safety monitoring area.

[0121] Taking the key area to be inspected as the priority inspection path and the ordinary area as the secondary inspection path can establish an orderly inspection resource allocation mechanism. When planning the priority inspection path, all key areas are covered and sufficient inspection time is reserved to ensure that high-risk areas can be monitored in detail. The secondary inspection path aims to connect key areas and cover ordinary areas, ensuring integrity while shortening unnecessary movement distance and avoiding waste of inspection time. This path division allows the robot to carry out inspection according to the risk priority, improving the monitoring depth and efficiency of high-risk areas.

[0122] Integrating the priority inspection path and the secondary inspection path into the adaptive inspection path of the robot can achieve a balance between comprehensive and efficient inspection. During integration, the two paths are connected in terms of spatial correlation and temporal continuity, allowing the robot to complete a key area inspection, then enter the adjacent ordinary area for inspection, and then move to the next key area, avoiding path repetition or omission. The adaptive inspection path ensures complete coverage of key areas and basic monitoring of ordinary areas, ensuring that the robot can complete comprehensive and efficient inspection tasks within a limited time.

[0123] According to the adaptive inspection path, the moving speed adjustment parameter and the inspection action sequence of the robot are determined, so that the robot inspection action can adapt to the safety requirements of different areas. In the high-risk area of the priority inspection path, a slower moving speed is set, and fine inspection actions such as device close-range detection and environment parameter multi-point collection are performed; in the ordinary area of the secondary inspection path, a faster moving speed is set, and basic actions such as routine environment monitoring and device appearance inspection are performed; for special road sections such as turning and climbing in the path, the moving speed is adjusted to ensure stability. This customized setting of parameters and actions enables the robot to accurately and efficiently inspect different areas.

[0124] The control strategy of the robot is obtained by reconstructing the moving speed adjustment parameter and the inspection action sequence, which can form a complete and executable inspection guidance scheme. During the reconstruction process, the speed parameter and the action sequence are integrated according to the time sequence and the logical relationship, and the speed adjustment timing and the action execution sequence of the robot in each section of the path are determined, while the emergency handling logic such as speed emergency adjustment and action emergency switching in case of parameter abnormalities is integrated. The final control strategy not only covers the conventional inspection process, but also has the ability to respond to risks, ensuring that the robot can accurately perform the inspection task according to the strategy and fully guarantee the reliability and effectiveness of the chemical environment safety monitoring.

[0125] The strategy analysis module 103 is configured to perform instruction analysis on the control strategy when the robot starts inspection, to control the moving adjustment path action and the path safety inspection action of the robot.

[0126] In the embodiment of the present application, when the strategy analysis module performs instruction analysis on the control strategy when the robot starts inspection, to control the moving adjustment path action and the path safety inspection action of the robot, it is specifically used for:

[0127] When the robot starts inspection, the environmental safety elements in the control strategy are analyzed to obtain the inspection task sequence of the robot.

[0128] According to the inspection task sequence, the task execution sequence of the robot is determined.

[0129] The key inspection positions and transition path point positions in the task execution sequence are connected to obtain the path node sequence of the robot.

[0130] The path node sequence is input to the robot to synchronize and coordinate the moving adjustment path action and the path safety inspection action of the robot.

[0131] When the robot starts the inspection program, the environmental safety elements contained in the control strategy are comprehensively disassembled, which include the risk level of each area in the chemical environment, the type of environmental parameters that need to be monitored, the inspection requirements of key equipment, and the safety operation specifications, etc. According to the inspection target corresponding to each environmental safety element, the control strategy is refined into a series of specific and executable independent inspection tasks, each of which clearly specifies the specific content of the inspection, the execution standard and the goal to be achieved, for example, a task requires to detect the gas concentration in a specific area and record the data, another task requires to check the running state of a key equipment. All these independent inspection tasks are sorted according to their internal correlation to form a logical and comprehensive inspection task sequence of the robot.

[0132] The importance, execution difficulty and dependency relationship between tasks in the inspection task sequence are analyzed one by one. For the inspection tasks involving high-risk areas and core components of key equipment, the highest execution priority is given; for the routine inspection tasks of ordinary areas, the routine priority is set; at the same time, the connection logic between tasks is considered, if the execution result of a task is the premise of another task, the former is placed in front. According to these priorities and dependencies, the sequence of the robot to complete each inspection task is determined, ensuring that high-priority tasks are executed first, and the execution process of all tasks is logical, smooth and efficient, and finally the execution order of the robot is determined.

[0133] According to the determined task execution order, the key inspection positions corresponding to each inspection task are extracted, including the core monitoring points of high-risk areas, the key inspection parts of key equipment, and the points with high incidence of environmental parameter abnormalities, etc. Each key inspection position corresponds to a specific physical space coordinate. At the same time, combined with the spatial layout of the chemical environment, reasonable transition path point positions are planned between adjacent key inspection positions, which should avoid obstacles, ensure smooth movement path, and enable the robot to reach the next key inspection position from the current one with the shortest distance. According to the task execution order, all key inspection positions and transition path point positions are connected in order to form a continuous and orderly path node sequence of the robot, providing clear coordinate guidance for the movement of the robot.

