Tunnel entrance multifunctional cooperative response processing system and method based on environmental perception

Abstract

The application discloses a kind of based on environmental perception's tunnel portal multifunctional cooperative response processing method and system, system includes multimodal environmental perception module, integrated control unit, multifunctional execution module, communication module, remote platform and auxiliary support module;Method is through multimodal environmental perception module and collects temperature, humidity, smoke concentration, traffic state and three-dimensional point cloud etc., after pre-processing and fusion analysis, accurately identify icing risk, fire outbreak, traffic anomaly etc., based on preset priority rule triggers multifunctional execution module and switches snow-melting agent / fire extinguishing agent spraying mode, synchronous linkage traffic guidance equipment issues warning information, simultaneously realize local autonomous response and remote platform monitoring;It can effectively solve the defects of the prior art function fragmentation, slow response, equipment redundancy, single perception, greatly improve the tunnel portal working condition identification accuracy and emergency disposal efficiency, save roadside space and operation and maintenance cost.

Description

Background Technology

[0001] Tunnels, as a key component of modern highway and urban transportation systems, are widely used in complex terrain environments such as mountainous areas, high-altitude cold regions, and hot and humid areas. The tunnel entrance area, as the transition zone between the natural climate environment and the enclosed tunnel space, has long faced multiple prominent safety hazards due to its unique geographical and climatic conditions, as follows: Winter snow and ice accumulation pose significant risks: near the entrance of the tunnel, the combined effects of wind speed changes, uneven sunlight, and topographic microclimate make it easy for stable "ice zones" or "snow accumulation areas" to form, leading to frequent traffic accidents such as vehicle skidding and loss of steering control, which seriously threatens road traffic safety.

[0002] Delayed fire emergency response: Tunnels are inherently enclosed and narrow in structure. Once a fire or vehicle explosion occurs in the tunnel entrance area, rescue access is restricted. Existing fire protection systems are mostly deployed in the middle and rear sections of the tunnel, which is insufficient to cover the area outside the tunnel entrance. The response speed of ordinary fire extinguishers or manual deployment is difficult to meet the needs of rapid fire suppression in the early stages of a fire, which can easily lead to the spread and expansion of the fire.

[0003] The level of intelligence in traffic guidance is low: existing traffic guidance methods mostly rely on fixed equipment such as LED guidance screens, lane light arrows, and broadcasting systems, and most of them are in a "manually preset" operation state. They lack a linkage mechanism with the real-time environmental conditions and cannot issue personalized and accurate guidance instructions in real time according to emergencies such as freezing or fire in the tunnel entrance area, making it difficult to effectively guide vehicles to avoid risks.

[0004] Current mainstream technologies for addressing the above issues suffer from significant drawbacks such as functional fragmentation and poor synergy, as detailed below: In terms of de-icing and snow melting technology at tunnel entrances, traditional manual methods of spreading salt and sand are cumbersome and cause serious corrosion to the road surface and surrounding environment. Although new electric heating road surface systems and spray-type de-icing fluid devices have a certain degree of automation, they are not linked with weather systems and traffic guidance systems and cannot independently complete the closed-loop processing of "perception-judgment-response", thus limiting their application effectiveness.

[0005] In terms of tunnel fire extinguishing equipment, common high-pressure water guns and dry powder automatic spraying systems are mostly deployed in the middle and rear sections of the tunnel. They have a slow response speed to initial fires in the tunnel entrance area, making it difficult to achieve rapid local suppression and failing to meet the timeliness requirements for emergency response to fires at the tunnel entrance.

[0006] In terms of traffic guidance technology, existing devices lack a linkage mechanism with the real-time environment (such as freezing or fire), resulting in untimely and inaccurate updates to guidance information. Furthermore, they lack effective integration with environmental response equipment, making it difficult to achieve synchronized information dissemination and emergency response when accidents or severe weather occur.

[0007] Furthermore, the aforementioned systems are currently operating independently, exhibiting technical challenges such as modular deployment without unified control, fragmented response mechanisms lacking coordination, and difficulty in adapting to rapid response needs under extreme weather conditions or emergencies. Traditional de-icing methods rely on manual spreading or burying heating cables, resulting in long operation cycles, high operating costs, and susceptibility to terrain limitations. Spray-type de-icing systems have a simple structure, limited to specific functional scenarios, and cannot accurately select different types of treatment agents for precise application based on actual needs. Simultaneously, the control modes of existing systems are mostly independent, information cannot be shared, and the use of multiple devices in a decentralized deployment leads to redundancy, wasted space resources, and complex system maintenance, making them unsuitable for the "high-risk, low-space" application scenarios at tunnel entrances. Even in systems that integrate sensing and response functions, their judgment methods often rely on single-parameter triggers, lacking comprehensive sensing and intelligent judgment of multiple environmental factors such as temperature, humidity, smoke, and wind speed, and their response strategies cannot be dynamically adjusted according to the severity of the event.

[0008] In summary, existing technologies lack an intelligent system solution that can integrate de-icing, fire extinguishing, and guidance functions onto a single platform, conduct collaborative control based on different functional characteristics, and remotely monitor and provide status feedback through a unified management system. This makes it difficult to meet the complex response needs of various emergencies in tunnel entrance areas and effectively ensure road traffic safety and emergency response efficiency. Summary of the Invention

[0009] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a multi-functional collaborative response processing method and system for tunnel entrances based on environmental perception. It integrates basic environmental parameters, traffic status data, and 3D point cloud data through a multi-modal environmental perception module, and combines this with specialized data processing algorithms to achieve multi-dimensional and accurate operational condition identification. It innovatively integrates three core functions—snow melting, fire extinguishing, and traffic guidance—onto a single system platform. It employs an integrated spraying assembly with switchable media and standardized guidance equipment deployment. A scientific operational condition response priority control logic resolves functional conflict issues, while simultaneously constructing a "local autonomous triggering +..." The dual-layer control architecture of "remote platform monitoring" not only significantly reduces the number of devices and roadside space occupation, and lowers procurement, installation and maintenance costs, but also significantly improves the accuracy of working condition identification, the timeliness of emergency response and the targeted nature of handling. It enables the precise delivery of snow melting agents and fire extinguishing agents and the synchronous linkage of guidance information and emergency actions. It effectively solves the defects of existing technologies such as fragmented functions, slow response, equipment redundancy, single perception and poor coordination. It comprehensively improves the traffic safety guarantee capability and emergency response efficiency in the tunnel entrance area, and is suitable for the typical application scenario of "high risk and low space" at the tunnel entrance. It has good engineering practicality and promotion value.

