A robot rescue emergency control system based on voice interaction

Abstract

The present invention relates to the field of robot control technology, and in particular to a robot rescue emergency control system based on voice interaction. The system is used to: utilize the robot's built-in voice receiving function to collect control voice signals and perform structured analysis of rescue emergency voice control tasks, and generate rescue emergency voice control task structured data; obtain the panoramic image data, environmental perception data, and long-range laser radar data monitored by the robot; perform global image point cloud model construction and collaborative optimization processing of multi-level environmental changes based on the panoramic image data, environmental perception data, and long-range laser radar data monitored by the robot, and generate an optimized global image perception model; perform rescue emergency intelligent control operations on the robot based on the rescue emergency voice control task structured data and the optimized global image perception model. The present invention receives voice commands through the voice interaction function to achieve automated and efficient robot rescue emergency control.

Description

Technical Field

[0001] The present invention relates to the technical field of robot control, and in particular to a robot rescue emergency control system based on voice interaction. Background Art

[0002] Emergency rescue robot systems usually rely on fixed control methods, such as manual operation or preset path planning, which are difficult to meet the needs of complex and changeable rescue environments. With the rapid development of artificial intelligence technology, the application of robots in various scenarios has gradually become popular. In actual rescue emergency scenarios, fast and accurate command transmission and execution are the key to ensuring the success of the mission, so that robots with automated rescue emergency control have the characteristics of rapid response, efficient execution and reduced risk of casualties, which has attracted widespread attention. However, existing robot rescue emergency control systems often find it difficult to meet the needs of real-time and high interactivity. There are limitations in the image processing and sensor technology of rescue emergency robots in environmental modeling, such as insufficient response to multi-level environmental changes or inaccurate data processing, and lack of high-precision perception and intelligent adaptation capabilities of the environment, which limits their efficiency in practical applications. Summary of the invention

[0003] Based on this, the present invention provides a robot rescue emergency control system based on voice interaction to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a robot rescue emergency control system based on voice interaction includes the following modules:

[0005] The voice interaction analysis module is used to collect control voice signals using the robot's built-in voice receiving function to generate control voice signals; perform structured analysis of rescue emergency voice control tasks based on control voice signals to generate rescue emergency voice control task structured data;

[0006] The robot image perception modeling module is used to obtain the panoramic image data, environmental perception data and long-range laser radar data monitored by the robot; perform global image point cloud modeling processing based on the panoramic image data and long-range laser radar data to generate a global image point cloud model; perform three-dimensional model collaborative optimization processing of multi-level environmental changes on the global image point cloud model based on the environmental perception data to generate an optimized global image perception model;

[0007] The robot rescue emergency intelligent control module is used to perform rescue emergency intelligent control operations on the robot based on the rescue emergency voice control task structured data and the optimized global image perception model;

[0008] The robot rescue emergency feedback module is used to collect the immediate rescue emergency task status feedback data for the rescue emergency intelligent control operation to obtain the immediate rescue emergency task status feedback data; transmit the immediate rescue emergency task status feedback data to the terminal to perform the immediate rescue emergency task status feedback operation;

[0009] The real-time running posture analysis module is used to use the robot's built-in short-range laser radar equipment to perform short-range laser radar real-time monitoring and processing on the robot performing the rescue emergency intelligent control operation, and generate real-time short-range laser radar data; perform real-time rescue driving path state analysis based on the real-time short-range laser radar data, and generate real-time rescue driving path state data; perform real-time robot adaptive driving posture analysis based on the real-time rescue driving path state data, and generate real-time robot adaptive driving posture data;

[0010] The real-time control state intelligent optimization module is used to perform real-time robot adaptive driving rate analysis based on the real-time rescue driving path state data and the real-time robot adaptive driving posture data to generate real-time robot adaptive driving rate data; based on the real-time robot adaptive driving posture data and the real-time robot adaptive driving rate data, the real-time rescue emergency control state intelligent optimization operation is performed on the robot performing the rescue emergency intelligent control operation.

[0011] Furthermore, the voice interaction analysis module includes the following functions:

[0012] Use the robot's built-in voice receiving function to collect control voice signals and generate control voice signals;

[0013] Use automatic speech recognition technology to perform rescue emergency control demand text analysis on the control demand voice signal to generate rescue emergency control demand text data;

[0014] Perform semantic parsing on the rescue emergency control demand text data to generate rescue emergency control demand semantic data;

[0015] According to the rescue emergency control demand semantic data, a structured analysis of the rescue emergency voice control task is performed to generate the rescue emergency voice control task structured data.

[0016] Furthermore, the robot image perception modeling module includes the following functions:

[0017] Use the built-in panoramic image monitoring device of the robot to perform panoramic image monitoring processing and generate panoramic image data;

[0018] Use the robot's built-in environmental perception sensors to collect environmental perception data and generate environmental perception data;

[0019] Use the robot's built-in long-range laser radar device to collect long-range laser radar data and generate long-range laser radar data;

[0020] Perform long-range laser point cloud data analysis based on long-range laser radar data to generate long-range laser point cloud data;

[0021] Transmitting long-distance laser point cloud data to panoramic image data for global image point cloud modeling and processing to generate a global image point cloud model;

[0022] Performing hierarchical distance division processing on the environmental perception data to generate hierarchical environmental perception data;

[0023] Performing hierarchical environment perception time series feature analysis on the hierarchical environment perception data to generate hierarchical environment perception time series feature data;

[0024] Perform hierarchical environment perception trend analysis based on hierarchical environment perception time series feature data to generate hierarchical environment perception trend data;

[0025] The environmental perception data is transmitted to the global image point cloud model for environmental perception data mapping processing to generate a global image perception model; and the global image perception model is collaboratively optimized with a three-dimensional model of multi-level environmental changes through hierarchical environmental perception trend data to generate an optimized global image perception model.

[0026] Furthermore, the transmitting of the long-distance laser point cloud data to the panoramic image data for global image point cloud modeling processing includes:

[0027] Perform image matching node analysis based on panoramic image data to generate image matching node data;

[0028] Perform laser point cloud image matching data analysis based on image matching node data and long-distance laser point cloud data to generate laser point cloud image matching data;

[0029] Based on the laser point cloud image matching data, the long-distance laser point cloud data is transmitted to the panoramic image data for global image modeling processing to generate a global image point cloud model.

[0030] Furthermore, performing image matching node analysis based on the panoramic image data includes:

[0031] Performing global pixel blur interference evaluation processing on the panoramic image data to generate global pixel blur interference data;

[0032] Perform global image effective area analysis based on global pixel fuzzy interference data to generate global image effective area data;

[0033] Performing global image effective area gradient feature analysis on global image effective area data to generate global image effective area gradient feature data;

[0034] Image matching node analysis is performed based on the gradient feature data of the effective area of ​​the global image to generate image matching node data.

[0035] Furthermore, the robot rescue emergency intelligent control module includes the following functions:

[0036] Based on the structured data of rescue emergency voice control tasks and the optimized global image perception model, the rescue emergency automated control decision analysis is performed to generate rescue emergency automated control decisions;

[0037] Based on the rescue emergency automated control decision, the robot performs rescue emergency intelligent control operations.

[0038] Furthermore, the real-time running posture analysis module includes the following functions:

[0039] Use the robot's built-in short-range laser radar equipment to conduct real-time short-range laser radar monitoring and processing on the robot performing rescue emergency intelligent control operations to generate real-time short-range laser radar data;

[0040] Performing an immediate rescue driving path status analysis based on the immediate short-range laser radar data to generate immediate rescue driving path status data;

[0041] Obtain the physical design parameters of the rescue robot;

[0042] Analyze the controllable parameters of robot motion according to the physical design parameters of the rescue robot, and generate the controllable parameters of robot motion;

[0043] Analyze the robot simulation operation posture parameters based on the robot motion controllable parameters, and generate the robot simulation operation posture parameters;

[0044] Conducting robot simulation operation posture limit state analysis on robot simulation operation posture parameters to generate robot simulation operation posture limit state data;

[0045] Performing feasibility analysis on the state data of the immediate rescue driving path according to the extreme state data of the robot simulation operation posture, so as to obtain the state data of the feasible rescue driving path;

[0046] According to the feasible rescue driving path state data and the robot simulation operation posture parameters, the real-time robot adaptive driving posture analysis is performed to generate the real-time robot adaptive driving posture data.

[0047] Furthermore, the feasibility analysis of the immediate rescue path status data includes:

[0048] When the immediate rescue driving path status data conforms to the robot simulation operation posture limit state data, the immediate rescue driving path status data is marked as feasible rescue driving path status data; otherwise, the immediate rescue driving path status data is marked as abnormal rescue driving path status data, and the abnormal rescue driving path status data is fed back to the rescue emergency intelligent control operation to execute the rescue emergency driving path intelligent adjustment operation.

[0049] Furthermore, the robot simulation operation posture parameters include the robot simulation operation pitch angle, the robot simulation operation roll angle and the robot simulation operation heading angle.

[0050] Furthermore, the real-time control state intelligent optimization module includes the following functions:

[0051] Design the robot driving rate cost function based on the preset robot driving constraint decision and the robot simulation operation posture parameters, and generate the robot driving rate cost function;

[0052] The instant rescue driving path state data and the instant robot adaptive driving posture data are transmitted to the robot driving speed cost function to perform the instant robot adaptive driving speed analysis, and generate the instant robot adaptive driving speed data;

[0053] Based on the real-time robot adaptive driving posture data and the real-time robot adaptive driving speed data, the robot performing the rescue emergency intelligent control operation is intelligently optimized in the real-time rescue emergency control state.