[0134] The complete path node sequence is transmitted to the control system of the robot, and after the control system receives the node sequence, the coordinate information, movement direction and stay requirement of each node are analyzed. According to the guidance of the node sequence, the movement system of the robot adjusts the movement speed, turning angle and travel route of the robot to ensure accurate arrival at each path node; at the same time, the inspection system of the robot starts the preset inspection action at the corresponding key inspection position according to the task execution order, such as starting the sensor to collect environmental data, starting the camera to shoot the equipment state, and executing the equipment parameter detection. In the whole process, the control system synchronously coordinates the movement adjustment path action and the path safety inspection action in real time to ensure that the robot accurately completes the inspection task during movement, and the movement state is stable during the execution of the inspection action, and the two are seamlessly connected and cooperated to ensure that the inspection work is carried out efficiently, accurately and safely.

[0135] The beneficial effect is that when the robot starts the inspection, the environmental safety elements in the control strategy are analyzed to obtain the robot inspection task sequence, which can convert the abstract control strategy into specific executable task units. The environmental safety elements cover core information such as regional risk level, key monitoring parameters and equipment inspection requirements. By disassembling these elements, the control strategy is refined into a series of independent inspection tasks, each of which clearly specifies the inspection content, execution standard and target, such as "detecting the gas concentration in a high-risk area" and "checking the vibration state of a key equipment", avoiding the ambiguity of robot execution direction caused by abstract strategy, and providing a clear execution basis for subsequent task sequencing.

[0136] According to the inspection task sequence, the task execution order of the robot is determined, which can ensure that the inspection tasks are orderly pushed forward according to the safety priority. By analyzing the risk level of the task and the dependency relationship between tasks, high-risk area related tasks and core equipment inspection tasks are set as high priority and are prioritized, and ordinary area routine monitoring tasks are post-positioned according to the logical order, ensuring that the robot prioritizes the handling of key safety requirements and avoids the delay in checking high-risk hidden dangers caused by chaotic task order, thereby improving the safety and pertinence of the inspection.

[0137] Connecting the key inspection positions in the task execution order and the transition path point positions obtains the path node sequence of the robot, which can provide precise spatial coordinate guidance for the movement of the robot. The key inspection positions correspond to the core execution points of each task, and the transition path point positions are optimized movement nodes between adjacent key positions. By connecting the two types of nodes in turn according to the task order, a continuous and orderly coordinate sequence is formed, avoiding movement deviation caused by ambiguous path and ensuring that the robot can accurately arrive at each inspection point.

[0138] The path node sequence is input to the robot to synchronously coordinate movement adjustment path actions and path safety inspection actions, which can realize seamless connection of robot movement and inspection. After the robot control system analyzes the node sequence, the movement module is driven to adjust speed and steering according to node coordinates, to ensure smooth arrival at the target position; at the same time, preset inspection actions are automatically triggered at key inspection positions, and the coordination of movement and inspection actions is monitored in real time, such as slowing down the movement speed in high-risk areas to prolong the inspection time, and speeding up the movement in ordinary areas to improve efficiency, to avoid missed inspection or inefficient movement due to disconnection of actions, and to ensure accurate, efficient and safe inspection process.

[0139] The real-time monitoring module 104 is configured to encode the movement adjustment path action and the path safety inspection action into safety control instructions to control the robot to monitor real-time environmental change data of the chemical environment.

[0140] In the embodiment of the present application, when the real-time monitoring module encodes the movement adjustment path action and the path safety inspection action into safety control instructions to control the robot to monitor real-time environmental change data of the chemical environment, it is specifically used for:

[0141] The basic movement primitives in the movement adjustment path action and the monitoring enhancement primitives in the path safety inspection action are mapped into instructions to obtain a basic instruction set of the robot;

[0142] The environmental risk characteristics of the chemical environment are taken as a logical reference to eliminate logical inconsistency in the basic instruction set, to obtain a composite control instruction of the robot;

[0143] The effectiveness of instruction coordination in the composite control instruction is verified to obtain a safety control instruction of the robot;

[0144] According to the safety control instruction, the robot monitors real-time environmental change data of the chemical environment.

[0145] The movement adjustment path action is disassembled to extract basic movement primitives, which include straight line travel, turning, acceleration, deceleration, stopping and other core action units that constitute robot movement, and each primitive corresponds to a clear mechanical execution logic. At the same time, the path safety inspection action is split to extract monitoring enhancement primitives, which include sensor activation, data acquisition, device detection, image capture, parameter recording and other core operation units that directly serve environmental monitoring. According to the instruction mapping rules preset by the robot control system, each basic movement primitive and monitoring enhancement primitive is converted into a system recognizable instruction code, each primitive is accurately corresponding to a unique instruction code, all instruction codes are integrated to form a basic instruction set of the robot, and the instruction set can completely cover the core execution requirements of the two types of actions.