[0010] To achieve the above objectives, one aspect of the present invention provides a multi-functional collaborative response processing system for tunnel entrances based on environmental perception, including a multi-modal environmental perception module, an integrated control unit, a multi-functional execution module, a communication module, a remote platform, and an auxiliary support module. The modules work together to achieve a complete closed-loop function of "perception-decision-execution-monitoring". The multimodal environmental perception module includes a basic environmental sensing unit, a traffic and visual perception unit, and a 3D point cloud perception unit. The basic environmental sensing unit is used to collect temperature, humidity, smoke concentration, and wind speed, and to identify icing risks and fire hazards. The traffic and visual perception unit includes a video image acquisition unit, which is used to identify abnormal traffic conditions such as vehicle congestion and reduced visibility. The 3D point cloud perception unit includes an acquisition device composed of multiple lidar sensors, which is used to acquire 3D point cloud data of the environment and to accurately identify obstacles and road surface features through subsequent processing. The integrated control unit includes a data processing subunit, a working condition identification subunit, a priority control subunit, and a control command generation subunit, which are used to realize data fusion analysis, working condition identification, priority decision-making, and control command generation. The multi-functional execution module includes a spraying execution unit and a traffic guidance unit, which are used to respond to the control commands of the integrated control unit and complete the specific operations of snow melting, fire extinguishing and traffic guidance. The communication module includes a local communication subunit and a remote communication subunit. The local communication subunit is used to transmit the data collected by the environmental sensing module to the integrated control unit in real time and to accurately send the control commands of the integrated control unit to the multi-functional execution module. The remote communication subunit is used to upload the operating condition information and response status data obtained by the integrated control unit to the remote platform, and at the same time receive the control commands from the remote platform to support remote manual control functions.

[0011] Furthermore, the auxiliary support module includes a structural support module and a liquid storage module; the structural support module includes a stable support foundation and a protective shell; the liquid storage module has a built-in de-icing agent storage tank and a fire extinguishing agent storage tank, which store the two media respectively.

[0012] Furthermore, the remote platform includes a status monitoring subunit, a remote control subunit, a data storage and analysis subunit, and a parameter adjustment subunit; The status monitoring subunit is used to display the system's operating status, real-time environmental parameter sensing data, operating condition identification results, and the action status of execution modules in real time; the remote control subunit is used to support manual issuance of control commands to intervene in system operation, including starting / stopping snow melting spraying, starting / stopping fire extinguishing spraying, starting / stopping induction equipment, adjusting spraying parameters, and modifying operating condition identification thresholds; the data storage and analysis subunit is used to record historical event data; the parameter adjustment subunit is used to provide a visual parameter configuration interface to remotely modify temperature thresholds, humidity thresholds, smoke concentration thresholds, temperature rise thresholds, visibility thresholds, spraying flow range, and response priority rules, and to synchronize the modified parameters to the integrated control unit in real time.

[0013] Furthermore, the basic environmental sensing unit includes a temperature sensor, a humidity sensor, a smoke sensor, and an anemometer; The spraying unit includes an integrated spraying assembly, a snow-melting agent storage tank, a fire extinguishing agent storage tank, a delivery pipeline, a booster pump, and a flow control valve. The integrated spraying assembly uses a 360° rotatable universal nozzle to support rapid switching between snow-melting agent and fire extinguishing agent. The booster pump and flow control valve work together to achieve stepless adjustment of the spraying flow rate, and the spraying flow rate, range, and time of each spraying medium can be controlled independently. The traffic guidance unit includes flashing lights, a high-brightness LED guidance screen, and a voice prompt device, which are used to issue warning information, guide vehicles to slow down or divert traffic, and are synchronized with the spraying action.

[0014] A second aspect of the present invention provides a multi-functional collaborative response processing method for tunnel entrances based on environmental perception, implemented using the aforementioned multi-functional collaborative response processing system for tunnel entrances based on environmental perception, comprising the following steps: S1: Environmental parameters of the tunnel entrance area are collected by the multimodal environmental perception module and transmitted to the integrated control unit in real time; the environmental parameters include temperature, humidity, smoke concentration, wind speed and traffic status data; S2: The data processing subunit of the integrated control unit preprocesses and fuses the environmental parameters based on a preset multi-condition judgment logic to identify the current working condition type, which includes icing risk working condition, fire emergency working condition and traffic abnormality guidance working condition. S3: The control command generation subunit of the integrated control unit sends control commands to the multi-functional execution module according to the working condition identification result and the preset response priority control logic, triggering the corresponding functional module to perform response operations; S4: While performing the response operation, the integrated control unit uploads the operating condition type, environmental parameters, and response action details to the remote platform in real time through the communication module. The status monitoring subunit of the remote platform displays relevant information in real time, and the data storage subunit records historical data. Maintenance personnel monitor the response process in real time through the remote platform. If intervention is required, they can issue control commands through the remote control subunit to adjust response parameters or stop the response action, thereby realizing status feedback, remote monitoring, and historical event recording.

[0015] Furthermore, in step S2, the data processing subunit of the integrated control unit performs preprocessing and fusion analysis on the environmental parameters based on a preset multi-condition criterion judgment logic, including sensor data preprocessing, video image preprocessing, three-dimensional point cloud data processing, and data fusion. The preset multi-condition judgment logic includes the following conditions: icing risk conditions: temperature collected by temperature sensor ≤2℃ and humidity collected by humidity sensor ≥85%RH; fire emergency conditions: smoke concentration ≥50ppm or temperature rise ≥15℃ within 1 minute; traffic anomaly guidance conditions: the video image acquisition unit identifies vehicles stationary in the same area for 3 consecutive frames and determines it as vehicle stagnation, or the visibility is determined to be ≤200m based on wind speed, humidity and video image analysis. Sensor data preprocessing includes using the Kalman filter algorithm to filter and denoise temperature, humidity, smoke concentration, and wind speed data, and to remove outliers; Video image preprocessing includes grayscale enhancement, noise reduction, and distortion correction to improve image clarity and support traffic condition recognition. 3D point cloud data processing specifically includes: Voxel rasterization is performed on the 3D point cloud data to obtain simplified point cloud data; The simplified point cloud data is processed to obtain ordered point cloud data, and the ground point cloud data and non-ground point cloud data in the ordered point cloud data are identified and the ground point cloud data is deleted. The non-ground point cloud data is clustered based on the mean-shift clustering algorithm to identify obstacles in the environment; Data fusion involves combining preprocessed sensor data, video analysis results, and 3D point cloud processing results to generate a unified environmental status dataset.

[0016] Furthermore, the three-dimensional point cloud data is subjected to voxel rasterization to obtain simplified point cloud data. The specific process is as follows: Determine the minimum bounding box of the 3D point cloud data; The minimum bounding box is divided into multiple voxel grids according to the preset voxel grid side length; Determine the voxel grid to which each point data belongs in the point cloud data; Determine the centroid of the voxel grid, and use the point data corresponding to the centroid as all point data in the voxel grid, or use the point data closest to the centroid as all point data in the voxel grid, to obtain the simplified point cloud data. The formula for calculating the number of the multiple voxel grids is: ; The number of the plurality of voxel grids; for The number of voxel grids along the axis; for The number of voxel grids along the axis; for The number of voxel grids along the axis; , , The calculation formula is:

[0017] in, for The length of the minimum bounding box in the axial direction; for The length of the minimum bounding box in the axial direction; for The length of the minimum bounding box along the axis; cell is the side length of the voxel grid; , , The calculation formula is as follows:

[0018] in, , The point cloud data are respectively in Maximum and minimum values ​​on the axis; , The point cloud data are respectively in Maximum and minimum values ​​on the axis; , The point cloud data are respectively in Maximum and minimum values ​​on the axis.