[0054] The beneficial effect of the present application is that the voice interaction analysis module realizes accurate analysis of the rescue emergency control needs by introducing automatic speech recognition (ASR) technology and semantic analysis technology. Through the steps of voice signal acquisition, text analysis and semantic analysis, the system can quickly convert natural language into structured data, thereby improving the response efficiency of rescue emergency control tasks. Voice interaction greatly simplifies the human-computer interaction process. Especially in emergency situations, operators can directly issue instructions to the robot in natural language, shortening the response time. In addition, the module has the ability to perform structured analysis on complex voice instructions, so that the robot can understand multi-level task requirements, thereby improving the accuracy and flexibility of task execution. The robot image perception modeling module realizes multi-dimensional perception and high-precision modeling of the rescue environment by combining panoramic image monitoring equipment, environmental perception sensors and long-range laser radar equipment. The combination of global image point cloud modeling processing and three-dimensional model collaborative optimization enables the robot to generate an optimized global image perception model, thereby fully grasping the dynamic change characteristics of complex environments. This multi-level optimization processing not only improves the ability to adapt to the environment, but also enhances the path planning and obstacle avoidance performance of the robot in complex terrain. In addition, the module can identify and predict potential environmental change trends by analyzing the hierarchical temporal characteristics of environmental perception data, providing the robot with more intelligent action decision support, thereby ensuring the efficient execution and safety of rescue tasks. The robot rescue emergency intelligent control module realizes highly automated rescue emergency control decisions by comprehensively analyzing the rescue emergency voice control task structured data and optimizing the global image perception model. The module has autonomous decision-making capabilities, can quickly generate rescue emergency automation control plans that meet actual needs, and guide the robot to perform complex tasks. This intelligent control method reduces human intervention and significantly improves the robot's emergency response speed in emergencies. In addition, the module supports real-time adjustment of control strategies in dynamic environments to ensure the accuracy and stability of robots when performing tasks in complex and changing environments. This feature greatly improves rescue efficiency, especially in disaster relief or emergency medical situations, which is of great significance to the protection of life safety. The robot rescue emergency feedback module builds an efficient information closed loop by real-time collection and feedback of task execution status. The module can obtain feedback data on the status of rescue emergency tasks in real time and transmit it to the terminal for further analysis or display. This instant feedback mechanism helps operators or control systems to understand the progress of task execution in real time and optimize subsequent control strategies based on feedback information. Through continuous monitoring and feedback of task status, the module effectively reduces control errors caused by information lag and improves the success rate of rescue missions. At the same time, the module also provides important data support for post-analysis and improvement of rescue missions, further improving the overall reliability and intelligence level of the robot emergency control system.The real-time running posture analysis module realizes real-time dynamic analysis of the robot's driving path state and posture by combining real-time short-range laser radar data and robot motion parameters. It can make feasibility judgments on the rescue path state, and can also quickly generate adaptive posture data according to environmental changes to ensure the robot's travel stability in complex terrain. This real-time analysis capability significantly improves the flexibility and safety of the robot in disaster rescue scenarios, especially in narrow passages, large slopes or environments with many obstacles. The module also improves the adaptability of the robot under special conditions through comprehensive evaluation of simulated running posture parameters, providing a strong guarantee for the efficient and stable operation of the robot in tasks. The real-time control state intelligent optimization module adjusts the robot's driving rate and control strategy in real time through multi-dimensional analysis based on path state and posture data. The rate cost function designed by the module can balance the relationship between the robot's travel speed and task requirements, while ensuring task efficiency and reducing the complexity of operating energy consumption and posture adjustment. Through intelligent optimization of real-time status, the robot can quickly adapt to environmental changes or emergencies, reducing the risk of task interruption or failure caused by unstable status. The optimization function of this module significantly enhances the robot's control accuracy and dynamic response capabilities in emergency rescue missions, providing reliable technical support for the efficient execution of complex tasks.

[0055] Therefore, the robot rescue emergency control system based on voice interaction of the present invention combines voice interaction technology with advanced image perception modeling technology. The voice interaction technology can automatically execute the target rescue mission through its voice instructions. By analyzing the response characteristics of multi-level environmental changes, it can ensure the real-time and high interactivity requirements of the image processing and sensor technology of the rescue emergency robot in environmental modeling, so that the robot has the ability to provide real-time feedback and automatically execute rescue missions. In addition, the robot's motion posture and path planning in the rescue mission can achieve adaptive dynamic adjustment, meet the high-precision perception and intelligent adaptability of the environment, avoid mission failure due to path planning errors or environmental interference, and improve the actual application efficiency of rescue missions. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 A schematic diagram of a module flow of a robot rescue emergency control system based on voice interaction according to the present invention;

[0057] Figure 2 for Figure 1 Detailed implementation functional flow chart of the robot image perception modeling module;

[0058] Figure 3 for Figure 1 Detailed implementation function flow diagram of the real-time running posture analysis module;

[0059] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0060] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0061] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0062] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0063] To achieve this, please refer to Figures 1 to 3 The present invention provides a robot rescue emergency control system based on voice interaction. In the embodiments of the present invention, please refer to Figure 1 FIG. 1 is a schematic diagram of a module flow of a robot rescue emergency control system based on voice interaction according to the present invention. The robot rescue emergency control system based on voice interaction includes the following modules:

[0064] S1: Voice interaction analysis module, used to collect control voice signals using the robot's built-in voice receiving function to generate control voice signals; perform structured analysis of rescue emergency voice control tasks based on control voice signals to generate rescue emergency voice control task structured data;

[0065] In the embodiment of the present invention, the control voice signal is collected by the microphone array built into the robot. The microphone array can realize 360-degree sound perception, automatically capture the sound in the surrounding environment, and perform preliminary signal preprocessing, such as removing background noise, echo suppression, etc. The collected control voice signal is transmitted to the signal processing unit to generate the original digital voice signal. At this time, the system can distinguish the sound signals from different directions, and ensure that the quality of the signal is not affected by environmental factors, and ensure that the collected control voice signal has high clarity and accuracy. Once the control voice signal is successfully collected, it will enter the automatic speech recognition (ASR) processing stage. The robot uses efficient automatic speech recognition technology to convert the voice signal into text data. The control demand text data is semantically parsed, and the system will further analyze the recognized text content. Semantic parsing also includes correlation analysis of the context in the text to ensure that the user's real needs are understood, and generate structured semantic data to indicate the task objectives, task types, specific operation instructions, etc. This process can accurately identify the implicit task requirements in the text by relying on a pre-trained natural language model. After completing the semantic parsing, the rescue emergency control demand semantic data enters the structured analysis stage. At this stage, the system first classifies, filters and sorts the semantic data, and generates task priorities based on the urgency of the task, the order of operations, etc. These structured task data are converted into robot control instruction sets, including action execution instructions, task target locations, task types, etc., to form structured data that can be directly transmitted to the robot control system. This process not only ensures the accurate understanding of the user's intentions, but also provides clear task guidance for the robot's subsequent intelligent decision-making and behavior execution.

[0066] S2: Robot image perception modeling module, used to obtain panoramic image data, environmental perception data and long-range laser radar data monitored by the robot; perform global image point cloud modeling processing based on the panoramic image data and long-range laser radar data to generate a global image point cloud model; perform three-dimensional model collaborative optimization processing of multi-level environmental changes on the global image point cloud model based on environmental perception data to generate an optimized global image perception model;

[0067] In an embodiment of the present invention, panoramic image monitoring processing is performed by a panoramic image monitoring device built into the robot. The panoramic image monitoring device includes multiple high-resolution cameras arranged at a fixed angle to cover a 360-degree all-round view around the robot. By synchronously collecting image data from these cameras, the robot can obtain a panoramic image. While collecting panoramic image data, the robot also uses built-in environmental perception sensors to collect environmental perception data. Environmental perception sensors generally include temperature and humidity sensors, gas sensors, acoustic wave sensors, etc., which can monitor the temperature, humidity, gas concentration and other information of the surrounding environment in real time. In order to further enhance the robot's ability to perceive the environment, a long-range laser radar device (LiDAR) built into the robot is used to collect long-range laser radar data. By emitting a laser beam and receiving the reflected signal, the laser radar can accurately measure the distance to the surrounding objects and generate three-dimensional point cloud data of the environment. The built-in laser radar of the robot can cover a certain range of environments, obtain multi-dimensional spatial information by scanning at different angles, and generate long-range laser radar data. After the long-range laser radar data is collected, the system will analyze the laser point cloud data. According to the distance information in the laser point cloud data, the robot uses the point cloud processing algorithm to perform data denoising, point cloud reconstruction, and sparse point supplementation. The processed long-range laser point cloud data can more accurately describe obstacles, buildings, and terrain features in the environment. After analysis, the long-range laser point cloud data will be combined with the panoramic image data to perform global image point cloud modeling. First, the system will align and match the common viewpoints in the panoramic image data and the laser point cloud data to ensure the spatial consistency of the image and point cloud data. The system uses the image matching algorithm to accurately match the feature points in the image with the corresponding points in the point cloud to form a set of corresponding relationships between the image and the point cloud data. Next, based on the matching results, the system generates a comprehensive global image point cloud model through a three-dimensional modeling algorithm. This model not only contains the image data of the surrounding environment, but also integrates accurate three-dimensional spatial information. With the establishment of the global image point cloud model, the system continues to perform hierarchical distance division processing on the environmental perception data. The system performs layered processing on the environment based on the environmental data collected by the environmental perception sensor, such as temperature, humidity, gas concentration, and other information. The environmental data of each layer is divided according to its impact range on the robot, and multiple levels of data sets are formed. The data of each level includes environmental parameters in a specific area, such as temperature range, gas concentration, etc. This processing process helps the system to judge environmental changes at each level and make more accurate responses and decisions. The environmental data is sorted in time series and classified into different time periods (such as hourly, minute-by-minute, or second-by-second data) according to the timestamp of the sensor data collection. Next, time series analysis methods (such as sliding windows or Fourier transforms) are used to extract time series features from these data.The generated hierarchical environmental perception time series feature data includes the change amplitude, frequency and other relevant statistical characteristics of variables such as temperature and gas concentration in each time period, which serves as the basic data for subsequent trend analysis. Analyze the change trend of environmental data in different time periods. By comparing the difference between the data of the current time period and the historical data, regression analysis, exponential smoothing or other trend prediction methods are used to identify the change pattern of environmental data. In this way, the system can analyze the change trend of the environment in real time and generate hierarchical environmental perception trend data. These trend data not only reflect the speed and amplitude of environmental changes, but also predict the change trajectory of the future environment, providing a basis for risk prediction and decision-making in robot rescue missions. In the process of mapping environmental perception data to the global image point cloud model, the data from the environmental perception sensor must first be converted into a format compatible with the panoramic image point cloud model. Specifically, the image data obtained by the robot through panoramic images and the three-dimensional point cloud data collected by the lidar need to be converted into a unified spatial coordinate system. The system associates each level of environmental perception data (such as temperature, humidity, gas concentration, etc.) with the corresponding position in the image point cloud data. For example, if the gas sensor detects that the gas concentration in a certain area exceeds the standard, the system will map the gas concentration data to the point cloud coordinates corresponding to the area. Through this data mapping, the system can combine real-time environmental perception information with the three-dimensional environmental model to form a real-time updated global image perception model. This model not only contains the image data and physical space information of the environment, but also reflects the immediate state of the surrounding environment, such as high temperature areas or harmful gas leakage areas. After completing the environmental perception data mapping, the system then performs a multi-level environmental change three-dimensional model collaborative optimization process of the global image perception model. The specific implementation steps are as follows: First, based on the hierarchical environmental perception trend data, the system analyzes the change trend of the current environment. For example, if the gas concentration continues to rise in a certain area and the temperature gradually rises, the system will determine that there may be a fire risk in the area and give priority to the panoramic image point cloud data of the area. Through the weighted average method, the system corrects the point cloud data of the area, adjusts the density, accuracy and reliability of the point cloud, and thus optimizes the global image perception model of the area.