[0146] The environmental risk characteristics of the chemical environment are retrieved, including the risk levels of different areas, the attributes of high-risk substances, the safety thresholds of equipment operation, the critical ranges of environmental parameters, etc., which are used as the basis for judging the logical rationality of the instructions. Each instruction in the basic instruction set is checked one by one to analyze whether there is a logical conflict between the instructions, for example, one instruction requires the robot to accelerate through a high-risk area, while another instruction requires the robot to slow down in that area for detailed monitoring. Such contradictory instructions are logically inconsistent. For the discovered logical conflicts, adjustments are made by referring to the environmental risk characteristics, keeping the instructions that meet the safety monitoring requirements, and modifying or removing the conflicting instructions to ensure that all instructions are consistent and coordinated around the core goal of environmental risk monitoring. Finally, the composite control instructions for the robot are obtained.

[0147] A verification standard for instruction coordination effectiveness is established, which covers the rationality of the timing of instruction execution, the fluency of action connection, the integrity of monitoring data, and the safety of robot operation, etc. According to the execution order of the composite control instructions, the running process of the instructions in the robot control system is simulated to check whether the execution result of the previous instruction can provide a good foundation for the next instruction, for example, whether the moving instruction can accurately deliver the robot to the monitoring position to ensure the effective execution of the monitoring instruction. At the same time, it is verified whether the combination of instructions will cause the robot to appear mechanical jam, data collection omission or safety hazards, such as whether fast turning will affect the stable monitoring of the sensor. Through comprehensive simulation and verification, instruction combinations that are not effective in coordination are removed, and instruction combinations that are effective in coordination are retained and optimized to obtain safe control instructions for the robot.

[0148] The safe control instructions are transmitted to the core control unit of the robot, which parses and executes each instruction in order according to the instruction sequence. Under the control of the moving-related instructions, the robot accurately adjusts the moving state and smoothly travels along the preset path to ensure reaching each monitoring point. Under the drive of the monitoring-related instructions, various sensors and monitoring devices of the robot start on time to accurately collect real-time environmental change data such as gas concentration, temperature, humidity, and equipment vibration in the chemical environment. The control unit monitors the execution status of the instructions in real time to ensure that the moving actions and monitoring actions are coordinated, the data collection process is continuous and complete, and the collected real-time environmental change data are transmitted to the background system in real time to realize dynamic monitoring and data recording of the chemical environment, fully ensuring the accuracy and timeliness of the monitoring work.

[0149] The beneficial effect is that the basic moving primitives and monitoring enhancement primitives are mapped to instructions to obtain a basic instruction set, which can realize the conversion of patrol actions into machine recognizable instructions. Each action primitive is accurately mapped to a unique code through preset rules to ensure that the instruction set completely covers the moving and monitoring requirements, avoiding execution deviations caused by missing action-instruction correspondence, and laying a foundation for subsequent instruction optimization.

[0150] Eliminating the logical inconsistency of the basic instruction set based on the environmental risk characteristics can ensure that the instruction logic is consistent with the safety requirements. Referring to the risk level of each area, the safety threshold of the equipment and other characteristics, the conflicting instructions such as 'accelerate through high-risk areas' and 'decelerate monitoring' are corrected to ensure that all instructions are coordinated around risk monitoring, and to avoid instruction conflicts causing inspection failure or safety hazards.

[0151] Verifying the cooperative effectiveness of the composite control instruction obtains the safety control instruction, which can ensure that the instruction combination is executable and efficient. Simulation verification is performed from the dimensions of time sequence connection, action fluency and data integrity, and the combination of cooperative invalidity is eliminated, the efficient instruction is reserved, and the monitoring data missing or equipment failure caused by cooperative problems is avoided, and the smooth driving of the instruction is ensured.

[0152] According to the safety control instruction, the robot can monitor the real-time environmental change data, and the dynamic and accurate monitoring of the environmental safety state can be realized. The moving system driven by the instruction travels along the preset path, triggers the monitoring system to collect data, and transmits the data to the background in real time, ensures the continuity and accuracy of the data, provides high-quality real-time data support for subsequent instruction optimization and safety evaluation, and ensures the timeliness and reliability of the monitoring.

[0153] The optimization instruction module 105 is configured to perform multi-level optimization on the safety control instruction based on the real-time environmental change data, and obtain the optimization control instruction of the robot.

[0154] In the embodiment of the present application, when the optimization instruction module performs multi-level optimization on the safety control instruction based on the real-time environmental change data, and obtains the optimization control instruction of the robot, it is specifically used for:

[0155] According to the dynamic characteristics in the real-time environmental change data, the state change trend of the chemical environment is determined.

[0156] Based on the state change trend, the safety control instruction is optimized and weighted, and the optimization priority of the safety control instruction is obtained.

[0157] According to the optimization priority, the instruction optimization sequence of the robot is constructed, and the preliminary optimization instruction of the robot is obtained.

[0158] The data consistency of the preliminary optimization instruction is verified, and the optimization control instruction of the robot is obtained.