[0019] Furthermore, the specific process of ordering the simplified point cloud data to obtain ordered point cloud data is as follows: In the Cartesian coordinate system - The plane is transformed into a circle with an infinite radius, and the circle is divided into multiple sectors based on the radian parameter; Based on the aforementioned sector, the simplified point cloud data is divided into multiple point cloud data sectors; Based on preset distance parameters and preset division conditions, each point cloud data sector in the multiple point cloud data sectors is divided into multiple point cloud data sector rings to obtain ordered point cloud data. The formula for calculating the number of point cloud data sectors is:

[0020]

[0021] in, Point data within a sector of point cloud data; To simplify point cloud data; It is simplified point cloud data The first in Data points; Represents a dotted cloud fan shape; For point data The position within the dotted cloud fan surface; Point The corresponding sector number is ; Points are expressed in radians The arctangent value in a planar coordinate system; For radians; The preset partitioning condition expression is:

[0022] in, The first preset distance parameter; This is the second preset distance parameter; For the first sector of point cloud data The x-coordinate of each data point; For the first sector of point cloud data The vertical coordinates of the data points.

[0023] Furthermore, determining the ground point cloud data and non-ground point cloud data within the ordered point cloud data specifically includes: The ordered point cloud data is dimensionality reduced to obtain dimensionality-reduced point cloud data. Based on preset constraints, linear fitting is performed on the dimensionality-reduced point cloud data to obtain multiple fitted lines. It is then determined whether the distance between each point cloud and the fitted line is greater than a threshold height. Points greater than the threshold height are considered non-ground point cloud data, while those less than or equal to the threshold height are considered ground point cloud data. The expression for the dimensionality-reduced point cloud data is:

[0024] in, The coordinates of the points in the ordered point cloud data; The coordinates of the point cloud in the reduced-dimensional point cloud data; This refers to the angle between the point cloud and the origin in the reduced-dimensional point cloud data.

[0025] Furthermore, the response priority control logic in step S3 is as follows: if multiple working conditions are identified at the same time, only the response action corresponding to the high-priority working condition is triggered; if it is a single working condition, the corresponding response action is triggered directly; that is, when multiple working condition types are identified at the same time, the fire emergency working condition is responded to first, followed by the icing risk working condition, and finally the traffic abnormality guidance working condition. When an icing risk condition is identified, the de-icing agent spraying device is controlled to release de-icing agent; when a fire emergency condition is identified, the fire extinguishing agent spraying device is controlled to release fire extinguishing agent and the traffic guidance module is activated to issue a warning; when an abnormal traffic guidance condition is identified, the traffic guidance module is activated to guide vehicles to slow down or divert. The traffic guidance module in step S3 includes a flashing light, a high-brightness LED guidance screen, and a voice prompt device. The flashing light, the high-brightness LED guidance screen, and the voice prompt device are all connected to the integrated control unit and operate on the network. They execute corresponding warning or guidance operations according to the instructions of the integrated control unit.

[0026] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects: (1) This invention integrates basic environmental parameters such as temperature, humidity, smoke concentration, and wind speed, video image traffic status data, and three-dimensional point cloud terrain obstacle data through a multimodal environmental perception module. Combined with specialized data processing algorithms such as voxel rasterization simplification, ordered division, and mean drift clustering, it achieves multi-dimensional and accurate perception of tunnel entrance icing risk, sudden fire, traffic anomalies, and obstacles. The accuracy of working condition identification is significantly improved, effectively solving the problems of low reliability and large identification deviation caused by the reliance on single parameter perception in existing technologies.

[0027] (2) This invention innovatively integrates three core functions—snow melting agent spraying, fire extinguishing agent spraying, and traffic guidance—into the same system platform. It uses an integrated spraying assembly with switchable media and standardized interfaces to deploy guidance equipment. Compared with the traditional decentralized deployment scheme, it significantly reduces the number of equipment and the space occupied by roadside deployment, reduces the cost of equipment procurement, installation, and subsequent operation and maintenance, and simplifies the maintenance process. It is perfectly suited to the typical application scenario of "high risk and low space" at the tunnel entrance.

[0028] (3) This invention establishes a scientific working condition response priority control logic, which clarifies that the working condition of sudden fire is given priority over the working condition of icing risk and the working condition of icing risk is given priority over the working condition of abnormal traffic. When multiple working conditions occur at the same time, key functional modules can be scheduled to perform response actions first, avoiding conflicts and interference between different functions, ensuring the effectiveness and pertinence of emergency response, significantly improving the timeliness of emergency response such as initial fire fighting and road snow melting and de-icing, and solving the technical pain points of existing systems being independent and lacking coordination.

[0029] (4) The present invention adopts a two-layer control architecture of “local autonomous triggering + remote platform monitoring”. The system can respond quickly and autonomously when the environment changes. At the same time, it supports real-time status monitoring, control command issuance, parameter dynamic adjustment and historical data storage and analysis through the remote platform. This not only ensures the ability to deal with emergencies quickly, but also provides data support for optimizing response strategies and planning maintenance cycles in the later stage, improves the intelligence level and flexibility of system operation and maintenance, and reduces long-term operating costs.

[0030] (5) The spraying execution unit of the present invention can independently control the spraying flow rate, range and time of the snow melting agent and fire extinguishing agent according to different working conditions and environmental parameters, so as to achieve precise delivery of the medium and avoid the problems of resource waste and poor effect in traditional de-icing and fire extinguishing methods. At the same time, the traffic guidance module is linked with the spraying action to ensure that the information release and emergency response actions are coordinated and consistent, effectively guide vehicles to avoid risks, and comprehensively improve the traffic safety guarantee capability and emergency response efficiency of the tunnel entrance area. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the structure of a multi-functional collaborative response processing system for tunnel entrances based on environmental perception, according to an embodiment of the present invention. Figure 2 This is a schematic diagram illustrating a multi-functional collaborative response processing method for tunnel entrances based on environmental perception, according to an embodiment of the present invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0033] like Figure 1As shown, one aspect of the present invention provides a multi-functional collaborative response processing system for tunnel entrances based on environmental perception. It adopts a five-layer modular integrated design consisting of a perception layer, a control layer, an execution layer, a communication layer, and a management layer. The system is deployed in the guardrails, ditches, or slope areas on both sides of the tunnel. The overall structure is compact, including a multi-modal environmental perception module, an integrated control unit, a multi-functional execution module, a communication module, a remote platform, and an auxiliary support module. These modules work together to achieve a complete closed-loop function of "perception-decision-execution-monitoring." The present invention enables multi-dimensional and accurate perception of the tunnel entrance environment, collaborative response to multiple emergency functions, and dual local and remote control. It solves problems such as functional fragmentation, slow response, equipment redundancy, and limited perception in existing technologies, thereby improving traffic safety and emergency response efficiency in the tunnel entrance area.

