Accident detection early warning method and system based on intelligent induction object

Abstract

The invention provides an accident detection early warning method and system based on an intelligent induction object. According to the method, track space-time characteristic data are collected, and a dynamic time warping principle is utilized to align a multi-vehicle track time sequence to generate a synchronization sequence; meanwhile, the intelligent induction mark is driven to dynamically adjust the curve lamplight projection direction, the tunnel design geometric curvature is extracted, and the track sequence is matched with the curve space structure; fusing the curve adaptive track and the vehicle space distribution characteristics, and when the abnormal position aggregation degree exceeds a dynamic threshold value, generating an accident upstream blind area coverage instruction and linkage light parameters; after responding to the instruction, dynamically adjusting the physical projection attitude to generate real-time parameters; and a control signal is generated based on real-time matching of the light parameters and the attitude parameters, and millisecond-level collaborative early warning of tunnel curve accidents is realized. According to the method, through cross-layer cooperation of trajectory curvature adaptation, space aggregation degree judgment and magneto-rheological posture regulation and control, dynamic lamplight accurate coverage of the upstream blind area of the tunnel curve accident is achieved.

Description

Technical Field

[0001] The present application relates to the field of intelligent traffic active safety prevention and control technology, and in particular to an accident detection and early warning method and system based on intelligent induction targets. Background Art

[0002] Due to the unique spatial structure and driving environment of tunnel curves, multiple risks such as obstructed vision, vehicle centrifugal deviation, and insufficient braking distance are present. Therefore, a real-time, accurate accident detection and dynamic guidance mechanism is required. This scenario requires a technical solution that overcomes three core bottlenecks: first, overcoming the monitoring blind spots caused by the curvature of the curve to achieve comprehensive accident detection; second, dynamically adjusting the warning coverage based on the accident location to accurately guide oncoming vehicles; and third, integrating multi-dimensional data such as vehicle behavior, road geometry, and environmental conditions to achieve a closed-loop response from accident identification to proactive intervention in a very short time, thereby preventing a chain reaction of secondary accidents.

[0003] The current mainstream technology uses a computer vision-based event detection system linked to a static guidance device. This system deploys intelligent cameras and millimeter-wave radar fusion sensing devices on the tunnel sidewalls, using deep learning algorithms to identify abnormal vehicle behavior and trigger fixed-angle guidance beacons at pre-set locations to provide light warnings. This solution, which utilizes edge computing devices for localized analysis and decision-making, avoids data forwarding delays from central platforms and has achieved initial success in some straight tunnel sections.

[0004] However, existing solutions still have fundamental flaws in curve accident scenarios: First, fixed-mounted guidance beacons cannot dynamically adjust their projection angles, resulting in a serious mismatch between lighting coverage and the spatial distribution of blind spots on curves, leaving numerous warning blind spots in sections with smaller curvature radii. Second, visual perception is obstructed by curve walls and interfered with by light refraction, significantly increasing the missed detection rate for low-speed accident targets close to the inside lane, and false triggering is frequent in rainy and foggy weather. Finally, the system response chain still relies on multi-level software decision-making, with a difficult-to-eliminate delay window from event confirmation to hardware action, making it impossible to meet the three-second safety braking time requirement in curve scenarios. These flaws collectively lead to existing solutions frequently experiencing problems such as delayed warnings, coverage deviations, and inaccurate responses in real-world curve accident handling. Summary of the Invention

[0005] The embodiments of the present application provide an accident detection and early warning method and system based on intelligent induction targets to solve the problem of low accuracy in the prior art.

[0006] In a first aspect, an embodiment of the present application provides an accident detection and early warning method based on intelligent guidance targets, comprising:

[0007] Acquire trajectory spatiotemporal feature data of multiple target vehicles, wherein the trajectory spatiotemporal feature data includes position, velocity, and acceleration information;

[0008] Performing time sequence alignment processing on the spatiotemporal feature data of the trajectory using the principle of dynamic time warping to generate a synchronized trajectory sequence;

[0009] Dynamically adjust the light projection direction of the target vehicle at the curve using an intelligent induction beacon, extract the geometric curvature of the tunnel curve, and use the geometric curvature of the tunnel curve to perform spatial coordinate transformation on the synchronized trajectory sequence to generate curve adaptation trajectory data;