[0068] S3: Robot rescue emergency intelligent control module, used to perform rescue emergency intelligent control operations on the robot based on rescue emergency voice control task structured data and optimized global image perception model;

[0069] In an embodiment of the present invention, the system analyzes the control requirements contained in the task based on the rescue emergency voice control task structured data obtained from the voice interaction module. The content in the task structured data is converted into a set of standardized instruction sets that can express the actions that the robot needs to complete, such as moving, picking, switching, etc. On this basis, combined with the environmental perception data provided in the optimized global image perception model, the system evaluates the current environmental state, such as the obstacle distribution, spatial layout, channel smoothness and other information obtained through the panoramic image perception model and laser point cloud data. According to the needs of the rescue task, the system performs environmental adaptability analysis to determine which areas are accessible, which areas need to be avoided, and the optimal path planning. Combined with the above information, the system enters the automated control decision stage. At this stage, the system performs decision analysis through an algorithm model based on information such as task requirements, the current state of the environment, and the robot's own motion capabilities, and generates the most appropriate automated control decision. After receiving the rescue emergency automated control decision, the robot begins to perform the corresponding intelligent control operation. The system commands the robot to move along the specified path based on the previously generated automated control decision. This path will be updated in real time through panoramic images and lidar data to ensure that the robot avoids obstacles and maintains the effectiveness of the path during execution.

[0070] S4: A robot rescue emergency feedback module, which is used to collect the immediate rescue emergency task status feedback data for the rescue emergency intelligent control operation to obtain the immediate rescue emergency task status feedback data; transmit the immediate rescue emergency task status feedback data to the terminal to perform the immediate rescue emergency task status feedback operation;

[0071] In an embodiment of the present invention, during the period when the robot performs a rescue emergency task, the robot rescue emergency feedback module collects the status data of the robot's current task in real time. Through a variety of built-in sensors (such as visual sensors, distance sensors, accelerometers, temperature sensors, etc.), the robot's various operating states, such as current position, task progress, power state, environmental changes, etc., are obtained. For example, when performing the task "moving to the disaster site", the robot collects data simultaneously through a panoramic image monitoring device and a laser radar device, and feeds back the image of the current environment and the distance information of obstacles in real time. Assuming that the robot encounters an obstacle during the execution process, the sensor will immediately detect this change and transmit feedback data back. At the same time, the robot's position, speed, tilt angle (lateral tilt angle and longitudinal tilt angle) and other data will also be collected and fed back in real time. These feedback data include the robot's real-time position, state and task progress (such as whether it has reached the destination, whether it has successfully avoided obstacles, etc.). In addition, when the robot performs a rescue task, the feedback module will also collect data on the robot's current working state, such as whether operations such as grabbing and removing objects have been performed. Each operation step will be recorded and converted into data, and the feedback module will continuously track the progress of the task and generate task status data. The acquisition process of these data is automated and does not rely on human intervention, ensuring the efficiency and accuracy of the rescue process. Once the robot feedback module collects the status feedback data of the immediate rescue emergency task, the data will be transmitted to the rescue task terminal through the wireless communication module. Specifically, the data transmission method is to transmit the status feedback information to the remote control terminal through a wireless network (such as Wi-Fi, 4G / 5G). After receiving the data, the terminal performs corresponding operations according to the task requirements and feedback content. For example, if the robot fails to successfully avoid obstacles during the execution of the task, or the task is delayed, the terminal control system will receive these feedback information. The terminal system processes the received feedback data in real time, generates new control instructions or adjusts the task plan. For example, during transmission, if the system detects that the robot task progress is delayed, the terminal will determine whether it is necessary to adjust the robot's path planning or task execution order based on the feedback data. This process ensures the instant feedback of the task status, so that the task controller can clearly understand the execution status of the robot and make corresponding adjustments. The feedback data includes all important status information during the task execution process, such as progress, success rate, task success or failure flag, etc., which ensures the timely transmission and accurate analysis of information in the rescue mission. After the robot completes a certain task, the terminal uses the immediate rescue emergency task status feedback data to make real-time adjustments to the robot's task execution. When the terminal receives the feedback data, the system will decide whether to continue, suspend, or adjust the task content based on the execution progress and status of the current task. For example, suppose the robot is performing the "rescue the wounded" task, and the feedback data shows that the robot has successfully reached the destination, but due to the complex surrounding environment, the robot has failed to successfully open the rescue channel.At this time, after receiving the data, the terminal will display the interruption status of the task, and the control system will instruct the robot to take other measures, such as reselecting the path or calling the robotic arm to perform other operations. The terminal adjusts the robot's task execution based on the immediate task status feedback data to ensure the smooth completion of the rescue operation. The execution of feedback operations not only includes real-time adjustment of tasks, but also involves the status recording and alarm of the robot. For example, if the robot encounters a system failure or environmental abnormality (such as low battery power) during the task, the feedback module will automatically trigger the alarm mechanism and transmit the information to the terminal. The terminal will adjust the task progress or trigger the emergency plan in time according to the alarm data to ensure the safe progress of the task.

[0072] S5: Real-time operation posture analysis module, used to use the robot's built-in short-range laser radar equipment to perform short-range laser radar real-time monitoring and processing on the robot performing the rescue emergency intelligent control operation, and generate real-time short-range laser radar data; perform real-time rescue driving path state analysis based on the real-time short-range laser radar data, and generate real-time rescue driving path state data; perform real-time robot adaptive driving posture analysis based on the real-time rescue driving path state data, and generate real-time robot adaptive driving posture data;

[0073] In the embodiment of the present invention, the short-range laser radar device built into the robot performs a real-time scan of the surrounding environment through a laser beam during the rescue emergency intelligent control operation, obtains distance data and generates real-time short-range laser radar data. The device calculates the distance by emitting lasers and measuring the time of the reflected laser pulses, and accurately detects the relative position and distance between the robot and the surrounding obstacles. According to the real-time short-range laser radar data, the system performs a real-time rescue driving path state analysis on the path taken by the robot when performing the rescue mission. By analyzing the obstacle positions, distances between obstacles and their impact on the robot provided by the laser radar data, the robot control system evaluates the feasibility and safety of the current path in real time. The real-time rescue driving path state analysis mainly determines whether there are objects that hinder the robot's progress, whether the path needs to be adjusted, and how to optimize the driving path to avoid collisions by analyzing the surrounding obstacles and terrain of the robot's current travel path. The generated real-time rescue driving path state data will include information such as the feasibility analysis results of the path, the detection of whether there are obstacles, the impact range of obstacles, and whether obstacle avoidance is required. The rescue robot obtains its own physical design parameters through built-in sensors and design parameter libraries, including but not limited to the robot's maximum load, maximum speed, range of motion, reach of the robot arm, output power of the drive motor, tire friction coefficient, structural stability parameters, etc. These design parameters usually come from the preset technical standards when the robot hardware is manufactured, or data collected in real time by sensors. Through the mechanical structure model stored in the system, each design parameter is obtained in real time and transmitted to the control system for subsequent motion controllability analysis. Based on the obtained physical design parameters of the rescue robot, the system uses an algorithm to perform controllable parameter analysis on the robot's motion performance. This analysis process focuses on the robot's motion ability under different loads and different environments, and mainly analyzes the robot's maximum motion speed, acceleration, travel stability, path turning radius, controllability of the body tires and other motion performance data. After obtaining the robot's controllable motion parameters, the system performs simulation operation posture parameter analysis based on these parameters. This analysis considers how the robot adapts to different terrain changes and obstacle avoidance when performing rescue missions. The simulation operation posture parameter analysis mainly includes the robot's tilt angle (lateral tilt angle and longitudinal tilt angle), center of gravity position, angle change of the robot arm, and the state of the tire in contact with the ground, etc., and simulates the movement performance in different environments through the simulation model. At this time, the simulation analysis program will analyze the robot's posture changes under different operating conditions according to different parameter settings. The robot simulation operation posture limit state analysis conducts an in-depth limit evaluation of the robot's simulation operation posture parameters, focusing on determining whether the robot will exceed its physical limits in extreme environments (such as severe slopes, high loads, fast turns, etc.). Through this analysis, the system evaluates the robot's posture change range and maximum tolerable pressure, and the stability of the robot under extreme conditions.The analysis will generate the robot's simulated operation posture limit state data and provide the robot's stability data under different extreme conditions. For example, when the robot performs a rescue mission, if the robot needs to travel on an extreme slope (such as a 30-degree slope), the simulation will analyze whether the robot will overturn due to unstable center of gravity or insufficient tire friction. By simulating and calculating various parameters, the system outputs the robot's limit state data, such as the maximum feasible climbing angle of 25 degrees and the maximum load of 10 kg when turning. These data provide a detailed reference for the limit performance for subsequent decisions, ensuring that the rescue mission can be safely performed within the physical range. When the robot performs a rescue mission, the robot's posture limit state data is first obtained through simulation calculation. This data reflects the maximum angle, load and stability under external conditions that the robot can withstand during the mission. Based on this data, the system analyzes the instant rescue driving path state data in the current environment to determine whether there is an area in the path that the robot cannot safely pass through. The process includes calculating factors such as possible obstacles in the path, terrain undulations, slopes, and the minimum friction required for the robot to operate, and evaluating the feasibility of the path. For example, if the path contains a 25-degree slope, and the robot's simulated running posture limit state data shows that the robot can only withstand a maximum slope of 20 degrees, then this path will be marked as infeasible. The system will optimize the originally infeasible path to a new path suitable for the robot based on the robot's simulated running posture limit state data. Finally, the feasible rescue driving path state data is generated to ensure that the robot can perform the rescue mission smoothly and safely. After obtaining the feasible rescue driving path state data, the control system further performs real-time adaptive driving posture analysis on the robot based on the robot's simulated running posture parameters. During this analysis, the system adjusts the robot's traveling posture in real time based on the robot's current running state, path data, and motion controllable parameters to cope with possible sudden changes in the external environment. This analysis not only involves the robot's own motion control, but also includes how to dynamically adjust the robot's posture according to the current environmental changes. For example, when the robot turns from a flat ground to a narrow passage with obstacles, the system will perform real-time analysis based on the feasible rescue driving path state data and simulated running posture parameters to calculate the optimal turning angle, travel speed, travel stability and other adaptive parameters. When there is a 0.5-meter pothole on the road surface where the robot is located, and the current robot is evaluated to be in a drivable state and the corresponding driving state is that its own lateral deviation is 0.5 meters, the current robot's situation and posture can be analyzed, such as the right fuselage tire of the robot is 0.5 meters sunken from the horizontal line, the center node is 0.3 meters sunken from the horizontal line, and the left fuselage tire is parallel to the horizontal line, etc.If the system analysis finds that the robot is facing a ramp or a sharp turn, the control system will adjust the robot's travel posture according to the simulated attitude limit state data, such as reducing speed, adjusting tire pressure, or performing dynamic balance compensation to ensure that the robot can move adaptively in complex environments. Ultimately, the real-time robot adaptive driving posture data is generated, including the robot's various adjustment parameters (such as speed, tilt angle (lateral tilt angle and longitudinal tilt angle), balance state, etc.), which will be used to control the robot's movement in real time, thereby ensuring that the task can be completed smoothly and responding to various changes in the actual environment.