[0034] Furthermore, the multimodal environmental perception module serves as the perception layer, acting as the system's "perception terminal." It is responsible for comprehensively collecting environmental and traffic status data at the tunnel entrance, providing a basis for subsequent decision-making. This module includes a basic environmental sensing unit, a traffic and visual perception unit, and a 3D point cloud perception unit. The basic environmental sensing unit includes temperature sensors, humidity sensors, smoke sensors, and an anemometer, used to collect key environmental parameters such as temperature, humidity, smoke concentration, and wind speed, and to identify icing risks and fire hazards. The traffic and visual perception unit includes a video image acquisition unit, used to identify abnormal traffic conditions such as vehicle congestion and reduced visibility. The 3D point cloud perception unit includes a collection device composed of multiple lidar sensors, used to acquire 3D point cloud data of the environment, and through subsequent processing, accurately identify obstacles and road surface features. Furthermore, the temperature sensor has a measurement range of -40℃ to 85℃ and a measurement accuracy of ±0.5℃; the humidity sensor has a measurement range of 0 to 100%RH and a measurement accuracy of ±3%RH; the smoke sensor adopts a photoelectric detection principle with a response time ≤10s and can detect smoke concentrations ranging from 0 to 1000ppm; the anemometer adopts an ultrasonic type with a measurement range of 0 to 60m / s and a measurement accuracy of ±0.1m / s. All of the above sensors are connected to the integrated control unit through standardized interfaces and output environmental parameter electrical signals in real time. The traffic and visual perception unit adopts a high-definition network camera video image acquisition unit with a resolution of not less than 1920×1080, a frame rate of ≥25fps, low-light imaging capability (minimum illumination ≤0.01 lux), supports real-time video stream transmission and image analysis, and can identify abnormal traffic conditions such as vehicle congestion, abnormal vehicle speed, and reduced visibility. The three-dimensional point cloud perception unit consists of at least two lidars. The lidars have a ranging range of 0.5 to 100 m, a ranging accuracy of ±2 cm, and a point cloud density of ≥100,000 points / second. The multiple lidars are symmetrically distributed and, through time synchronization and spatial calibration technology, acquire three-dimensional point cloud data of the tunnel entrance area for accurate identification of road obstacles, terrain undulations, and vehicle outline information.

[0035] Furthermore, the integrated control unit serves as the control layer, acting as the core decision-making and control module. It employs an industrial-grade embedded controller (CPU clock speed ≥ 1.8GHz, memory ≥ 4GB, storage capacity ≥ 64GB) and incorporates dedicated control software. This software enables data fusion analysis, condition identification, priority decision-making, and control command generation. It includes a data processing subunit, a condition identification subunit, a priority control subunit, and a control command generation subunit. The data processing subunit preprocesses the parameters (including 3D point cloud data) collected by the multimodal sensing module, including sensor data filtering and noise reduction, video image enhancement, and 3D point cloud data processing (voxel rasterization simplification, ordered partitioning, ground point cloud removal, non-ground point cloud clustering, and mean-shift clustering), filtering effective data and converting it to a unified format. The condition identification subunit uses… Based on preset multi-condition judgment logic, the system comprehensively analyzes the preprocessed fused data to accurately identify the current operating condition type, including icing risk conditions, fire emergency conditions, and traffic anomaly guidance conditions. The priority control subunit is used to establish operating condition response priority rules (priority order: fire emergency conditions > icing risk conditions > traffic anomaly guidance conditions). When multiple operating conditions are identified simultaneously, the functional modules corresponding to the higher priority conditions are prioritized to execute response actions to avoid functional conflicts. The control command generation subunit is used to generate precise control commands based on the operating condition identification results and priority decisions, including spray medium switching commands, spray flow adjustment commands, spray range control commands, and guidance equipment start commands, with command transmission latency ≤100ms. It supports three modes: timed operation, autonomous triggering, and remote manual control.

[0036] Furthermore, the specific criteria for determining the icing risk condition, sudden fire condition, and traffic anomaly guidance condition are as follows: Icing risk conditions: Temperature collected by temperature sensor ≤2℃ (threshold can be adjusted remotely) and humidity collected by humidity sensor ≥85%RH (threshold can be adjusted remotely). Emergency fire conditions: Smoke concentration ≥ 50 ppm (threshold can be remotely adjusted) or temperature rise ≥ 15℃ within 1 minute (threshold can be remotely adjusted). Traffic anomaly guidance conditions: The video image acquisition unit identifies vehicles stationary in the same area for 3 consecutive frames (determined as vehicle stagnation), or based on wind speed, humidity and video image analysis, the visibility is determined to be ≤200m (the threshold can be adjusted remotely). Furthermore, the multi-functional execution module serves as the execution layer, responsible for responding to control commands from the integrated control unit and completing specific operations such as snow melting, fire extinguishing, and traffic guidance. It includes a spraying execution unit and a traffic guidance unit. The spraying execution unit comprises an integrated spraying assembly, a snow melting agent storage tank (capacity ≥ 500L), a fire extinguishing agent storage tank (capacity ≥ 300L), a delivery pipeline, a booster pump (working pressure ≥ 0.8MPa), and a flow control valve. The integrated spraying assembly uses a 360° rotatable universal nozzle (rotation angle accuracy ±1°), with a nozzle range of 5–20m, supporting rapid switching between snow melting agent and fire extinguishing agent (switching time ≤ 2s). The booster pump works in conjunction with the flow control valve to achieve… The spray flow rate is infinitely adjustable (adjustment range 5-50 L / min), and the spray flow rate, range, and time for each spraying medium can be independently controlled. The spraying execution unit of this invention integrates a snow-melting agent spraying device and a fire extinguishing agent spraying device, sharing the same rotatable / directional spraying assembly. It supports switching between the two media and can independently control the spray flow rate, range, and time according to the instructions of the integrated control unit. It is connected to an internal storage tank via pipeline. The traffic guidance unit includes flashing lights, a high-brightness LED guidance screen, and a voice prompt device, used to issue warning information, guide vehicles to slow down or divert traffic, and synchronize with the spraying action. The flashing light has a flashing frequency of 5-20 Hz and a brightness ≥10000 cd / m². 2 It supports switching between red, yellow, and blue colors; the LED guidance screen has a display resolution of ≥320×160 and a brightness of ≥5000cd / m². 2 It can display text prompts (such as "Ice ahead, slow down" and "Fire ahead, do not pass") and graphic symbols (arrows, warning symbols); the voice prompt device has an output power of ≥30W and a voice clarity of ≥85dB, and can play preset voice warning messages (such as "Caution: Icy road surface, drive carefully" and "Fire at tunnel entrance, please detour immediately").

[0037] Furthermore, the communication module, as a communication layer, adopts a "wired + wireless" dual communication mode to ensure the stability and reliability of data transmission. It enables data interaction between local modules and information exchange with a remote platform, including a local communication subunit and a remote communication subunit. The local communication subunit connects the perception layer, control layer, and execution layer, ensuring that data collected by the environmental perception module is transmitted to the integrated control unit in real time, and that control commands from the integrated control unit are accurately issued to the multi-functional execution module. The local communication subunit uses a CAN bus (communication rate ≥ 500kbps) to connect the integrated control unit with the multi-modal environmental perception module and the multi-functional execution module. The integrated control unit enables real-time data interaction between local modules. The remote communication subunit is used to upload the operating condition information and response status data obtained by the integrated control unit to the remote platform, and at the same time receive control commands from the remote platform to support remote manual control functions. The remote communication subunit integrates a 4G / 5G communication module (supporting full network compatibility) and an Ethernet interface. The downlink rate of the 4G / 5G communication module is ≥100Mbps and the uplink rate is ≥50Mbps. The Ethernet interface supports the 1000M Ethernet protocol to realize bidirectional data transmission between the integrated control unit and the remote platform, including environmental data uploading, operating condition information reporting, control command reception, and parameter configuration synchronization.