[0159] The weight factor of the optimization weight distribution is calculated as follows:

[0160] ;

[0161] The formula is, a weight factor for the state change trend, a risk sensitivity coefficient for the state change trend, a stability control coefficient for the state change trend, a timeliness decay coefficient for the state change trend, a risk value change amount of the dynamic feature, a unit time change amount corresponding to the dynamic feature, an environmental parameter change amount of the real-time environmental change data, an instantaneous time change amount corresponding to the real-time environmental change data, a time decay factor, a logarithmic function.

[0162] Deep analysis of the dynamic features contained in the real-time environmental change data, including the rise and fall amplitude of gas concentration, the fluctuation law of temperature and humidity, the change trend of equipment operation parameters, the appearance and diffusion trajectory of risk factors, etc. Combined with the safe operation standards of chemical environment and the change law of historical data, these dynamic features are systematically sorted out to determine whether each data tends to be safe and stable, maintains the existing state, or develops towards risk aggravation. For example, if the gas concentration in a certain area continues to rise and exceeds the normal fluctuation range, it is determined that the state of this area tends to high risk; if the equipment vibration parameter remains within the safety threshold and has no obvious fluctuation, it is determined that the equipment state tends to be stable, and the state change trend of the chemical environment is determined through comprehensive analysis.

[0163] Based on the determined state change trend of the chemical environment, the weight of the safety control instruction is optimized according to the different needs of different trends for safety monitoring. For the area whose state develops towards risk aggravation, the corresponding patrol monitoring instruction is given a higher weight to ensure that the robot prioritizes the monitoring frequency and data acquisition accuracy of this area; for the area whose state tends to be stable, the original monitoring instruction weight is maintained or the non-core monitoring instruction weight is appropriately reduced; for the area whose state fluctuates greatly but does not reach the risk threshold, the weight of the mobile adjustment instruction is increased to ensure that the robot can flexibly adjust the path to cover more monitoring points. Through this targeted weight allocation, the importance of each safety control instruction in the optimization process is determined, and the optimization priority of the safety control instruction is obtained.

[0164] The safety control instructions are sorted and integrated in order of optimization priority from high to low to build the optimized instruction sequence of the robot. The instructions with the highest priority are arranged at the front end of the sequence to ensure that the robot performs critical optimization operations first, such as intensive monitoring instructions for high-risk trend areas. The instructions with medium priority are arranged in logical order to ensure the coordination of monitoring and movement actions. The instructions with lower priority are arranged at the back end of the sequence to avoid occupying critical monitoring resources. During the sorting process, the logical coherence of instruction execution is also considered to ensure that the execution results of a previous instruction can support the next instruction, for example, executing a path adjustment instruction before executing the corresponding monitoring instruction, forming a preliminary optimized instruction of the robot with clear logic and explicit priority.

[0165] A data consistency verification standard is established, which covers the data collection integrity after instruction execution, the matching degree of monitoring data and actual environmental state, the correlation of data collected by different sensors, the adaptability of instruction parameters and data changes, etc. The preliminary optimized instructions are input into the robot simulation system or tested in a small range during actual inspection to collect monitoring data and robot operation data generated after instruction execution. The consistency of these data and real-time environmental change data is compared and analyzed to check whether the instructions can accurately respond to environmental state changes, for example, whether the intensive monitoring instructions can collect more comprehensive risk area data and whether the path adjustment instructions can accurately guide the robot to the target monitoring point. If the data is inconsistent, the corresponding instruction parameters or execution logic are traced back and corrected until all the data after instruction execution is consistent with the actual environmental state, and finally the optimized control instructions of the robot are obtained.

[0166] The risk sensitivity coefficient of the state change trend is derived from the preset definition of the safety monitoring requirements of the chemical environment. Combined with the safety standards of the chemical industry, historical risk disposal experience, and the influence degree of different risk types on elevator safety, the fixed value of this coefficient is determined to ensure that it can accurately reflect the sensitivity of the safety control instruction to the risk factor in the state change trend, providing a risk correlation benchmark for weight factor calculation.

[0167] The stability control coefficient of the state change trend is set based on the state stability requirements of the normal operation of the chemical environment. Referring to the stability threshold of the chemical equipment operation and the normal fluctuation range of the environmental parameters, the specific value of this coefficient is determined by analyzing the execution effect of the safety control instruction in the stable state, so that it can effectively measure the action strength of the safety control instruction on maintaining the stability of the environmental state.

[0168] The time-effectiveness decay coefficient of the state change trend is determined according to the execution time-effectiveness feature of the safety control instruction. In combination with the change speed of the risk factors in the chemical environment, the effective time window of the data monitoring, the decay law of the instruction execution effect with the lapse of time is analyzed, and then the fixed value of the coefficient is set to ensure that it can accurately reflect the influence of the time factor on the optimization weight of the instruction.

[0169] The risk value change amount of the dynamic feature is derived from the risk quantization analysis of the dynamic feature in the real-time environmental change data. The core indexes related to the risk in the dynamic feature, such as the exceeding amplitude of the gas concentration and the abnormal value of the equipment vibration, are extracted, and the numerical difference of the indexes at different time points is compared to calculate the change amplitude of the risk value, which is the risk value change amount of the dynamic feature.