[0074] S6: An instant control state intelligent optimization module is used to perform an instant robot adaptive driving rate analysis based on the instant rescue driving path state data and the instant robot adaptive driving posture data to generate instant robot adaptive driving rate data; based on the instant robot adaptive driving posture data and the instant robot adaptive driving rate data, an instant rescue emergency control state intelligent optimization operation is performed on the robot performing the rescue emergency intelligent control operation.

[0075] In the embodiment of the present invention, the robot driving constraint decision generally includes the maximum driving speed, maximum acceleration, slope limit, stability requirement, obstacle avoidance requirement and energy consumption limit that the robot can bear. These constraint decisions play a vital role in the driving performance of the robot in a complex environment. And consider the robot's motion parameters and simulation operation posture parameters. Combined with the robot simulation operation posture parameters, such as the robot's stability data under different slopes and different loads, tire friction coefficient, robot center of mass position, joint angle, driving angle and other information, and using the supervised learning algorithm and the historical robot rescue emergency control task rate driving data to design the cost function in the driving process. After obtaining the instant rescue driving path state data and the instant robot adaptive driving posture data, the system will transfer these data to the designed robot driving rate cost function. Through calculation and analysis, the cost function will evaluate the robot's operating performance at different speeds under the current path state. The system will generate instant robot adaptive driving rate data based on these analysis results. Based on the instant robot adaptive driving posture data and the instant robot adaptive driving rate data, the system will perform an instant control state intelligent optimization operation on the robot performing the rescue emergency intelligent control operation. The core task of this optimization operation is to enable the robot to move efficiently and stably while ensuring the completion of the task. The system will use the robot's driving rate cost function and path state data to dynamically adjust the robot's motion parameters, such as adjusting the robot's travel speed, posture, direction, etc. If the system detects that the robot's speed on a certain path is too fast, resulting in a decrease in stability, the system will appropriately reduce the speed based on the instant robot's adaptive driving rate data to avoid tilting or instability. At the same time, the system will continue to monitor various parameters of the robot's movement (such as speed, acceleration, direction of travel, and the lateral tilt angle and longitudinal tilt angle of the fuselage), and adjust the control strategy according to real-time changes in the environment. If obstacles or other unforeseen factors affect the robot's movement, the system will re-plan the path or adjust the speed according to the preset path and robot status to ensure that the rescue mission can be successfully completed.

[0076] Furthermore, the voice interaction analysis module includes the following functions:

[0077] Use the robot's built-in voice receiving function to collect control voice signals and generate control voice signals;

[0078] Use automatic speech recognition technology to perform rescue emergency control demand text analysis on the control demand voice signal to generate rescue emergency control demand text data;

[0079] Perform semantic parsing on the rescue emergency control demand text data to generate rescue emergency control demand semantic data;

[0080] According to the rescue emergency control demand semantic data, a structured analysis of the rescue emergency voice control task is performed to generate the rescue emergency voice control task structured data.

[0081] In the embodiment of the present invention, the voice interaction analysis module first collects the control voice signal through the built-in microphone array of the robot. The microphone array can realize 360-degree sound perception, automatically capture the sound in the surrounding environment, and perform preliminary signal preprocessing, such as removing background noise and echo suppression. The collected control voice signal is transmitted to the signal processing unit to generate the original digital voice signal. At this time, the system can distinguish the sound signals from different directions and ensure that the quality of the signal is not affected by environmental factors, so as to ensure that the collected control voice signal has high clarity and accuracy. Once the control voice signal is successfully collected, the automatic speech recognition (ASR) processing stage is entered next. The robot uses efficient automatic speech recognition technology to convert the voice signal into text data. In this process, the control voice signal is first extracted by the speech recognition engine for speech feature extraction and acoustic model matching, and the language pattern in the audio is compared with the preset vocabulary. Specific words and sentences are identified and converted into control demand text data, for example, "start a rescue mission" or "go to a designated location". This text data records the basic information of the user's intention and is the basis for subsequent analysis. The recognition process uses a speech recognition model based on deep learning, which has a high recognition accuracy and can ensure a low misrecognition rate even in noisy environments. The control demand text data is semantically parsed, and the system will further analyze the recognized text content. Through natural language processing (NLP) technology, the system first extracts keywords and phrases from the text data, such as "rescue mission", "go to", "specified location", etc. On this basis, the system applies a semantic analysis model to deeply understand the intention, action and goal in the text. For example, the input text "start the rescue mission" will be parsed as "action: start", "goal: rescue mission". Semantic parsing also includes correlation analysis of the context in the text to ensure that the user's real needs are understood, and structured semantic data is generated to indicate the task goal, task type, specific operation instructions, etc. This process can accurately identify the implicit task requirements in the text by relying on a pre-trained natural language model. After completing the semantic parsing, the rescue emergency control demand semantic data enters the structured analysis stage. In this stage, the system first classifies, filters and sorts the semantic data, and generates task priorities based on the urgency of the task, the order of operations, etc. For example, for the two commands "go to the designated location" and "start emergency rescue", the system will determine which task should be executed first based on the predetermined priority rules, and determine the robot's action path based on the geographic location information. Then, the system converts these structured task data into a robot control instruction set, including action execution instructions, task target location, task type, etc., to form structured data that can be directly transmitted to the robot control system. This process not only ensures the accurate understanding of the user's intentions, but also provides clear task guidance for the robot's subsequent intelligent decision-making and behavior execution.

[0082] Further, as an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed functional flow diagram of the robot image perception modeling module. In this embodiment, the robot image perception modeling module includes the following functions:

[0083] S201: Perform panoramic image monitoring processing using the built-in panoramic image monitoring device of the robot to generate panoramic image data;

[0084] In an embodiment of the present invention, panoramic image monitoring processing is performed by a panoramic image monitoring device built into the robot. The panoramic image monitoring device includes multiple high-resolution cameras arranged at fixed angles to cover a 360-degree all-round field of view around the robot. By synchronously collecting image data from these cameras, the robot can obtain a panoramic image. The image data will undergo preliminary image processing, including distortion correction and color balance, to ensure the accuracy and clarity of the image. Finally, the processed panoramic image data will be transmitted to the data storage module for subsequent analysis and processing.

[0085] S202: Collecting environmental perception data using the built-in environmental perception sensor of the robot to generate environmental perception data;

[0086] In an embodiment of the present invention, while collecting panoramic image data, the robot also uses built-in environmental perception sensors to collect environmental perception data. Environmental perception sensors generally include temperature and humidity sensors, gas sensors, acoustic wave sensors, etc., which can monitor the temperature, humidity, gas concentration and other information of the surrounding environment in real time. In rescue missions, environmental perception data is crucial for judging the safety of the surrounding environment and risk assessment. For example, gas sensors can detect the leakage of harmful gases, and temperature and humidity sensors can monitor changes in special environments such as fires or floods. The collected environmental perception data will be converted into digital format, saved as environmental perception data files, and transmitted to the system for subsequent processing.

[0087] S203: collecting long-range laser radar data using a long-range laser radar device built into the robot to generate long-range laser radar data;

[0088] In the embodiment of the present invention, in order to further enhance the robot's ability to perceive the environment, the robot's built-in long-range laser radar device (LiDAR) is used to collect long-range laser radar data. By emitting laser beams and receiving reflected signals, the laser radar can accurately measure the distance to surrounding objects and generate three-dimensional point cloud data of the environment. The robot's built-in laser radar can cover a certain range of environments, obtain multi-dimensional spatial information by scanning at different angles, and generate long-range laser radar data. The data includes the specific location, shape, distance and reflection intensity information of the object, and is stored in the form of a point cloud. These data will be sent to the subsequent point cloud analysis module to generate a subsequent point cloud model.