[0038] Furthermore, the remote platform serves as the management layer, acting as the system's remote management terminal. Built on a B / S architecture, it supports access from multiple terminals (computers, mobile phones, tablets) to achieve remote monitoring, operation and maintenance management, and data analysis. It includes a status monitoring subunit, a remote control subunit, a data storage and analysis subunit, and a parameter adjustment subunit. The status monitoring subunit displays the real-time system operating status (normal, fault, offline), real-time environmental parameter sensing data, operating condition identification results, and the action status of execution modules. The remote control subunit supports manual issuance of control commands to intervene in system operation, including starting / stopping snow melting spraying, starting / stopping fire extinguishing spraying, starting / stopping induction equipment, adjusting spraying parameters (flow rate, range, time), and modifying... The system includes features such as operating condition identification thresholds and remote control command response latency of ≤1s. The data storage and analysis subunit records historical event data to provide a basis for subsequent response strategy optimization. Specifically, it employs cloud storage technology to store historical environmental data, operating condition event records, and execution action logs (storage duration ≥1 year). It incorporates built-in data statistical analysis algorithms to generate monthly / quarterly / annual statistical reports on operating condition frequency, system runtime, and consumable consumption (snow melting agent, fire extinguishing agent, etc.). The parameter adjustment subunit provides a visual parameter configuration interface, allowing remote modification of key parameters such as temperature threshold, humidity threshold, smoke concentration threshold, temperature rise threshold, visibility threshold, spray flow range, and response priority rules. The modified parameters are then synchronized to the integrated control unit in real time.

[0039] Furthermore, the auxiliary support module includes a structural support module and a liquid storage module. The structural support module is made of stainless steel (rust-proof grade ≥304) and includes a stable support foundation and a protective shell, used to fix all functional modules, adapting to the deployment scenarios of guardrails, ditches, or slopes on both sides of the tunnel, ensuring equipment stability. The support foundation is fixed to the ground with expansion bolts (fixing strength ≥50kN). The protective shell has a protection grade ≥IP65, with dustproof, waterproof, corrosion-proof, and impact-resistant functions, adapting to the complex outdoor environment of the tunnel entrance. The liquid storage module has a built-in de-icing agent storage tank and a fire extinguishing agent storage tank, storing two media respectively, providing material supply for the spraying execution unit. During operation, the multimodal environmental perception module collects data and transmits it to the integrated control unit via the communication module. The integrated control unit analyzes and identifies the operating conditions and determines the priorities, then sends instructions to the multi-functional execution module in the execution layer. The spraying execution unit and traffic guidance unit of the multi-functional execution module perform actions to complete specific operations such as snow melting, fire extinguishing, and traffic guidance, while simultaneously transmitting the status back to the remote platform via the communication module. Monitoring and maintenance are then achieved through the remote platform.

[0040] This invention integrates the three core safety functions required in the tunnel entrance area—snow melting, fire extinguishing, and guidance—into a single intelligent system. Through multimodal perception fusion, priority decision-making, and local and remote collaborative control, it breaks through the technical barriers of existing systems, such as functional isolation, redundant deployment, and slow response.

[0041] like Figure 2 As shown, a second aspect of the present invention provides a multi-functional collaborative response processing method for tunnel entrances based on environmental perception, comprising the following steps: S1: Environmental parameters of the tunnel entrance area are collected by the multimodal environmental perception module and transmitted to the integrated control unit in real time; the environmental parameters include temperature, humidity, smoke concentration, wind speed and traffic status data; S2: The data processing subunit of the integrated control unit preprocesses and fuses the environmental parameters based on a preset multi-condition judgment logic to identify the current working condition type, which includes icing risk working condition, fire emergency working condition and traffic abnormality guidance working condition. S3: The control command generation subunit of the integrated control unit sends control commands to the multi-functional execution module based on the working condition identification result and the preset response priority control logic, triggering the corresponding functional module to perform response operations; when the working condition is identified as an icing risk, the de-icing agent spraying device is controlled to release de-icing agent; when the working condition is identified as a fire emergency, the fire extinguishing agent spraying device is controlled to release fire extinguishing agent and the traffic guidance module is linked to issue a warning; when the working condition is identified as an abnormal traffic guidance, the traffic guidance module is activated to guide vehicles to slow down or divert. S4: While executing the response operation, the integrated control unit uploads the operating condition type, environmental parameters, and response action details (spraying mode, flow rate, range, time, and induction equipment operating status) to the remote platform in real time through the communication module. The status monitoring subunit of the remote platform displays relevant information in real time, and the data storage subunit records historical data. Maintenance personnel monitor the response process in real time through the remote platform. If intervention is required, they can issue control commands through the remote control subunit to adjust response parameters or stop the response action, thereby realizing status feedback, remote monitoring, and historical event recording.

[0042] Furthermore, the step S1, which involves collecting environmental parameters of the tunnel entrance area through the multimodal environmental perception module and transmitting them to the integrated control unit in real time, specifically includes the following steps: The basic environmental sensing unit collects temperature, humidity, smoke concentration, and wind speed data in real time, for example, at a collection frequency of 1 time / second, and transmits the collected data to the integrated control unit via the CAN bus. The traffic and visual perception unit continuously captures video images of the tunnel entrance area at a video stream rate of 25 frames per second. Simultaneously, it analyzes the video images in real time to extract information such as vehicle position, speed, and visibility, and transmits it to the integrated control unit. The 3D point cloud sensing unit synchronously collects 3D point cloud data of the environment. For example, the collection frequency is 10 times / second. The initial point cloud data of multiple lidars are processed through time synchronization (based on PTP protocol) and spatial calibration (based on calibration board calibration) to generate 3D point cloud data in a unified coordinate system, which is then transmitted to the integrated control unit.

[0043] Furthermore, step S1 also includes activating the multi-functional collaborative response processing system at the tunnel entrance based on environmental perception. The integrated control unit performs self-tests on each module (including sensor communication status, actuator action status, communication link connectivity status, and power supply status). After the self-test passes, the system enters standby mode, and the remote platform displays "System is running normally". If the self-test detects a module fault, the integrated control unit sends a fault alarm message (including the faulty module name, fault type, and fault occurrence time) to the remote platform through the communication module, and displays a fault prompt on the LED guidance screen.

[0044] Further, the multimodal environment perception module in step S1 includes a temperature sensor, a smoke sensor, a humidity sensor, an anemometer, and a video image acquisition unit; the traffic status data is acquired through the video image acquisition unit, and the multimodal environment perception module also includes a three-dimensional point cloud data acquisition unit based on lidar; the specific process of the three-dimensional point cloud data acquisition unit acquiring environmental three-dimensional point cloud data includes: acquiring multiple initial point cloud data of the environment based on multiple lidars; preprocessing the multiple initial point cloud data to generate the three-dimensional point cloud data; Furthermore, in step S2, the data processing subunit of the integrated control unit performs preprocessing and fusion analysis on the environmental parameters based on a preset multi-condition criterion judgment logic, including sensor data preprocessing, video image preprocessing, three-dimensional point cloud data processing, and data fusion. Sensor data preprocessing includes using the Kalman filter algorithm to filter and denoise temperature, humidity, smoke concentration, and wind speed data, and to remove outliers (such as data that are outside the sensor's measurement range or data that suddenly change). Video image preprocessing includes grayscale enhancement, noise reduction, and distortion correction to improve image clarity and support traffic condition recognition. 3D point cloud data processing specifically includes: The three-dimensional point cloud data is subjected to voxel rasterization to obtain simplified point cloud data; The simplified point cloud data is processed to obtain ordered point cloud data, and the ground point cloud data and non-ground point cloud data in the ordered point cloud data are identified and the ground point cloud data is deleted. The non-ground point cloud data is clustered based on the mean-shift clustering algorithm to identify obstacles in the environment; Data fusion involves combining preprocessed sensor data, video analysis results, and 3D point cloud processing results to generate a unified environmental status dataset.