[0170] The unit time change amount corresponding to the dynamic feature is calculated by the dynamic feature data of the fixed time interval. A preset fixed time period is selected as the unit time, the risk value change amount of the dynamic feature in the time period is counted, and the risk value change amount is divided by the unit time length to obtain the change degree of the dynamic feature in the unit time, which is the unit time change amount corresponding to the dynamic feature.

[0171] The environmental parameter change amount of the real-time environmental change data is derived from the comparative analysis of the real-time monitored environmental parameter data. The key environmental parameters in the chemical environment, such as temperature, humidity, and pressure, are selected, and the specific values of the parameters at adjacent two monitoring time points are recorded to calculate the difference between the two values, which is the environmental parameter change amount of the real-time environmental change data.

[0172] The instantaneous time change amount corresponding to the real-time environmental change data is the time interval between the adjacent two environmental parameter monitoring. Based on the sampling frequency of the robot monitoring device, the specific time points of the adjacent two collected real-time environmental change data are recorded, and the time difference between the two time points is calculated, which is the instantaneous time change amount corresponding to the real-time environmental change data.

[0173] The time decay factor is determined based on the correlation between the execution time of the safety control instruction and the monitoring period. The preset period of the chemical environment monitoring and the duration of the data effective after the instruction execution are referred to, and the value of the time decay factor is set to reflect the decay influence of the distance between the instruction execution time and the current monitoring time on the weight factor.

[0174] The formula is significant in integrating the risk change speed, state stability degree and time decay effect of the comprehensive chemical environment, and accurately quantifying the optimization weight factor of the safety control instruction. In the calculation process, the risk change rate is obtained by the ratio of the risk value change amount and the unit time change amount of the dynamic characteristic, and then multiplied by the risk sensitivity coefficient to highlight the positive influence of risk change on the weight; the parameter change rate is obtained by the ratio of the environmental parameter change amount and the instantaneous time change amount of the real-time environmental change data, and then multiplied by the stability control coefficient to reflect the positive contribution of state stability demand to the weight; the time decay term is obtained by taking the logarithm of the time decay factor plus one, and then multiplied by the timeliness decay coefficient to adjust the weight in the opposite direction and weaken the weight proportion of outdated instructions.

[0175] The weight factor is obtained by integrating the three parts of the calculation results, which not only ensures that the high-risk and strong fluctuation environment state corresponds to a higher instruction optimization weight, but also takes into account the influence of time factors on the effectiveness of instructions, providing a scientific and quantitative basis for the optimization priority allocation of safety control instructions, making the instruction optimization more in line with the real-time state changes of the chemical environment.

[0176] The beneficial effect is that the state change trend of the chemical environment is determined according to the dynamic characteristics in the real-time environmental change data, which can accurately capture the dynamic evolution law of the environmental safety state. Dynamic characteristics include gas concentration rise and fall, temperature and humidity fluctuation, equipment parameter change, etc. By analyzing its matching degree with safety standards and historical laws, it can be determined that the environment is developing in the direction of risk intensification, safety stability or fluctuation adjustment, avoiding misjudgment caused by relying on static data, and providing a trend basis for instruction optimization that fits the actual situation.

[0177] Based on the state change trend, the optimization weight allocation of safety control instructions can make the instruction optimization more targeted. Higher weight is given to monitoring and path adjustment instructions in risk intensification areas to ensure priority execution; non-core instruction weight is appropriately reduced in safety stable areas to avoid resource waste; mobile flexibility instruction weight is increased in fluctuation areas to adapt to dynamic changes, so that the instruction optimization priority is directly linked to environmental demand.

[0178] According to the optimization priority, the preliminary optimization instruction is obtained by constructing the instruction optimization sequence, which can form a logical and clear execution framework. The highest priority instruction is placed at the front end to ensure priority response to high-risk operations; the medium priority instruction is interleaved logically to ensure action coordination; the low priority instruction is placed at the back to avoid occupying key resources, while considering the instruction dependency to ensure smooth guidance of the preliminary optimization instruction for inspection.

[0179] Verifying the data consistency of initial optimization commands yields optimized control commands, ensuring the accuracy of data after command execution. Establish verification standards covering data matching and correlation; collect data through simulations or small-scale tests; compare and analyze consistency with real-time environmental data; correct deviations until the data accurately reflects the environmental state; and ensure optimized control commands precisely adapt to dynamic environmental changes, improving inspection accuracy.

[0180] The risk sensitivity coefficient is determined by combining industry safety standards and historical risk experience. It can accurately reflect the sensitivity of instructions to risk factors, ensure that the calculation of risk items in the weights matches the actual impact, and avoid the weight allocation from deviating from safety requirements.

[0181] The stability control coefficient is set based on environmental stability requirements, referencing equipment operating thresholds and parameter fluctuation ranges. It measures the role of instructions in maintaining environmental stability, ensuring that weight calculations fully reflect the impact of stability requirements. The timeliness decay coefficient, combined with the rate of risk change and the effective data window, reflects the weakening effect of time on weights, preventing outdated instructions from receiving unreasonably high weights.