[0089] S204: performing long-range laser point cloud data analysis according to the long-range laser radar data to generate long-range laser point cloud data;

[0090] In the embodiment of the present invention, after the long-range laser radar data is collected, the system will analyze the laser point cloud data. Specifically, the robot will use the point cloud processing algorithm to perform denoising, point cloud reconstruction, and sparse point supplementation on the data based on the distance information in the laser point cloud data. The processed long-range laser point cloud data can more accurately describe obstacles, buildings, and terrain features in the environment. The refined processing of point cloud data also includes alignment and optimization of the point cloud and the environmental model to ensure that the generated point cloud data is highly consistent with the object morphology in the actual environment.

[0091] S205: Transmitting the long-distance laser point cloud data to the panoramic image data for global image point cloud modeling processing to generate a global image point cloud model;

[0092] In an embodiment of the present invention, after the long-distance laser point cloud data is analyzed, it will be combined with the panoramic image data to perform global image point cloud modeling processing. First, the system will align and match the common viewpoints in the panoramic image data and the laser point cloud data to ensure the spatial consistency of the image and point cloud data. The system uses an image matching algorithm to accurately match the feature points in the image with the corresponding points in the point cloud to form a set of correspondences between the image and the point cloud data. Next, based on the matching results, the system generates a comprehensive global image point cloud model through a three-dimensional modeling algorithm. This model not only contains the image data of the surrounding environment, but also integrates accurate three-dimensional spatial information.

[0093] S206: performing a hierarchical distance division process on the environmental perception data to generate hierarchical environmental perception data;

[0094] In the embodiment of the present invention, as the global image point cloud model is established, the system continues to perform hierarchical distance division processing on the environmental perception data. The system performs hierarchical processing on the environment based on the environmental data collected by the environmental perception sensor, such as temperature, humidity, gas concentration and other information. The environmental data of each layer will be divided according to its impact range on the robot, and multiple levels of data sets will be formed. The data of each level includes environmental parameters in a specific area, such as temperature range, gas concentration, etc. This processing process helps the system to judge environmental changes at each level and make more accurate responses and decisions.

[0095] S207: Performing hierarchical environment perception time series feature analysis on the hierarchical environment perception data to generate hierarchical environment perception time series feature data;

[0096] In an embodiment of the present invention, the environmental data is sorted according to the time series, and is classified into different time periods (such as data per hour, per minute, or per second) according to the timestamp of the sensor collecting the data. Next, a time series analysis method (such as a sliding window or Fourier transform) is used to extract time series features from these data. By analyzing these data, the system can identify periodic changes in environmental data, sudden events (such as a sharp rise in temperature), and long-term trends (such as a continuous increase in gas concentration). For example, when the gas concentration continues to increase over a period of time, the system will mark the event as a potential danger signal. The generated hierarchical environmental perception time series feature data includes the change amplitude, frequency and other relevant statistical characteristics of variables such as temperature and gas concentration in each time period, which serve as the basic data for subsequent trend analysis.

[0097] S208: performing hierarchical environment perception trend analysis according to the hierarchical environment perception time series feature data to generate hierarchical environment perception trend data;

[0098] In an embodiment of the present invention, the changing trends of environmental data in different time periods are analyzed. By comparing the difference between the data of the current time period and the historical data, regression analysis, exponential smoothing or other trend prediction methods are used to identify the changing pattern of environmental data. For example, when the temperature rises significantly in a short period of time and the gas concentration also changes abnormally, the system will judge it as a potential fire risk and calculate the time window when the danger may occur. In this way, the system can analyze the changing trends of the environment in real time and generate hierarchical environmental perception trend data. These trend data not only reflect the speed and magnitude of environmental changes, but also predict the changing trajectory of the future environment, providing a basis for risk prediction and decision-making in robot rescue missions. For example, the temperature may reach a dangerous critical point in the next few minutes. The system generates an early warning signal through this data to guide the robot to adjust the path or take corresponding actions.

[0099] S209: Transmitting environmental perception data to the global image point cloud model for environmental perception data mapping processing to generate a global image perception model; and performing collaborative optimization processing of a three-dimensional model of multi-level environmental changes on the global image perception model through hierarchical environmental perception trend data to generate an optimized global image perception model.

[0100] In the embodiment of the present invention, in the process of mapping environmental perception data to the global image point cloud model, it is first necessary to convert the data from the environmental perception sensor into a format compatible with the panoramic image point cloud model. Specifically, the image data obtained by the robot through the panoramic image and the three-dimensional point cloud data collected by the laser radar need to be converted into a unified spatial coordinate system. The system associates each level of environmental perception data (such as temperature, humidity, gas concentration, etc.) with the corresponding position in the image point cloud data. For example, if the gas sensor detects that the gas concentration in a certain area exceeds the standard, the system will map the gas concentration data to the point cloud coordinates corresponding to the area. Through this data mapping, the system can combine real-time environmental perception information with the three-dimensional environmental model to form a real-time updated global image perception model. The model not only contains the image data and physical space information of the environment, but also reflects the immediate state of the surrounding environment, such as high temperature areas or harmful gas leakage areas. After completing the environmental perception data mapping, the system then performs a multi-level environmental change three-dimensional model collaborative optimization process of the global image perception model. The specific implementation steps are as follows: First, based on the hierarchical environmental perception trend data, the system analyzes the change trend of the current environment. For example, if the gas concentration continues to rise in a certain area and the temperature gradually increases, the system will determine that there may be a fire risk in the area and give priority to the panoramic image point cloud data of the area. Through the weighted average method, the system corrects the point cloud data of the area, adjusts the density, accuracy and reliability of the point cloud, and thus optimizes the global image perception model of the area. Secondly, the system uses algorithms (such as image-based deep learning methods) to further process the optimized point cloud data, combines the data of each environmental perception level, such as temperature and humidity data, gas concentration data, etc., to identify potential dangerous areas in the environment, and mark them according to different risk levels. For example, in areas with high gas concentrations, the system will mark the danger level of the area in the point cloud model and provide dynamic safe path suggestions. In this way, the optimized global image perception model can more accurately reflect the real-time changes in the environment, especially those key changing factors that affect the execution of robot rescue missions, such as leakage of hazardous materials, fire, and water accumulation.

[0101] Furthermore, the transmitting of the long-distance laser point cloud data to the panoramic image data for global image point cloud modeling processing includes:

[0102] Perform image matching node analysis based on panoramic image data to generate image matching node data;

[0103] Perform laser point cloud image matching data analysis based on image matching node data and long-distance laser point cloud data to generate laser point cloud image matching data;

[0104] Based on the laser point cloud image matching data, the long-distance laser point cloud data is transmitted to the panoramic image data for global image modeling processing to generate a global image point cloud model.

[0105] In an embodiment of the present invention, when performing image matching node analysis, the panoramic image data is first processed. The panoramic image data comes from the built-in panoramic camera of the robot. These images are two-dimensional image data. By extracting the color, texture, edge and other features of the image, the system calculates the feature value of each pixel or area. These feature values ​​include color brightness, edge gradient, texture details, etc., and the matching nodes with significant features in the image are identified by algorithms (such as SIFT or SURF). Each matching node consists of a set of feature values ​​and corresponding spatial coordinates, representing an important position in the image. The selection of matching nodes is based on the local saliency of the image, and those areas with little change are selected to ensure that matching analysis can be effectively performed. For example, in a panoramic image containing the facade of a building and a street scene, the system can extract windows or doors on the wall as matching nodes. Through these matching nodes, the key positions in the image can be determined, and basic data can be provided for subsequent point cloud matching and global modeling. The generated image matching node data includes the two-dimensional position coordinates of each node, the relevant image feature vector, and the relative importance of the node in the entire panoramic image. When performing laser point cloud image matching data analysis based on image matching node data and long-range laser point cloud data, it is first necessary to convert the long-range laser point cloud data into coordinate system information in three-dimensional space. The laser radar obtains the three-dimensional spatial coordinate data of the surrounding environment by emitting laser beams and receiving reflected signals. These data form a point cloud data set containing attributes such as point coordinates and reflection intensity. When pairing image matching node data with laser point cloud data, the difference in spatial coordinate systems between the two must be considered first. Coordinate transformation methods (such as Euclidean transformation or affine transformation) are usually used to align the image data and laser point cloud data. Then, by analyzing the projection position of each image matching node in the point cloud space, the points in the laser point cloud closest to these nodes are found. These points are marked as potential matching points based on features such as position, color, and reflection intensity. By comparing image features (such as texture, edges, color changes, etc.) with point cloud features (such as distance, reflection intensity, etc.), the system further verifies the accuracy of the match and eliminates the possibility of mismatching. For example, suppose there is a window area in the panoramic image data. After image matching node analysis, the four corners of the window are identified as matching nodes. By comparing the laser point cloud data, the area corresponding to the window will be mapped to the corresponding position in the laser point cloud. Each point of the laser point cloud will be compared with the node in the image according to its spatial coordinates to generate laser point cloud image matching data. When transferring the long-distance laser point cloud data to the panoramic image data and performing global image point cloud modeling processing, the accurate correspondence between the panoramic image and the laser point cloud is first established through the laser point cloud image matching data. Using the image matching node data and the laser point cloud image matching data, the system accurately aligns the spatial position of each laser point cloud point with the corresponding position in the image.At this point, the laser point cloud and image data share a unified spatial coordinate system to ensure that the laser point cloud data can reflect the actual object position in the image. Then, using the image matching node data and point cloud data, the system further processes each laser point cloud data and adjusts its position, size and density to ensure that the point cloud accurately reproduces the physical environment in three-dimensional space. For example, if the density of the laser point cloud is too low, the system will use an interpolation algorithm to supplement the missing points to ensure the continuity and integrity of the point cloud. If there is noise or error in the point cloud data, the system will filter and denoise according to the eigenvalues ​​of the image data to ensure the precision and accuracy of the point cloud. The generated global image point cloud model combines the high-precision three-dimensional data of panoramic images and lidar, and can accurately describe the structure of the physical environment. For example, in a complex urban building environment, the system generates a three-dimensional point cloud model containing buildings, streets, trees, vehicles and other objects by matching images and laser point clouds. This model not only has visual information, but also integrates the depth and reflection intensity information provided by the lidar, which can achieve high-precision environmental modeling and real-time perception.