[0045] Furthermore, the three-dimensional point cloud data is subjected to voxel rasterization to obtain simplified point cloud data. The specific process is as follows: Determine the minimum bounding box of the 3D point cloud data ( Axis range ~ , Axis range , Axis range ~ ); The minimum bounding box is divided into multiple voxel grids according to the preset voxel grid side length; Determine the voxel grid to which each point data belongs in the point cloud data; The centroid of the voxel grid is determined, and the point data corresponding to the centroid is used as all point data in the voxel grid, or the point data closest to the centroid is used as all point data in the voxel grid, to obtain the simplified point cloud data and reduce the amount of data processing. In one specific embodiment of the present invention, the number of the plurality of voxel grids is calculated as follows: ; The number of the plurality of voxel grids; for The number of voxel grids along the axis; for The number of voxel grids along the axis; for The number of voxel grids along the axis; , , The calculation formula is:

[0046] in, for The length of the minimum bounding box in the axial direction; for The length of the minimum bounding box in the axial direction; for The length of the minimum bounding box along the axis; cell is the side length of the voxel grid; , , The calculation formula is as follows:

[0047] , The point cloud data are respectively in Maximum and minimum values ​​on the axis; , The point cloud data are respectively in Maximum and minimum values ​​on the axis; , The point cloud data are respectively in Maximum and minimum values ​​on the axis; Furthermore, the specific process of ordering the simplified point cloud data to obtain ordered point cloud data is as follows: In the Cartesian coordinate system - The plane is transformed into a circle with an infinite radius, and the circle is divided into multiple sectors based on the radian parameter; Based on the aforementioned sector, the simplified point cloud data is divided into multiple point cloud data sectors; Based on preset distance parameters and preset division conditions, each point cloud data sector in the multiple point cloud data sectors is divided into multiple point cloud data sector rings to obtain ordered point cloud data. Furthermore, the formula for calculating the number of point cloud data sectors is as follows:

[0048]

[0049] in, Point data within a sector of point cloud data; To simplify point cloud data; It is simplified point cloud data The first in Data points; Represents a dotted cloud fan shape; For point data The position within the dotted cloud fan surface; Point The corresponding sector number is ; Points are expressed in radians The arctangent value in a planar coordinate system; This is a radian parameter used to divide a plane into multiple sectors according to angles; The preset partitioning condition expression is:

[0050] in, The first preset distance parameter; This is the second preset distance parameter; For the first sector of point cloud data The x-coordinate of each data point; For the first sector of point cloud data The vertical coordinate of each data point; Furthermore, determining the ground point cloud data and non-ground point cloud data within the ordered point cloud data specifically includes: The ordered point cloud data is dimensionality reduced to obtain dimensionality-reduced point cloud data. Based on preset constraints, linear fitting is performed on the dimensionality-reduced point cloud data to obtain multiple fitted lines. Based on the fitted lines and threshold height, ground point cloud data and non-ground point cloud data in the dimensionality-reduced point cloud data are determined. The expression for the dimensionality-reduced point cloud data is:

[0051] in, The coordinates of the points in the ordered point cloud data; The coordinates of the point cloud in the reduced-dimensional point cloud data; The angle between the point cloud and the origin in the reduced-dimensional point cloud data; The process of determining the ground point cloud data and non-ground point cloud data in the dimensionality-reduced point cloud data based on the fitted straight line and the threshold height includes the following steps: Determine whether the distance between each point cloud and the fitted line is greater than the threshold height. Points greater than the threshold height are non-ground point cloud data, while those less than or equal to the threshold height are ground point cloud data. Furthermore, the clustering of the non-ground point cloud data based on the mean-drift clustering algorithm to determine obstacles in the environment includes: clustering the non-ground point cloud data based on the mean-drift clustering algorithm, setting a bandwidth parameter (e.g., bandwidth = 0.5m, which can be adjusted), obtaining cluster centers through iterative calculation, and determining obstacles in the environment (e.g., stranded vehicles, scattered objects, etc.) based on the clustering results. Furthermore, the identification conditions for the icing risk condition in step S2 are: the temperature collected by the temperature sensor is lower than the preset temperature threshold, and the humidity collected by the humidity sensor is higher than the preset humidity threshold; the identification conditions for the fire emergency condition are: the smoke concentration collected by the smoke sensor is abnormally high, or the temperature collected by the temperature sensor shows a sharp temperature rise; the identification conditions for the traffic anomaly guidance condition are: the video image acquisition unit identifies vehicle stagnation, or the visibility is determined to be lower than the preset visibility threshold based on environmental parameters. In a specific embodiment of the present invention, if the temperature is ≤2℃ and the humidity is ≥85%RH, it is determined to be a working condition with a risk of icing. If the smoke concentration is ≥50ppm or the temperature rise is ≥15℃ within 1 minute, it is determined to be a fire emergency. If a vehicle is detected to be stuck or visibility is ≤200m, it is determined to be a traffic abnormality guidance condition; If any of the above criteria are not met, the system is considered to be in normal operating condition and remains in standby mode. Furthermore, the integrated control unit mentioned in steps S2 and S3 supports three control modes: timed operation, autonomous triggering, and remote manual control. The remote manual control is achieved by sending control commands through the remote platform. The remote platform can store, analyze, and adjust parameters for operating conditions and response status data. The response priority control logic in step S3 is as follows: if multiple working conditions are identified at the same time, only the response action corresponding to the high-priority working condition is triggered; if it is a single working condition, the corresponding response action is triggered directly; that is, when multiple working condition types are identified at the same time, the fire emergency working condition is responded to first, followed by the icing risk working condition, and finally the traffic abnormality guidance working condition. When the identification result indicates an icing risk condition response: the control spraying execution unit switches to the de-icing agent spraying mode, and automatically adjusts the spraying flow rate (5-50L / min), spraying range (5-20m), and spraying time (1-30 minutes, adjustable) based on temperature, humidity, and wind speed data. The strobe lights are activated (flashing yellow, frequency 10Hz), the LED guidance screen displays "Ice ahead, slow down," and the voice prompt device plays "Caution: Icy road surface, drive carefully" in a loop. When the identification result is a fire emergency response: the control spraying execution unit switches to the fire extinguishing agent spraying mode, and adjusts the spraying flow rate (20-50L / min) and spraying range (5-15m) according to the smoke concentration, temperature rise rate and the location of the fire area identified by the three-dimensional point cloud, and accurately sprays the fire extinguishing agent towards the fire area. At the same time, the strobe lights are activated (red flashing, frequency 20Hz), the LED guidance screen displays "Fire ahead, no passage allowed", the voice prompt device plays "Fire at the tunnel entrance, please detour immediately", and the lane indicator lights at the tunnel entrance are turned off in conjunction with the operation. When the identification result is a traffic anomaly guidance condition response: the flashing lights are activated (blue flashing, frequency 15Hz), the LED guidance screen displays "Low visibility, slow down" or "Congestion ahead, please detour" (switching according to the specific anomaly type), and the voice prompt device plays the corresponding warning voice to guide vehicles to slow down or divert. The traffic guidance module in step S3 includes a flashing light, a high-brightness LED guidance screen, and a voice prompt device. The flashing light, the high-brightness LED guidance screen, and the voice prompt device are all connected to the integrated control unit and operate on the network. They execute corresponding warning or guidance operations according to the instructions of the integrated control unit.