[0182] The change in risk value is obtained by comparing the numerical differences of core risk indicators, intuitively reflecting risk changes and providing accurate data support for the calculation of risk items in the weighting, ensuring that the impact of risk can be quantified. The change per unit time converts the change in risk value into a rate per unit time, making risk changes in different time periods comparable, avoiding misjudging the speed of risk, and improving the scientific nature of the calculation.

[0183] Environmental parameter changes are obtained by recording the differences in key parameters between adjacent monitoring times, directly reflecting parameter fluctuations and providing a true basis for calculating the environmental stability item in the weighting, ensuring that stability requirements are accurately reflected. Instantaneous time changes are determined based on the sampling frequency of the monitoring equipment, accurately reflecting the temporal rhythm of parameter changes and avoiding weighting calculation errors caused by inaccurate time intervals.

[0184] The time decay factor is set with reference to the monitoring period and the effective duration of the data. After logarithmic processing, the decay effect is more in line with the actual law, taking into account both timeliness and rationality, and ensuring the scientific calculation of weights.

[0185] The overall formula integrates three factors: risk changes, environmental stability, and time decay, enabling precise quantification of instruction optimization weight factors. The synergistic effect of these three factors ensures that the weights not only match the real-time risk and stability requirements of the chemical environment but also consider the impact of time on instruction effectiveness. This provides a scientific basis for prioritizing instruction optimization, ensuring that instruction optimization better aligns with dynamic environmental changes and improves the accuracy and adaptability of robot inspection control.

[0186] The event determination module 106 is configured to control the inspection behavior of the robot according to the optimized control instruction, and obtain the environment safety summary and the abnormal event record of the chemical environment when the robot completes the inspection.

[0187] In the embodiment of the present application, when the event determination module performs the operation of controlling the inspection behavior of the robot according to the optimized control instruction, and obtaining the environment safety summary and the abnormal event record of the chemical environment when the robot completes the inspection, the event determination module is specifically configured to:

[0188] drive the robot to complete the predetermined inspection task by using the optimized control instruction, and obtain the execution trajectory of the robot;

[0189] spatially match the execution trajectory with the real-time environment change data, and obtain the action-environment association combination of the robot;

[0190] evolve the safety state of the action-environment association combination, and obtain the environment safety situation distribution of the chemical environment;

[0191] refine the multi-dimensional features of the environment safety situation distribution, and obtain the environment safety summary and the abnormal event record of the chemical environment.

[0192] The robot gradually develops the predetermined inspection task according to the adaptive inspection path, the movement speed adjustment parameter and the inspection action sequence set in the optimized control instruction. In the inspection process, the robot accurately executes each movement adjustment action and path safety inspection action, strictly follows the task execution sequence specified in the instruction, and sequentially covers the key areas and ordinary areas to be inspected. At the same time, the positioning system of the robot records the spatial position coordinates and movement trajectory of the robot in the chemical environment in real time, and combines the execution time stamp of the inspection action to form the execution trajectory of the robot which records the inspection process of the robot, so as to ensure that the trajectory can clearly restore the inspection path and action execution of the robot.

[0193] The execution trajectory of the robot and the real-time environment change data are matched in time and space based on the unified time reference and spatial coordinate system. In the time dimension, the execution time of each action in the execution trajectory is accurately aligned with the monitoring data at the corresponding time point in the real-time environment change data, so that each inspection action can correspond to the environment state data at the same time; in the spatial dimension, the real-time environment change data is associated to the specific physical position of the robot inspection according to the position coordinates in the execution trajectory, such as binding the gas concentration data of a coordinate point with the monitoring action performed by the robot at the position. Through this two-dimensional matching, the action of the robot is one-to-one corresponding to the environment data at the corresponding time and space, forming the action-environment association combination of the robot.

[0194] Based on the action-environment association combination, combined with the safety evolution law of the chemical environment, the change of the environment safety state after the execution of different inspection actions is analyzed. The change trend of the environment parameters, the evolution path of the risk factors, and the influence of the inspection action on the environment safety state in the association combination are combed, such as whether the diffusion or mitigation of the risk factors is found after the robot executes the intensive monitoring action in a certain area. According to the time sequence and spatial distribution, the discrete association combination is connected in series to restore the complete evolution process of the chemical environment safety state from the beginning to the end of the inspection, and the environment safety situation distribution of the chemical environment which comprehensively presents the safety state of each area of the chemical environment is formed.

[0195] The environment safety situation distribution is multi-dimensionally characterized to focus on the core information of the safety state. The overall safety level of each area, the distribution range and intensity of the risk factors are extracted from the spatial dimension; the change law of the safety state, the appearance and disappearance time of the risk factors are combed from the time dimension; different types of safety information such as environment parameter abnormalities and equipment state abnormalities are classified and summarized from the risk type dimension, forming a concise and clear environment safety abstract of the chemical environment with core highlights. At the same time, the abnormal data exceeding the safety threshold in the situation distribution is screened out, and the key information such as the location, time, abnormal type, data change amplitude of the abnormal occurrence is recorded in detail, and the abnormal event record of the chemical environment is arranged according to the chronological order of the abnormal occurrence, to ensure that the abstract and record can comprehensively and accurately reflect the safety state of the chemical environment.