[0010] By integrating the curve adaptation trajectory data with the spatial position distribution characteristics of the target vehicle acquired in real time by the accident monitoring scanning equipment, when the abnormal position concentration of the target vehicle exceeds a preset dynamic threshold, an accident upstream blind spot coverage instruction and linkage light warning parameters are generated;

[0011] Responding to the accident upstream blind spot coverage instruction, and dynamically adjusting the physical projection posture of the smart guidance beacon through the magnetorheological damper built into the smart guidance beacon to generate real-time projection posture parameters;

[0012] Based on the linkage light warning parameters and the real-time projection posture parameters, a real-time matching light control signal is generated.

[0013] Optionally, in response to the accident upstream blind spot coverage instruction, the physical projection posture of the smart guidance beacon is dynamically adjusted by a magnetorheological damper built into the smart guidance beacon to generate real-time projection posture parameters, including:

[0014] parsing the spatial orientation information contained in the accident upstream blind spot coverage instruction and converting it into attitude adjustment requirement parameters, wherein the attitude adjustment requirement parameters include a target pitch angle value and a target yaw angle value;

[0015] determining an input current intensity of the magnetorheological damper according to a difference between a target pitch angle value and a current pitch angle value of the attitude adjustment requirement parameter, and a difference between a target yaw angle value and a current yaw angle value;

[0016] Applying the input current intensity to the magnetorheological damper so that the magnetic particles inside the magnetorheological damper form a chain structure under the action of the magnetic field, thereby generating a controllable damping force that is positively correlated with the current intensity;

[0017] The controllable damping force drives a supporting link mechanism connected to the base of the intelligent induction marker, so that the supporting link mechanism generates an angular displacement change around the rotation axis, and obtains an angular displacement change amount;

[0018] The angular displacement variation is converted into real-time projection attitude parameters in digital form, wherein the real-time projection attitude parameters include an actual pitch angle value and an actual yaw angle value.

[0019] Optionally, applying the input current intensity to the magnetorheological damper so that magnetic particles inside the magnetorheological damper form a chain structure under the action of the magnetic field and generate a controllable damping force positively correlated with the current intensity includes:

[0020] Transmitting the input current intensity through a wire to a multi-layer electromagnetic coil tightly wound around the surface of the annular inner cavity of the magnetorheological damper;

[0021] Converting electrical energy into magnetic energy through the electromagnetic coil to generate a spatially uniformly distributed axial magnetic field in the annular inner cavity of the magnetorheological damper;

[0022] Under the action of the axial magnetic field, the magnetic particles suspended in the magnetorheological fluid are driven to overcome the Brownian motion force and migrate in a directional manner along the magnetic lines of force. When the magnetic particles migrate to the area with the maximum magnetic flux density, magnetic dipole moments of attraction are generated between adjacent magnetic particles, so that multiple particles are connected end to end to form a columnar chain structure that penetrates the fluid gap.

[0023] The columnar chain structure bridges two opposite sides of the inner cavity of the magnetorheological damper, thereby generating a mechanical obstruction to the fluid flow perpendicular to the direction of the chain structure;

[0024] Based on the principle of viscous resistance in fluid mechanics, the mechanical obstruction is converted into a controllable damping force that corresponds linearly to the current intensity.

[0025] Optionally, using an intelligent induction beacon to dynamically adjust the light projection direction of the target vehicle at the curve, extracting the geometric curvature of the tunnel curve, and using the geometric curvature of the tunnel curve to perform spatial coordinate transformation on the synchronized trajectory sequence to generate curve adaptation trajectory data, including:

[0026] Obtain the centerline curvature radius and pitch slope parameters based on the tunnel design drawings, and combine them to generate the geometric curvature characteristics of the tunnel curve;

[0027] Establishing a curve spatial coordinate system based on the geometric curvature characteristics, wherein the origin of the curve spatial coordinate system is located at the curve entrance tangent point, the horizontal axis is perpendicular to the road surface, and the vertical axis is along the tangent direction of the centerline;

[0028] Mapping the vehicle position data in the synchronized trajectory sequence from a preset earth coordinate system to the curve space coordinate system to generate an intermediate conversion position;

[0029] According to the actual driving trajectory of the tunnel curve and the lateral offset law of the theoretical center line of the tunnel curve, the curvature compensation correction is performed on the intermediate conversion position to generate curve adaptation trajectory data.