[0052] Furthermore, step S4 also includes: when the data collected by the multimodal environmental perception module shows that the working condition release conditions are met (icing risk condition: temperature > 5℃ and humidity < 70%RH; fire emergency condition: smoke concentration < 10ppm and stable temperature; traffic anomaly guidance condition: vehicle congestion resolved and visibility > 500m), the integrated control unit automatically issues a termination command, the multi-functional execution module stops responding, and the system resets to standby mode; if the working condition release conditions are not met, the system continues to execute response actions until the preset maximum response time is reached (maximum snow melting spraying time = 120 minutes, maximum fire extinguishing spraying time = 60 minutes, which can be remotely adjusted), then automatically stops and alarms prompting manual intervention.

[0053] A third aspect of the present invention also provides an environment sensing device based on point cloud data, comprising: A 3D point cloud data acquisition unit is used to acquire 3D point cloud data of the environment. A simplified point cloud element is used to perform voxel rasterization processing on the three-dimensional point cloud data to obtain simplified point cloud data. The point cloud segmentation unit is used to process the simplified point cloud data into ordered point cloud data, and to determine the ground point cloud data and non-ground point cloud data in the ordered point cloud data, and delete the ground point cloud data. A point cloud clustering unit is used to cluster the non-ground point cloud data based on the mean-shift clustering algorithm to identify obstacles in the environment.

[0054] The environmental perception device based on point cloud data provided by this invention obtains simplified point cloud data by voxel rasterization of 3D point cloud data. This reduces point cloud density while maintaining the basic dimensional characteristics of the point cloud data, thereby reducing the time required for environmental perception based on point cloud data. Furthermore, by ordering the simplified point cloud data to obtain ordered point cloud data, and then identifying ground and non-ground point cloud data within the ordered point cloud data, this invention avoids the poor robustness of environmental perception caused by the disorder of point cloud data, thus improving the reliability of environmental perception. Finally, by clustering the non-ground point cloud data using a mean-shift algorithm to identify obstacles in the environment, this invention addresses the technical problem of poor clustering results for distant point cloud data due to the near-dense and far-sparse characteristics of point cloud data, further improving the robustness of the point cloud data-based environmental perception method and thus further enhancing the reliability of environmental perception.

[0055] This application also provides a computer-readable storage medium storing a computer program that is executed by a processor to implement... Figure 2 The methods provided in each step are detailed in the implementation methods provided in the above steps, and will not be repeated here.

[0056] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A multi-functional collaborative response processing system for tunnel entrances based on environmental perception, characterized in that: It includes a multimodal environment perception module, an integrated control unit, a multi-functional execution module, a communication module, a remote platform, and an auxiliary support module. These modules work together to achieve a complete closed-loop function of "perception-decision-execution-monitoring". The multimodal environmental perception module includes a basic environmental sensing unit, a traffic and visual perception unit, and a 3D point cloud perception unit. The basic environmental sensing unit is used to collect temperature, humidity, smoke concentration, and wind speed, and to identify icing risks and fire hazards. The traffic and visual perception unit includes a video image acquisition unit, which is used to identify abnormal traffic conditions such as vehicle congestion and reduced visibility. The 3D point cloud perception unit includes an acquisition device composed of multiple lidar sensors, which is used to acquire 3D point cloud data of the environment and to accurately identify obstacles and road surface features through subsequent processing. The integrated control unit includes a data processing subunit, a working condition identification subunit, a priority control subunit, and a control command generation subunit, used to realize data fusion analysis, working condition identification, priority decision-making, and control command generation; the multi-functional execution module includes a spraying execution unit and a traffic guidance unit, used to respond to the control commands of the integrated control unit and complete the specific operations of snow melting, fire extinguishing, and traffic guidance. The communication module includes a local communication subunit and a remote communication subunit. The local communication subunit is used to transmit the data collected by the environmental sensing module to the integrated control unit in real time and to accurately send the control commands of the integrated control unit to the multi-functional execution module. The remote communication subunit is used to upload the operating condition information and response status data obtained by the integrated control unit to the remote platform, and at the same time receive the control commands from the remote platform to support remote manual control functions.

2. The multi-functional collaborative response processing method for tunnel entrances based on environmental perception according to claim 1, characterized in that: The auxiliary support module includes a structural support module and a liquid storage module; the structural support module includes a stable support foundation and a protective shell; the liquid storage module has a built-in de-icing agent storage tank and a fire extinguishing agent storage tank, which store the two media respectively.

3. The multi-functional collaborative response processing system for tunnel entrances based on environmental perception according to claim 2, characterized in that: The remote platform includes a status monitoring subunit, a remote control subunit, a data storage and analysis subunit, and a parameter adjustment subunit; The status monitoring subunit is used to display the system's operating status, real-time environmental parameter sensing data, operating condition identification results, and the action status of the execution module in real time; the remote control subunit is used to support the manual issuance of control commands to intervene in the system's operation, including starting / stopping snow melting spraying, starting / stopping fire extinguishing spraying, starting / stopping induction equipment, adjusting spraying parameters, and modifying operating condition identification thresholds. The data storage and analysis subunit is used to record historical event data; the parameter adjustment subunit is used to provide a visual parameter configuration interface to remotely modify temperature threshold, humidity threshold, smoke concentration threshold, temperature rise threshold, visibility threshold, spray flow range, response priority rules, and synchronize the modified parameters to the integrated control unit in real time.

4. The multi-functional collaborative response processing system for tunnel entrances based on environmental perception as described in claim 3, characterized in that: The basic environmental sensing unit includes a temperature sensor, a humidity sensor, a smoke sensor, and an anemometer; The spraying unit includes an integrated spraying assembly, a snow-melting agent storage tank, a fire extinguishing agent storage tank, a delivery pipeline, a booster pump, and a flow control valve. The integrated spraying assembly uses a 360° rotatable universal nozzle to support rapid switching between snow-melting agent and fire extinguishing agent. The booster pump and flow control valve work together to achieve stepless adjustment of the spraying flow rate, and the spraying flow rate, range, and time of each spraying medium can be controlled independently. The traffic guidance unit includes flashing lights, a high-brightness LED guidance screen, and a voice prompt device, which are used to issue warning information, guide vehicles to slow down or divert traffic, and are synchronized with the spraying action.