[0196] The beneficial effect is that the execution trajectory of the robot is obtained by driving the robot to complete the predetermined inspection task with the optimized control instruction, which can completely restore the dynamic path and action details of the robot inspection. The optimized control instruction clearly defines the inspection path, speed parameter and action sequence, and the positioning system will record the spatial position and action timestamp in real time during the execution of the robot, forming a trajectory associated with the movement path and inspection action, which provides a comprehensive basis for subsequent space-time matching and avoids the inability to trace the inspection process.

[0197] The action-environment association combination is obtained by matching the execution trajectory with the real-time environment change data, which can establish a precise correspondence between the inspection action and the environment state. The action execution time and the environment data of the same period are aligned in time, and the environment data and the robot inspection position are bound in space, avoiding the disconnection of the action and the environment data due to space-time dislocation, so that each inspection action can correspond to clear environment feedback.

[0198] The environment safety situation distribution of the chemical environment is obtained by safety state evolution of the action-environment association combination, which can present the dynamic change of the environment safety state. Combined with the safety evolution law, the influence of the action on the environment parameters and risk factors is analyzed, the association combination is connected in series according to the time and space sequence to restore the safety state evolution of the whole inspection process, avoiding the inability to grasp the overall change law caused by isolated analysis of the data.

[0199] By extracting multi-dimensional features from the distribution of environmental safety status, an environmental safety summary and abnormal event records for the chemical environment can be obtained, enabling efficient extraction of core safety information. Information is refined from spatiotemporal and risk type dimensions to form a concise safety summary, and details of abnormal data records are filtered out, avoiding information overload from obscuring key safety points and providing clear and practical results for safety management.

[0200] Reference Figure 2 The diagram shown is a flowchart illustrating a sensor fusion-based control method for a chemical safety inspection robot according to an embodiment of the present invention. In this embodiment, the sensor fusion-based control method for a chemical safety inspection robot includes:

[0201] S1. The raw data from multi-source heterogeneous data in the chemical environment are fused into standardized data, and the structured features in the standardized data are analyzed to identify the risk factors in the chemical environment;

[0202] S2. Using the risk factors as the construction benchmark, the safety status of the chemical environment is reconstructed to obtain a safety situation map of the chemical environment, and the safety situation map is mapped to the robot to determine the control strategy of the robot.

[0203] S3. When the robot starts its inspection, the control strategy is parsed to control the robot's movement, path adjustment, and path safety inspection actions.

[0204] S4. Encode the movement adjustment path action and the path safety inspection action into safety control instructions to control the robot to monitor real-time environmental change data of the chemical environment;

[0205] S5. Based on the real-time environmental change data, the safety control command is optimized at multiple levels to obtain the optimized control command for the robot;

[0206] S6. According to the optimized control instructions, control the inspection behavior of the robot. When the robot completes the inspection, obtain the environmental safety summary and abnormal event record of the chemical environment.

[0207] 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.

[0208] 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.

[0209] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application but not to limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the spirit and scope of the present application.

Claims

1. A chemical safety inspection robot control system based on sensor fusion, characterized in that, The system includes a data fusion module, a strategy construction module, a strategy parsing module, a real-time monitoring module, an optimization instruction module, and an event determination module, wherein: The data fusion module is used to fuse raw data from multiple heterogeneous sources in the chemical environment into standardized data, and to analyze the structured features in the standardized data to identify risk factors in the chemical environment. The strategy construction module is used to reconstruct the safety status of the chemical environment based on the risk factors, obtain a safety situation map of the chemical environment, and map the safety situation map to the robot to determine the control strategy of the robot. Specifically, it is used for: Using the high-risk factors among the risk factors as the upper layer of the map, the medium-risk factors among the risk factors as the middle layer of the map, and the low-risk factors among the risk factors as the bottom layer of the map, a risk level distribution map of the chemical environment is constructed. Spatial smoothing interpolation is performed on the risk level distribution map to obtain a continuous risk surface of the risk level distribution map; The environmental safety boundaries in the continuous risk surface are extracted to obtain a safety zoning map of the chemical environment; The safety zoning map is superimposed on the real-time environment of the chemical environment to obtain a safety situation map of the chemical environment. When mapping the safety situation map to the robot and determining the robot's control strategy, it is specifically used for: The security zone information in the security situation map is analyzed to determine the key areas and ordinary areas to be inspected by the robot; The key areas to be inspected are designated as the priority inspection paths for the robot, while the ordinary areas are designated as the secondary inspection paths for the robot. The priority inspection path and the secondary inspection path are integrated into the robot's adaptive inspection path; Based on the adaptive inspection path, determine the robot's movement speed adjustment parameters and inspection action sequence; The control strategy of the robot is obtained by reconstructing the movement speed adjustment parameters and the inspection action sequence. The strategy parsing module is used to parse the control strategy when the robot starts its inspection, so as to control the robot's movement, path adjustment and path safety inspection actions. The real-time monitoring module is used to encode the movement path adjustment action and the path safety inspection action into safety control commands to control the robot to monitor real-time environmental change data of the chemical environment, specifically for: The basic instruction set of the robot is obtained by mapping the basic motion primitives in the movement adjustment path action and the monitoring enhancement primitives in the path safety inspection action. Using the environmental risk characteristics of the chemical environment as a logical benchmark, the logical inconsistencies in the basic instruction set are eliminated to obtain the composite control instructions for the robot. Verify the effectiveness of instruction coordination in the composite control command to obtain the robot's safety control command; According to the safety control instructions, the robot is controlled to monitor real-time environmental change data of the chemical environment; The optimization instruction module is used to perform multi-level optimization of the safety control instructions based on the real-time environmental change data to obtain the optimized control instructions for the robot, specifically for: Based on the dynamic characteristics in the real-time environmental change data, the trend of state change of the chemical environment is determined; Based on the state change trend, the safety control instructions are optimized by weight allocation to obtain the optimized priority of the safety control instructions; Based on the optimization priority, an instruction optimization sequence for the robot is constructed to obtain the initial optimization instructions for the robot; Verify the data consistency of the preliminary optimization instructions to obtain the optimized control instructions for the robot; The event determination module is used to control the robot's inspection behavior according to the optimized control instructions. When the robot completes the inspection, it obtains an environmental safety summary and abnormal event record of the chemical environment. Specifically, it is used for: The robot is driven by the optimized control commands to complete the predetermined inspection task, and the execution trajectory of the robot is obtained; The execution trajectory is spatiotemporally matched with the real-time environmental change data to obtain the robot's action-environment association combination. The safety state evolution of the action-environment association combination is performed to obtain the environmental safety status distribution of the chemical environment; Multidimensional feature extraction is performed on the environmental safety situation distribution to obtain an environmental safety summary and abnormal event record of the chemical environment.