[0106] Furthermore, performing image matching node analysis based on the panoramic image data includes:

[0107] Performing global pixel blur interference evaluation processing on the panoramic image data to generate global pixel blur interference data;

[0108] Perform global image effective area analysis based on global pixel fuzzy interference data to generate global image effective area data;

[0109] Performing global image effective area gradient feature analysis on global image effective area data to generate global image effective area gradient feature data;

[0110] Image matching node analysis is performed based on the gradient feature data of the effective area of ​​the global image to generate image matching node data.

[0111] In an embodiment of the present invention, when performing global pixel blur interference evaluation processing on panoramic image data, the panoramic image needs to be processed as a two-dimensional matrix first. Each pixel of the image represents a pixel area in the image, and its value is usually color or brightness information. The system evaluates the overall clarity of the image by calculating the blurriness of each pixel in the entire image. Global pixel blur interference data is generated by calculating the blurriness, smoothness and other characteristics of each pixel in the image. The specific blur metric value can be based on the Gaussian blur algorithm or other related spatial frequency calculation methods, reflecting the detail complexity of the image in the local area. For example, if there are a large number of small objects or areas in a panoramic image with relatively uniform texture and lack of edge changes, then these areas will show a higher blurriness. By evaluating all pixels, global pixel blur interference data can be generated, in which each value represents the blur feature intensity of the area. Such evaluation data provides an accurate blur reference for subsequent effective area analysis. When performing global image effective area analysis based on global pixel blur interference data, it is first necessary to determine the clear area in the image through the calculation result of the blurriness data. The system takes the global pixel fuzzy interference data as input and selects the area with high visual information contrast by setting a certain pixel fuzziness threshold. The effective area data is formed by screening the pixel set with clarity higher than the threshold in the panoramic image to form a set of areas with relatively concentrated image features. For example, in a panoramic image of the exterior of a building, the fuzziness value of each pixel is first calculated. If the fuzziness value of the pixel in an area is lower than the set threshold (such as 0.05), the area is defined as a valid area. These valid areas are mainly parts with obvious edges, rich textures and high contrast features, which can be more clearly recognized and processed by the machine. When performing global image effective area gradient feature analysis on the global image effective area data, it is first necessary to calculate the gradient features of each pixel in these valid areas. The gradient feature refers to the change direction and amplitude of the area where each pixel is located in the panoramic image. Using gradient calculation methods such as Sobel operator, Canny operator or Prewitt operator, the gradient information of the edge area in the image can be calculated, that is, the change direction of the edge and the sharpness of the boundary. Gradient feature data is the directional change and intensity value calculated by these algorithms, such as the horizontal and vertical gradient values ​​of pixels to reflect the gradient direction of the image. For example, in a panoramic image of a city street, the gradient calculation of each pixel in the effective area can obtain the direction and intensity of the edge in each area. For areas in the image where the edge changes of objects are more prominent, their gradient feature values ​​will be larger, reflecting a higher local boundary information intensity. Through these gradient feature data, the system can more effectively analyze the image matching nodes in subsequent steps.When performing image matching node analysis based on the gradient feature data of the effective area of ​​the global image, the gradient feature information in the effective area data of the global image should first be used to find the area with the most contrast and clear edges. Then, the statistical data of the gradient direction and intensity changes in these areas are used to match the nodes in the image. For example, the gradient direction and amplitude in each effective area can be calculated to find areas with similar gradient features, and these areas can be regarded as potential matching nodes. The data of the matching node includes the characteristic values ​​of each node, such as the two-dimensional coordinates, gradient direction, and gradient intensity, which are used to describe the specific position and boundary characteristics in the image. If there is an edge area of ​​a building in the panoramic image, after gradient feature analysis, it can be found that the edge of the top of the building has the most prominent gradient intensity change. After calculating the gradient feature value of the area, the system uses it as an image matching node for subsequent matching analysis of the laser point cloud and the image.

[0112] Furthermore, the robot rescue emergency intelligent control module includes the following functions:

[0113] Based on the structured data of rescue emergency voice control tasks and the optimized global image perception model, the rescue emergency automated control decision analysis is performed to generate rescue emergency automated control decisions;

[0114] Based on the rescue emergency automated control decision, the robot performs rescue emergency intelligent control operations.

[0115] In an embodiment of the present invention, the system analyzes the control requirements contained in the task based on the rescue emergency voice control task structured data obtained from the voice interaction module. The control task structured data contains key task instructions and rescue requirements after voice signal recognition, such as "move to the disaster site" or "open the rescue channel". The content in the task structured data is converted into a set of standardized instruction sets that can express the actions that the robot needs to complete, such as moving, picking up, switching, etc. On this basis, combined with the environmental perception data provided in the optimized global image perception model, the system evaluates the current environmental state, such as the obstacle distribution, spatial layout, channel smoothness and other information obtained through the panoramic image perception model and laser point cloud data. The system performs environmental adaptability analysis according to the needs of the rescue task, determines which areas are accessible, which areas need to be avoided, and the optimal path planning. Combined with the above information, the system enters the automated control decision stage. At this stage, the system performs decision analysis through an algorithm model based on information such as task requirements, the current state of the environment, and the robot's own motion capabilities, and generates the most appropriate automated control decision. For example, if the task requires the robot to go to the disaster area and open a passage, the system will generate a comprehensive decision-making plan including path planning, obstacle avoidance, motion control, etc., which will take into account factors such as the robot's current state, environmental layout, and mission objectives. After receiving the rescue emergency automated control decision, the robot begins to perform the corresponding intelligent control operation. The system commands the robot to move along the specified path based on the previously generated automated control decision. This path will be updated in real time through panoramic images and lidar data to ensure that the robot avoids obstacles and maintains the effectiveness of the path during execution. For example, if the control decision requires the robot to go to an area covered by collapsed objects and perform rescue, the system will plan the path based on the current environmental perception data to ensure that the robot can avoid unstable or blocked areas. When the robot is driving, the control system will monitor the changes in the surrounding environment in real time, and make dynamic adjustments based on the real-time data of the panoramic image perception model and the laser point cloud to ensure that the robot can flexibly respond to changes in the environment. At the same time, if the control decision requires the robot to perform specific tasks, such as picking up objects or opening passages, the robot will automatically perform corresponding actions according to the control instructions. For example, when picking up an object, the robot will perform precise grasping movements to ensure that the object is firmly grasped, and adjust the movement path and posture according to the task requirements to ensure the completion of the rescue mission. During the execution process, the robot will also communicate with the control system to feedback the current status and task progress, such as whether it successfully avoids obstacles and whether the task is completed. In this way, through the combination of voice-based commands and environmental perception data, the robot can efficiently and accurately complete the rescue emergency intelligent control operation.

[0116] Further, as an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Detailed functional flow diagram of the real-time running posture analysis module in FIG. 1 . In this embodiment, the real-time running posture analysis module includes the following functions:

[0117] S501: using the robot's built-in short-range laser radar device to perform short-range laser radar real-time monitoring and processing on the robot performing the rescue emergency intelligent control operation, and generating real-time short-range laser radar data;

[0118] In an embodiment of the present invention, the short-range laser radar device built into the robot performs a real-time scan of the surrounding environment through a laser beam during the rescue emergency intelligent control operation, obtains distance data and generates real-time short-range laser radar data. The device calculates the distance by emitting a laser and measuring the time of the reflected laser pulse, accurately detecting the relative position and distance between the robot and the surrounding obstacles. Specifically, when the robot is performing an "emergency search" task, its short-range laser radar scans the surrounding environment and captures various distance data including walls, objects, obstacles, etc. These data contain distance information, the spatial position of the obstacle, the azimuth, and the size and shape of the obstacle, thereby providing the necessary real-time data support for path planning and obstacle avoidance.

[0119] S502: Performing an immediate rescue driving path state analysis based on the immediate short-range laser radar data to generate immediate rescue driving path state data;

[0120] In the embodiment of the present invention, based on the real-time short-range laser radar data, the system performs real-time rescue driving path state analysis on the path taken by the robot when performing the rescue mission. By analyzing the obstacle position provided by the laser radar data, the distance between obstacles and their impact on the robot, the robot control system evaluates the feasibility and safety of the current path in real time. The real-time rescue driving path state analysis mainly determines whether there are objects that hinder its forward movement, whether the path needs to be adjusted, and how to optimize the driving path to avoid collision by analyzing the surrounding obstacles and terrain of the robot's current travel path. For example, when the robot is performing the "search for trapped persons" task, an obstacle with a width of 1.2 meters and a height of 0.8 meters appears 2 meters in front. The control system will analyze in real time whether the obstacle will affect the robot's travel path, and decide whether it is necessary to adjust the driving direction or adopt a detour strategy based on the distance, shape and height data of the obstacle. The generated real-time rescue driving path state data will include the feasibility analysis results of the path, the detection of whether there are obstacles, the impact range of obstacles, whether obstacle avoidance is needed, and other information.

[0121] S503: Obtaining physical design parameters of the rescue robot;

[0122] In the embodiment of the present invention, the rescue robot obtains its own physical design parameters through built-in sensors and design parameter libraries, including but not limited to the robot's maximum load, maximum speed, range of motion, extension range of the robot arm, output power of the drive motor, tire friction coefficient, structural stability parameters, etc. These design parameters usually come from the preset technical standards when the robot hardware is manufactured, or data collected in real time by sensors. Through the mechanical structure model stored in the system, each design parameter is obtained in real time and transmitted to the control system for subsequent motion controllability analysis.