[0030] Optionally, according to the actual driving trajectory of the tunnel curve and the lateral deviation law of the theoretical center line of the tunnel curve, curvature compensation correction is performed on the intermediate conversion position to generate curve adaptation trajectory data, including:

[0031] The vehicle trajectory point set that passes through the target curve within a continuous preset period is retrieved from the tunnel curve monitoring system as the location data of the historical natural driving path;

[0032] Calculating a distance value of a normal plane perpendicular to a theoretical center line for each trajectory point in the vehicle trajectory point set, and separating the distance value to obtain a pure lateral position offset component;

[0033] Dividing the statistical segments into equally spaced segments according to the arc length of the tunnel curve, aggregating the pure lateral position offset components in each segment, and extracting the statistical characteristics of the lateral offset distribution under different curvature radii;

[0034] Constructing a mapping table of curvature radius and compensation amount based on the statistical characteristics of the lateral offset distribution, and storing the mapping table in the local memory of the intelligent guidance target;

[0035] Matching the curvature radius index value of the corresponding statistical segment according to the arc length coordinate of the curve where the intermediate conversion position is located;

[0036] Querying the curvature radius and compensation amount mapping table through the curvature radius index value to obtain a target lateral compensation amount;

[0037] The target lateral compensation amount is added to the lateral coordinate value of the intermediate conversion position, and the longitudinal and elevation coordinates are kept unchanged to generate the curve adaptation trajectory data after curvature compensation correction.

[0038] Optionally, the controllable damping force drives a supporting link mechanism connected to the base of the intelligent induction marker, so that the supporting link mechanism generates an angular displacement change around the rotation axis, and the angular displacement change amount is obtained, including:

[0039] The controllable damping force is transmitted to the front hinge point of the active arm of the supporting linkage mechanism, forming a lever torque around the rotation axis at the end of the active arm;

[0040] The lever torque is used to drive the driven arm connected to the active arm, so that the driven arm slides along the arc guide rail fixed to the induction mark base;

[0041] The center of the arc guide rail is constrained to coincide with the center of the rotation axis, ensuring that the movement trajectory of the end of the driven arm is a precise arc path with a constant radius. A displacement mark point is set at the end of the driven arm, and the arc length change data of the displacement mark point relative to the initial position is captured in real time by a laser ranging sensor;

[0042] The angular displacement variation of the support link mechanism is calculated based on the geometric proportional relationship between the arc length variation data and the radius of the arc guide rail.

[0043] Optionally, the curve adaptation trajectory data is integrated with the spatial position distribution characteristics of the target vehicle acquired in real time by the accident monitoring scanning device. When the abnormal position concentration of the target vehicle exceeds a preset dynamic threshold, an accident upstream blind spot coverage instruction and linkage light warning parameters are generated, including:

[0044] Dividing the tunnel curve into spatial partitioning units, wherein the inner area of ​​the tunnel curve adopts a high-density unit partitioning mode, and the outer area of ​​the tunnel curve adopts a low-density unit partitioning mode;

[0045] Scanning the number of vehicles in the spatial partition unit in real time to generate a location concentration base quantity bound to the location of the spatial partition unit;

[0046] Mapping the locations of abnormal speed mutation points identified in the curve adaptation trajectory data to corresponding spatial partition units, marking them as accident risk core units;

[0047] Taking the accident risk core unit as the geometric center, multi-level radiation expansion is performed to the surrounding adjacent units to form a step detection area, and a distance-decay weighted aggregation operation is performed on the position concentration basic quantity within the step detection area to generate a comprehensive position concentration index;

[0048] According to the traffic flow level fed back in real time by the preset tunnel entrance traffic flow detection device, the judgment benchmark value of the comprehensive position concentration index is self-adjusted. When the comprehensive position concentration index continues to exceed the dynamically adjusted judgment benchmark value for a preset time window, the accident upstream blind spot coverage instruction and linkage light warning parameters are generated synchronously.