2. The chemical safety inspection robot control system based on sensor fusion as described in claim 1, characterized in that, The data fusion module, when performing the fusion of raw data from multiple heterogeneous sources in a chemical environment into standardized data, and parsing the structured features in the standardized data to identify risk factors in the chemical environment, is specifically used for: The environmental parameter data, equipment status data and spatial perception data collected by sensors in the chemical environment are integrated into multi-source heterogeneous data of the chemical environment. The multi-source heterogeneous data is aligned in both spatiotemporal dimensions to obtain the spatiotemporally aligned data of the chemical environment. The spatiotemporal aligned data is dimensionally unified to obtain preliminary standardized data of the chemical environment; The data formats of different scales in the preliminary standardized data are unified to obtain the standardized data of the chemical environment; Tensor synthesis is performed on the environmental parameter features, equipment status features, and spatial distribution features in the standardized data to obtain a multi-dimensional feature vector of the standardized data. The multi-dimensional feature vector is used as a risk assessment benchmark to identify risk factors in the chemical environment.

3. The chemical safety inspection robot control system based on sensor fusion as described in claim 1, characterized in that, When the robot begins its inspection, the strategy parsing module parses the control strategy to control the robot's movement, path adjustment, and path safety inspection actions. Specifically, this is used for: When the robot begins its inspection, the environmental safety elements in the control strategy are analyzed to obtain the robot's inspection task sequence. The task execution order of the robot is determined based on the inspection task sequence. By connecting the key inspection locations and transition path points in the task execution sequence, the path node sequence of the robot is obtained; The path node sequence is input into the robot to synchronize and coordinate the robot's movement, path adjustment, and path safety inspection actions.

4. The chemical safety inspection robot control system based on sensor fusion as described in claim 1, characterized in that, The formula for calculating the weight factor in the optimized weight allocation is as follows: ； formula, The weighting factor is... The risk sensitivity coefficient for the stated state change trend. This is the stability control coefficient for the trend of the state change. The time-dependent attenuation coefficient of the state change trend is given. The change in the risk value of the dynamic feature. The change per unit time corresponding to the dynamic feature. The change in environmental parameters in the real-time environmental change data. This refers to the instantaneous time change corresponding to the real-time environmental change data. The time decay factor, It is a logarithmic function.

5. A control method for a chemical safety inspection robot based on sensor fusion, characterized in that, The method is for using the sensor fusion-based chemical safety inspection robot control system as described in claim 1. S1. The raw data from multi-source heterogeneous data in the chemical environment are fused into standardized data, and the structured features in the standardized data are analyzed to identify the risk factors in the chemical environment; S2. Using the risk factors as the construction benchmark, the safety status of the chemical environment is reconstructed to obtain a safety situation map of the chemical environment, and the safety situation map is mapped to the robot to determine the control strategy of the robot. S3. When the robot starts its inspection, the control strategy is parsed to control the robot's movement, path adjustment, and path safety inspection actions. S4. Encode the movement adjustment path action and the path safety inspection action into safety control instructions to control the robot to monitor real-time environmental change data of the chemical environment; S5. Based on the real-time environmental change data, the safety control command is optimized in multiple levels to obtain the optimized control command for the robot; S6. According to the optimized control instructions, control the inspection behavior of the robot. When the robot completes the inspection, obtain the environmental safety summary and abnormal event record of the chemical environment.

Images