5. A multi-functional collaborative response processing method for tunnel entrances based on environmental perception, characterized in that, The application of the environmentally aware multi-functional collaborative response processing system for tunnel entrances as described in any one of claims 1-4 includes the following steps: S1: Environmental parameters of the tunnel entrance area are collected by the multimodal environmental perception module and transmitted to the integrated control unit in real time; the environmental parameters include temperature, humidity, smoke concentration, wind speed and traffic status data; S2: The data processing subunit of the integrated control unit preprocesses and fuses the environmental parameters based on a preset multi-condition judgment logic to identify the current working condition type, which includes icing risk working condition, fire emergency working condition and traffic abnormality guidance working condition. S3: The control command generation subunit of the integrated control unit sends control commands to the multi-functional execution module according to the working condition identification result and the preset response priority control logic, triggering the corresponding functional module to perform response operations; S4: While performing the response operation, the integrated control unit uploads the operating condition type, environmental parameters, and response action details to the remote platform in real time through the communication module. The status monitoring subunit of the remote platform displays relevant information in real time, and the data storage subunit records historical data. Maintenance personnel monitor the response process in real time through the remote platform. If intervention is required, they can issue control commands through the remote control subunit to adjust response parameters or stop the response action, thereby realizing status feedback, remote monitoring, and historical event recording.

6. The multi-functional collaborative response processing method for tunnel entrances based on environmental perception according to claim 5, characterized in that, The data processing subunit of the integrated control unit in step S2 performs preprocessing and fusion analysis on the environmental parameters based on a preset multi-condition criterion judgment logic, including sensor data preprocessing, video image preprocessing, three-dimensional point cloud data processing and data fusion. The preset multi-condition judgment logic includes the following conditions: icing risk conditions: temperature collected by temperature sensor ≤2℃ and humidity collected by humidity sensor ≥85%RH; fire emergency conditions: smoke concentration ≥50ppm or temperature rise ≥15℃ within 1 minute; traffic anomaly guidance conditions: the video image acquisition unit identifies vehicles stationary in the same area for 3 consecutive frames and determines it as vehicle stagnation, or the visibility is determined to be ≤200m based on wind speed, humidity and video image analysis. Sensor data preprocessing includes using the Kalman filter algorithm to filter and denoise temperature, humidity, smoke concentration, and wind speed data, and to remove outliers; Video image preprocessing includes grayscale enhancement, noise reduction, and distortion correction to improve image clarity and support traffic condition recognition. 3D point cloud data processing specifically includes: Voxel rasterization is performed on the 3D point cloud data to obtain simplified point cloud data; The simplified point cloud data is processed to obtain ordered point cloud data, and the ground point cloud data and non-ground point cloud data in the ordered point cloud data are identified and the ground point cloud data is deleted. The non-ground point cloud data is clustered based on the mean-shift clustering algorithm to identify obstacles in the environment; Data fusion involves combining preprocessed sensor data, video analysis results, and 3D point cloud processing results to generate a unified environmental status dataset.

7. The multi-functional collaborative response processing method for tunnel entrances based on environmental perception according to claim 6, characterized in that, The three-dimensional point cloud data is subjected to voxel rasterization to obtain simplified point cloud data. The specific process is as follows: Determine the minimum bounding box of the 3D point cloud data; The minimum bounding box is divided into multiple voxel grids according to the preset voxel grid side length; Determine the voxel grid to which each point data belongs in the point cloud data; Determine the centroid of the voxel grid, and use the point data corresponding to the centroid as all point data in the voxel grid, or use the point data closest to the centroid as all point data in the voxel grid, to obtain the simplified point cloud data. The formula for calculating the number of the multiple voxel grids is: ; The number of the plurality of voxel grids; for The number of voxel grids along the axis; for The number of voxel grids along the axis; for The number of voxel grids along the axis; , , The calculation formula is: ， in, for The length of the minimum bounding box in the axial direction; for The length of the minimum bounding box in the axial direction; for The length of the minimum bounding box along the axis; cell is the side length of the voxel grid; , , The calculation formula is as follows: ， in, , The point cloud data are respectively in Maximum and minimum values ​​on the axis; , The point cloud data are respectively in Maximum and minimum values ​​on the axis; , The point cloud data are respectively in Maximum and minimum values ​​on the axis.

8. The multi-functional collaborative response processing method for tunnel entrances based on environmental perception according to claim 6, characterized in that, The specific process of ordering the simplified point cloud data to obtain ordered point cloud data is as follows: In the Cartesian coordinate system - The plane is transformed into a circle with an infinite radius, and the circle is divided into multiple sectors based on the radian parameter; Based on the aforementioned sector, the simplified point cloud data is divided into multiple point cloud data sectors; Based on preset distance parameters and preset division conditions, each point cloud data sector in the multiple point cloud data sectors is divided into multiple point cloud data sector rings to obtain ordered point cloud data. The formula for calculating the number of point cloud data sectors is: ， ， in, Point data within a sector of point cloud data; To simplify point cloud data; It is simplified point cloud data The first in Data points; Represents a dotted cloud fan shape; For point data The position within the dotted cloud fan surface; Point The corresponding sector number is ; Points are expressed in radians The arctangent value in a planar coordinate system; For radians; The preset partitioning condition expression is: ， in, The first preset distance parameter; This is the second preset distance parameter; For the first sector of point cloud data The x-coordinate of each data point; For the first sector of point cloud data The vertical coordinates of the data points.

9. The multi-functional collaborative response processing method for tunnel entrances based on environmental perception according to claim 6, characterized in that, The determination of ground point cloud data and non-ground point cloud data in ordered point cloud data specifically includes: The ordered point cloud data is dimensionality reduced to obtain dimensionality-reduced point cloud data. Based on preset constraints, linear fitting is performed on the dimensionality-reduced point cloud data to obtain multiple fitted lines. It is then determined whether the distance between each point cloud and the fitted line is greater than a threshold height. Points greater than the threshold height are considered non-ground point cloud data, while those less than or equal to the threshold height are considered ground point cloud data. The expression for the dimensionality-reduced point cloud data is: ， in, The coordinates of the points in the ordered point cloud data; The coordinates of the point cloud in the reduced-dimensional point cloud data; This refers to the angle between the point cloud and the origin in the reduced-dimensional point cloud data.

10. A multi-functional collaborative response processing method for tunnel entrances based on environmental perception according to any one of claims 6-9, characterized in that, The response priority control logic in step S3 is as follows: if multiple working conditions are identified at the same time, only the response action corresponding to the high-priority working condition is triggered; if it is a single working condition, the corresponding response action is triggered directly; that is, when multiple working condition types are identified at the same time, the fire emergency working condition is responded to first, followed by the icing risk working condition, and finally the traffic abnormality guidance working condition. When an icing risk condition is identified, the de-icing agent spraying device is controlled to release de-icing agent; when a fire emergency condition is identified, the fire extinguishing agent spraying device is controlled to release fire extinguishing agent and the traffic guidance module is activated to issue a warning; when an abnormal traffic guidance condition is identified, the traffic guidance module is activated to guide vehicles to slow down or divert. The traffic guidance module in step S3 includes a flashing light, a high-brightness LED guidance screen, and a voice prompt device. The flashing light, the high-brightness LED guidance screen, and the voice prompt device are all connected to the integrated control unit and operate on the network. They execute corresponding warning or guidance operations according to the instructions of the integrated control unit.

Images