[0123] S504: Analyze the controllable motion parameters of the robot according to the physical design parameters of the rescue robot, and generate the controllable motion parameters of the robot;

[0124] In the embodiment of the present invention, based on the obtained physical design parameters of the rescue robot, the system performs controllable parameter analysis on the robot's motion performance through an algorithm. This analysis process focuses on the robot's motion ability under different loads and different environments, and mainly analyzes the robot's maximum motion speed, acceleration, travel stability, path turning radius, controllability of the fuselage tires and other motion performance data. During the analysis process, the system simulates and calculates the robot's motion response under different load conditions, taking into account ground friction, the flexibility of the steering system and the bearing capacity of the mechanical structure. For example, assuming that the robot can travel at a maximum speed of 1.5 meters per second in an unloaded state, the lateral bilateral deviation of the vehicle can be 0.8 meters, but if the load reaches 15 kilograms, the robot's motion speed will drop to 1 meter per second, and the lateral bilateral deviation of the vehicle can be 0.3 meters. Based on these physical parameters, the control system further calculates the controllability of the robot's motion under various conditions, generates controllable parameters for the robot's motion, and outputs analysis results including speed, steering angle, load capacity, etc.

[0125] S505: Analyze the robot simulation operation posture parameters based on the robot motion controllable parameters, and generate the robot simulation operation posture parameters;

[0126] In an embodiment of the present invention, after obtaining the controllable parameters of the robot's motion, the system performs a simulation operation posture parameter analysis based on these parameters. This analysis considers how the robot adapts to different terrain changes and obstacle avoidance when performing rescue tasks. The simulation operation posture parameter analysis mainly includes the robot's tilt angle (lateral tilt angle and longitudinal tilt angle), center of gravity position, angle change of the robotic arm, and the tire state in contact with the ground, etc., and simulates the motion performance in different environments through the simulation model. At this time, the simulation analysis program will analyze the robot's posture changes under different operating conditions according to different parameter settings. For example, when the robot performs a task in a narrow space, the system will analyze whether the robot can move stably or whether it will overturn due to excessive turning angles through simulation operation posture parameter analysis. During the analysis, if the robot needs to pass a ramp, the simulation will analyze whether it can maintain a stable posture to avoid center of gravity shift.

[0127] S506: performing a robot simulation operation posture limit state analysis on the robot simulation operation posture parameters to generate robot simulation operation posture limit state data;

[0128] In the embodiment of the present invention, the robot simulation operation posture limit state analysis conducts an in-depth limit evaluation of the robot simulation operation posture parameters, focusing on determining whether the robot will exceed its physical limit under extreme conditions (such as severe slopes, high loads, fast turns, etc.). Through this analysis, the system evaluates the robot's posture change range and maximum tolerable pressure, and the stability of the robot under extreme conditions. The analysis will generate the robot simulation operation posture limit state data, and provide the robot's stability data under different extreme conditions. For example, when the robot performs a rescue mission, if the robot needs to travel on an extreme slope (such as a 30-degree slope), the simulation will analyze whether the robot will overturn due to unstable center of gravity or insufficient tire friction. By simulating and calculating various parameters, the system outputs the robot's limit state data, such as the maximum feasible climbing angle of 25 degrees, the maximum load when turning is 10 kilograms, etc. These data provide a detailed reference for the limit performance for subsequent decision-making, ensuring that the rescue mission can be safely performed within the physical range.

[0129] S507: performing feasibility analysis on the immediate rescue driving path state data according to the robot simulation operation posture limit state data to obtain feasible rescue driving path state data;

[0130] In an embodiment of the present invention, when the robot performs a rescue mission, the robot's attitude limit state data is first obtained through simulation calculation. This data reflects the maximum angle, load and stability under external conditions that the robot can withstand during the mission. Based on this data, the system analyzes the instant rescue driving path state data in the current environment to determine whether there is an area in the path that the robot cannot pass safely. The process includes calculating factors such as possible obstacles, terrain undulations, slopes, and the minimum friction required for the robot to operate, and evaluating the feasibility of the path. For example, if the path contains a slope of 25 degrees, the robot simulation operation attitude limit state data indicates that the robot can only withstand a maximum slope of 20 degrees, then this path will be marked as infeasible. The system will optimize and adjust the originally infeasible path to a new path suitable for the robot based on the robot simulation operation attitude limit state data. Finally, the feasible rescue driving path state data is generated to ensure that the robot can smoothly and safely perform the rescue mission.

[0131] S508: Performing instant robot adaptive driving posture analysis according to the feasible rescue driving path state data and the robot simulation operation posture parameters to generate instant robot adaptive driving posture data.

[0132] In the embodiment of the present invention, after obtaining the state data of the feasible rescue driving path, the control system further performs an instant adaptive driving posture analysis on the robot according to the robot simulation running posture parameters. During the analysis process, the system adjusts the robot's traveling posture in real time according to the robot's current running state, path data and motion controllable parameters to cope with sudden changes that may occur in the external environment. This analysis not only involves the motion control of the robot itself, but also includes how to dynamically adjust the robot's posture according to the current environmental changes. For example, when the robot turns from a flat ground to a narrow passage with obstacles, the system will perform real-time analysis based on the state data of the feasible rescue driving path and the simulation running posture parameters, and calculate the optimal turning angle, travel speed, travel stability and other adaptive parameters. When the robot's situation road surface has a pit of 0.5 meters, and the current robot is evaluated to be in an operative state and the corresponding driving state is a lateral deviation of 0.5 meters, the current robot's situation posture can be analyzed, such as the robot's right fuselage tire is 0.5 meters sunken in the deviation horizontal line, the center node is 0.3 meters sunken in the deviation horizontal line, and the left fuselage tire is in a parallel form to the horizontal line, etc. If the system analysis finds that the robot is facing a ramp or a sharp turn, the control system will adjust the robot's travel posture according to the simulated attitude limit state data, such as reducing speed, adjusting tire pressure, or performing dynamic balance compensation to ensure that the robot can move adaptively in complex environments. Ultimately, the real-time robot adaptive driving posture data is generated, including the robot's various adjustment parameters (such as speed, tilt angle (lateral tilt angle and longitudinal tilt angle), balance state, etc.), which will be used to control the robot's movement in real time, thereby ensuring that the task can be completed smoothly and responding to various changes in the actual environment.

[0133] Furthermore, the feasibility analysis of the immediate rescue path status data includes:

[0134] When the immediate rescue driving path status data conforms to the robot simulation operation posture limit state data, the immediate rescue driving path status data is marked as feasible rescue driving path status data; otherwise, the immediate rescue driving path status data is marked as abnormal rescue driving path status data, and the abnormal rescue driving path status data is fed back to the rescue emergency intelligent control operation to execute the rescue emergency driving path intelligent adjustment operation.

[0135] In the embodiment of the present invention, when the robot performs a rescue mission, it first obtains the status data of the immediate rescue driving path. The data includes the terrain information, obstacle positions, slopes, space dimensions, and the current positioning information of the robot at each key point in the path. Next, according to the robot simulation operation attitude limit state data, the data reflects the maximum load, slope, speed, and other motion parameters that the robot can bear during the movement. The system compares the immediate rescue driving path status data with the simulation operation attitude limit state data to determine whether the path meets the actual operation capacity of the robot. First, the system analyzes the slope, obstacles, and travel space of each section in the path one by one according to the immediate rescue driving path status data. If the slope of any part of the path exceeds the maximum slope range in the robot simulation operation attitude limit state data, or the obstacle of the path blocks the size that exceeds the robot's passing capacity, the system will determine that the path is infeasible. For example, if there is a 30-degree ramp in the path, and the simulation operation attitude limit state data shows that the robot can only adapt to a maximum ramp of 20 degrees, the system will determine that the path is an abnormal state path. On the contrary, if the slope of the path is 15 degrees, which is in line with the robot's motion capability range, the system marks this path as a feasible rescue driving path. For the immediate rescue driving path state data that meets the robot's simulation operation attitude limit state data, the system marks this path as a feasible rescue driving path state data and continues to perform subsequent tasks. If the path data exceeds the limit of the simulation operation attitude limit state data, the path will be marked as abnormal rescue driving path state data. When the path data is marked as an abnormal rescue driving path, the system will start the intelligent adjustment operation of the driving path in the rescue emergency intelligent control module according to the abnormal data. The system feeds back the abnormal path information to the control module, and then recalculates and optimizes the path according to the obstacle position, slope information, etc. For example, the system may adjust the robot's travel angle, speed, or directly modify the path planning to ensure that the robot can find a suitable travel path that meets its operating limits. For the rescue driving path marked as abnormal, the control system performs intelligent path adjustment based on the obstacle information of the path and the simulation analysis results. Specific adjustments may include: replanning the path, avoiding collision detour strategies, or adjusting the robot's travel method (such as deceleration, changing posture, etc.). The adjusted path will be simulated again to ensure that the new path meets the constraints of the robot's motion.

[0136] Furthermore, the robot simulation operation posture parameters include the robot simulation operation pitch angle, the robot simulation operation roll angle and the robot simulation operation heading angle.

[0137] Furthermore, the real-time control state intelligent optimization module includes the following functions:

[0138] Design the robot driving rate cost function based on the preset robot driving constraint decision and the robot simulation operation posture parameters, and generate the robot driving rate cost function;

[0139] The instant rescue driving path state data and the instant robot adaptive driving posture data are transmitted to the robot driving speed cost function to perform the instant robot adaptive driving speed analysis, and generate the instant robot adaptive driving speed data;

[0140] Based on the real-time robot adaptive driving posture data and the real-time robot adaptive driving speed data, the robot performing the rescue emergency intelligent control operation is intelligently optimized in the real-time rescue emergency control state.