[0049] In a second aspect, an embodiment of the present application provides an accident detection and warning system based on intelligent guidance targets, comprising:

[0050] A data acquisition module is used to obtain the spatiotemporal feature data of the trajectories of multiple target vehicles, wherein the spatiotemporal feature data of the trajectories include position, velocity and acceleration information;

[0051] A timing synchronization module uses the principle of dynamic time warping to perform timing alignment processing on the temporal and spatial feature data of the trajectory to generate a synchronized trajectory sequence;

[0052] a curvature adaptation module, which uses an intelligent induction beacon to dynamically adjust the light projection direction of the target vehicle at the curve, extracts the geometric curvature of the tunnel curve, and uses the geometric curvature of the tunnel curve to perform spatial coordinate transformation on the synchronized trajectory sequence to generate curve adaptation trajectory data;

[0053] A decision generation module, which integrates the curve adaptation trajectory data with the spatial position distribution characteristics of the target vehicle acquired in real time by the accident monitoring scanning equipment, and generates an accident upstream blind spot coverage instruction and linkage light warning parameters when the abnormal position concentration of the target vehicle exceeds a preset dynamic threshold;

[0054] a posture control module, responding to the accident upstream blind spot coverage instruction, and dynamically adjusting the physical projection posture of the smart guidance beacon through the magnetorheological damper built into the smart guidance beacon to generate real-time projection posture parameters;

[0055] The optical communication joint control module generates a real-time matching light control signal based on the linkage light warning parameters and the real-time projection posture parameters.

[0056] In a third aspect, an embodiment of the present application provides a computing device comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement an accident detection and early warning method based on intelligent induction targets as described in the first aspect above.

[0057] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, it implements an accident detection and early warning method based on intelligent induction targets as described in the first aspect.

[0058] In an embodiment of the present application, spatiotemporal feature data of trajectories of multiple target vehicles are obtained, wherein the spatiotemporal feature data of the trajectories include position, velocity, and acceleration information; the spatiotemporal feature data of the trajectories are time-series aligned using the principle of dynamic time warping to generate a synchronized trajectory sequence; an intelligent guide beacon is used to dynamically adjust the light projection direction of the target vehicles at a curve, extract the geometric curvature of the tunnel curve, and use the geometric curvature of the tunnel curve to perform spatial coordinate transformation on the synchronized trajectory sequence to generate curve-adapted trajectory data; the curve-adapted trajectory data is integrated with the spatial position distribution characteristics of the target vehicles acquired in real time by an accident monitoring scanning device, and when the abnormal position concentration of the target vehicles exceeds a preset dynamic threshold, an accident upstream blind spot coverage instruction and a linkage light warning parameter are generated; in response to the accident upstream blind spot coverage instruction, the physical projection posture of the intelligent guide beacon is dynamically adjusted through the magnetorheological damper built into the intelligent guide beacon to generate real-time projection posture parameters; and a real-time matching light control signal is generated based on the linkage light warning parameters and the real-time projection posture parameters.

[0059] The technical solution of this application has the following beneficial effects: It solves the problem of the source of original behavior data of multiple target vehicles and provides key feature inputs such as speed mutation for accident judgment. It eliminates the phase offset of multiple vehicle trajectories caused by sampling time difference and generates a time-synchronized trajectory sequence to improve the accuracy of accident location. By dynamically adjusting the direction of light projection to match the direction of the curve, and using the geometric curvature of the tunnel to correct the spatial distortion of the trajectory, trajectory data adapted to the physical environment is generated. The authenticity of the accident is verified by combining the location distribution characteristics, and instructions and parameters are generated based on the dynamic threshold decision of the aggregation degree to reduce the false alarm rate. The physical projection posture of the induction mark is adjusted dynamically at the millisecond level to break through the hardware limitations of fixed angles. Ensure that the light signal matches the physical posture in real time to achieve a coordinated response of blind spot coverage and warning enhancement.

[0060] Furthermore, the target attitude parameters are generated by analyzing the blind spot instructions, and the current intensity of the magnetorheological damper is calculated based on the angle difference, which drives the magnetic particles to form a chain structure to generate a controllable damping force; this force acts on the supporting linkage mechanism and is converted into a rotational angular displacement, and finally outputs the digital real-time projection attitude parameters.