[0141] In the embodiment of the present invention, the robot driving constraint decision generally includes the maximum driving speed, maximum acceleration, slope limit, stability requirement, obstacle avoidance requirement and energy consumption limit that the robot can bear. These constraint decisions play a vital role in the driving performance of the robot in a complex environment. And consider the robot's motion parameters and simulation operation posture parameters. Combined with the robot simulation operation posture parameters, such as the robot's stability data under different slopes and different loads, tire friction coefficient, robot center of mass position, joint angle, driving angle and other information, and using the supervised learning algorithm and the historical robot rescue emergency control task rate driving data to design the cost function in the driving process. This cost function not only considers the maximum speed of the robot, but also needs to comprehensively consider the impact of factors such as obstacles, turning radius, and slope on the robot's driving performance. For example, in the case of a steep slope, the increase in the robot's driving speed will lead to a decrease in stability, thereby increasing the risk of overturning. Therefore, the cost function will constrain the driving rate to ensure that the robot always remains within a safe operating range. In addition, the robot's energy consumption and task efficiency also need to be considered when designing the cost function. For example, driving at a high speed for a long time will result in greater battery consumption, while too slow a speed may delay the completion time of the task. By balancing energy consumption and task efficiency, the cost function will calculate a suitable driving speed so that the robot can save as much energy as possible when completing the task and complete the rescue task in the shortest time. After obtaining the real-time rescue driving path state data and the real-time robot adaptive driving posture data, the system will transfer these data to the designed robot driving speed cost function. Through calculation and analysis, the cost function will evaluate the robot's operating performance at different speeds under the current path state. The system will generate real-time robot adaptive driving speed data based on these analysis results. For example, assuming that the real-time rescue driving path state data indicates that the current path has certain obstacles and a large slope, and the real-time robot adaptive driving posture data shows that the robot can pass through this area stably, the system will calculate the optimal travel speed of the robot in the current environment based on these data. If the path is relatively flat and there are fewer obstacles, the system will generate a higher travel speed to improve task efficiency. Based on the real-time robot adaptive driving posture data and the real-time robot adaptive driving speed data, the system will perform real-time control state intelligent optimization operations on the robot performing the rescue emergency intelligent control operation. The core task of this optimization operation is to enable the robot to travel efficiently and stably while ensuring the completion of the task. The system will use the robot's travel rate cost function and path state data to dynamically adjust the robot's motion parameters, such as the robot's travel speed, posture, direction, etc. If the system detects that the robot is moving too fast on a certain path, resulting in reduced stability, the system will appropriately reduce the speed based on the instant robot's adaptive travel rate data to avoid tilting or instability.At the same time, the system will continuously monitor various parameters of the robot during its movement (such as speed, acceleration, direction of movement, and lateral and longitudinal tilt angles of the body), and adjust the control strategy according to real-time changes in the environment. If obstacles or other unforeseen factors affect the robot's movement, the system will re-plan the path or adjust the speed according to the preset path and robot status to ensure that the rescue mission can be successfully completed.

[0142] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0143] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A robot rescue emergency control system based on voice interaction, characterized in that: Includes the following modules: The voice interaction analysis module is used to collect control voice signals using the robot's built-in voice receiving function to generate control voice signals; perform structured analysis of rescue emergency voice control tasks based on control voice signals to generate rescue emergency voice control task structured data; The robot image perception modeling module is used to obtain the panoramic image data, environmental perception data and long-range laser radar data monitored by the robot; perform global image point cloud modeling processing based on the panoramic image data and long-range laser radar data to generate a global image point cloud model; perform three-dimensional model collaborative optimization processing of multi-level environmental changes on the global image point cloud model based on the environmental perception data to generate an optimized global image perception model; The robot rescue emergency intelligent control module is used to perform rescue emergency intelligent control operations on the robot based on the rescue emergency voice control task structured data and the optimized global image perception model; The robot rescue emergency feedback module is used to collect the immediate rescue emergency task status feedback data for the rescue emergency intelligent control operation to obtain the immediate rescue emergency task status feedback data; transmit the immediate rescue emergency task status feedback data to the terminal to perform the immediate rescue emergency task status feedback operation; The real-time running posture analysis module is used to use the robot's built-in short-range laser radar equipment to perform short-range laser radar real-time monitoring and processing on the robot performing the rescue emergency intelligent control operation, and generate real-time short-range laser radar data; perform real-time rescue driving path state analysis based on the real-time short-range laser radar data, and generate real-time rescue driving path state data; perform real-time robot adaptive driving posture analysis based on the real-time rescue driving path state data, and generate real-time robot adaptive driving posture data; An instant control state intelligent optimization module is used to analyze the instant robot adaptive driving speed according to the instant rescue driving path state data and the instant robot adaptive driving posture data, and generate the instant robot adaptive driving speed data; Based on the real-time robot adaptive driving posture data and the real-time robot adaptive driving speed data, the robot performing the rescue emergency intelligent control operation is intelligently optimized in the real-time rescue emergency control state.

2. The robot rescue emergency control system based on voice interaction according to claim 1 is characterized in that: The voice interaction analysis module includes the following functions: Use the robot's built-in voice receiving function to collect control voice signals and generate control voice signals; Use automatic speech recognition technology to perform rescue emergency control demand text analysis on the control demand voice signal to generate rescue emergency control demand text data; Perform semantic parsing on the rescue emergency control demand text data to generate rescue emergency control demand semantic data; According to the rescue emergency control demand semantic data, a structured analysis of the rescue emergency voice control task is performed to generate the rescue emergency voice control task structured data.

3. The robot rescue emergency control system based on voice interaction according to claim 1 is characterized in that: The robot image perception modeling module includes the following functions: Use the built-in panoramic image monitoring device of the robot to perform panoramic image monitoring processing and generate panoramic image data; Use the robot's built-in environmental perception sensors to collect environmental perception data and generate environmental perception data; Use the robot's built-in long-range laser radar device to collect long-range laser radar data and generate long-range laser radar data; Perform long-range laser point cloud data analysis based on long-range laser radar data to generate long-range laser point cloud data; Transmitting long-distance laser point cloud data to panoramic image data for global image point cloud modeling and processing to generate a global image point cloud model; Performing hierarchical distance division processing on the environmental perception data to generate hierarchical environmental perception data; Performing hierarchical environment perception time series feature analysis on the hierarchical environment perception data to generate hierarchical environment perception time series feature data; Perform hierarchical environment perception trend analysis based on hierarchical environment perception time series feature data to generate hierarchical environment perception trend data; The environmental perception data is transmitted to the global image point cloud model for environmental perception data mapping processing to generate a global image perception model; and the global image perception model is collaboratively optimized with a three-dimensional model of multi-level environmental changes through hierarchical environmental perception trend data to generate an optimized global image perception model.

4. The robot rescue emergency control system based on voice interaction according to claim 3 is characterized in that: The transmitting of long-distance laser point cloud data to panoramic image data for global image point cloud modeling processing includes: Perform image matching node analysis based on panoramic image data to generate image matching node data; Perform laser point cloud image matching data analysis based on image matching node data and long-distance laser point cloud data to generate laser point cloud image matching data; Based on the laser point cloud image matching data, the long-distance laser point cloud data is transmitted to the panoramic image data for global image modeling processing to generate a global image point cloud model.

5. The robot rescue emergency control system based on voice interaction according to claim 4 is characterized in that: The image matching node analysis based on the panoramic image data includes: Performing global pixel point blur interference evaluation processing on the panoramic image data to generate global pixel point blur interference data; Perform global image effective area analysis based on global pixel fuzzy interference data to generate global image effective area data; Performing global image effective area gradient feature analysis on global image effective area data to generate global image effective area gradient feature data; Image matching node analysis is performed based on the gradient feature data of the effective area of ​​the global image to generate image matching node data.

6. The robot rescue emergency control system based on voice interaction according to claim 1, characterized in that: The robot rescue emergency intelligent control module includes the following functions: Based on the structured data of rescue emergency voice control tasks and the optimized global image perception model, the rescue emergency automated control decision analysis is performed to generate rescue emergency automated control decisions; Based on the rescue emergency automated control decision, the robot performs rescue emergency intelligent control operations.

7. The robot rescue emergency control system based on voice interaction according to claim 1, characterized in that: The real-time running posture analysis module includes the following functions: Use the robot's built-in short-range laser radar equipment to conduct real-time short-range laser radar monitoring and processing on the robot performing rescue emergency intelligent control operations to generate real-time short-range laser radar data; Performing an immediate rescue driving path status analysis based on the immediate short-range laser radar data to generate immediate rescue driving path status data; Obtain the physical design parameters of the rescue robot; Analyze the controllable parameters of robot motion according to the physical design parameters of the rescue robot, and generate the controllable parameters of robot motion; Analyze the robot simulation operation posture parameters based on the robot motion controllable parameters, and generate the robot simulation operation posture parameters; Conducting robot simulation operation posture limit state analysis on robot simulation operation posture parameters to generate robot simulation operation posture limit state data; Performing feasibility analysis on the state data of the immediate rescue driving path according to the extreme state data of the robot simulation operation posture, so as to obtain the state data of the feasible rescue driving path; According to the feasible rescue driving path state data and the robot simulation operation posture parameters, the real-time robot adaptive driving posture analysis is performed to generate the real-time robot adaptive driving posture data.

8. The robot rescue emergency control system based on voice interaction according to claim 7, characterized in that: The feasibility analysis of the immediate rescue path status data includes: When the immediate rescue driving path status data conforms to the robot simulation operation posture limit state data, the immediate rescue driving path status data is marked as feasible rescue driving path status data; otherwise, the immediate rescue driving path status data is marked as abnormal rescue driving path status data, and the abnormal rescue driving path status data is fed back to the rescue emergency intelligent control operation to execute the rescue emergency driving path intelligent adjustment operation.

9. The robot rescue emergency control system based on voice interaction according to claim 7, characterized in that: The robot simulation operation posture parameters include the robot simulation operation pitch angle, the robot simulation operation roll angle and the robot simulation operation heading angle.

10. The robot rescue emergency control system based on voice interaction according to claim 1, characterized in that: The real-time control state intelligent optimization module includes the following functions: Design the robot driving rate cost function based on the preset robot driving constraint decision and the robot simulation operation posture parameters, and generate the robot driving rate cost function; The instant rescue driving path state data and the instant robot adaptive driving posture data are transmitted to the robot driving speed cost function to perform the instant robot adaptive driving speed analysis, and generate the instant robot adaptive driving speed data; Based on the real-time robot adaptive driving posture data and the real-time robot adaptive driving speed data, the robot performing the rescue emergency intelligent control operation is intelligently optimized in the real-time rescue emergency control state.

Images