[0061] The magnetorheological effect is combined with a mechanical transmission mechanism to achieve instantaneous response adjustment of the physical posture of the induced marker, completely breaking through the delay bottleneck of traditional actuators; the direct drive mechanism based on current-angle difference mapping eliminates the closed-loop feedback link and achieves precise angle output without cumulative error; the magnetic particle chain structure forms a self-stabilizing damping field, effectively suppressing the posture drift caused by the vibration of the tunnel environment; the pitch angle and yaw angle coordinated adjustment mechanism adaptively matches the spatial distribution characteristics of the upstream blind area of ​​curves with different curvatures.

[0062] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0064] Figure 1 A flowchart of an accident detection and early warning method based on intelligent guidance targets provided by the present application is shown;

[0065] Figure 2 The present invention provides a schematic diagram of an accident detection and early warning system based on intelligent guidance targets;

[0066] Figure 3 A schematic structural diagram of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0067] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0068] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0069] Research has found that tunnel curve accident warning technology, constrained by static hardware architecture and centralized decision-making mechanisms, suffers from fundamental application flaws: fixed-angle guidance devices cannot adapt to the dynamic blind spot distribution caused by the changing curvature of the curve, resulting in a large number of blind spots in upstream warning coverage; video perception methods are obstructed by walls and interfered by light, resulting in a high rate of missed detection of low-speed targets in the inner lane; and multi-level software decision-making chains cause response delays, causing warning actions to lag behind the safe braking time window on the curve, resulting in low secondary accident prevention and control effectiveness. These shortcomings collectively stem from the lack of existing solutions in three dimensions: dynamic control at the physical layer, precise perception at the spatial layer, and real-time coordination at the timing layer. This makes it difficult to meet the closed-loop requirements for accident handling in curve scenarios.

[0070] In response to the above problems, this application proposes an accident detection and early warning method based on intelligent induction markers. Specifically, its core innovation lies in: realizing millisecond-level dynamic adjustment of the physical projection posture of the induction marker through magnetorheological dampers, combining the trajectory space transformation driven by tunnel curvature with the vehicle concentration fusion judgment mechanism, and building an integrated "perception-decision-execution" closed loop. Specifically, the geometric curvature of the curve is used to correct the spatial distortion of the vehicle trajectory, and the position distribution characteristics of the scanning equipment are synchronously integrated to accurately locate the accident; based on the positioning results, blind spot coverage instructions are generated to drive the magnetorheological damper to adjust the pitch and yaw angles of the induction marker in real time; finally, through the linkage control of posture parameters and light signals, the warning range is adaptively covered to the blind spot upstream of the accident. This method solves the coverage mismatch problem of static induction from the root, eliminates system delays with physical layer direct drive control, and overcomes perception blind spots with the help of curvature space mapping, forming a complete prevention and control closed loop for tunnel curve accident scenarios.

[0071] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0072] Figure 1 A flowchart of an accident detection and early warning method based on intelligent guidance targets is provided for the embodiment of the present application. Figure 1 As shown, the method includes:

[0073] 101. Acquire trajectory spatiotemporal feature data of multiple target vehicles, wherein the trajectory spatiotemporal feature data includes position, velocity, and acceleration information;

[0074] In this step, the trajectory spatiotemporal feature data refers to the vehicle position information, i.e., spatial coordinates;

[0075] Velocity information is the instantaneous rate of motion; acceleration information is a dynamic data set of the rate of change of velocity;

[0076] The data is collected by fusing millimeter-wave radar and video sensors installed on the tunnel wall.

[0077] In an embodiment of the present application, a detection signal is first emitted by a sensing device fixedly installed on the tunnel wall to capture the real-time spatial coordinates of the vehicle; secondly, the vehicle movement rate is extracted using continuous frame image analysis technology; then, the instantaneous acceleration is calculated by combining the spatial displacement and the time interval; finally, the position, velocity, and acceleration information are integrated according to the vehicle target identifier to form a trajectory spatiotemporal feature data set.

[0078] In an actual case, at the entrance of a tunnel curve, the sensing device detected the changes in the spatial coordinates of multiple vehicles, combined with video analysis to determine their movement rate and acceleration trend, forming a feature data packet containing three-dimensional position, movement speed and acceleration direction.

[0079] 102. Performing time sequence alignment processing on the trajectory spatiotemporal feature data using the principle of dynamic time warping to generate a synchronized trajectory sequence;

[0080] In this step, the synchronized trajectory sequence refers to the set of time-aligned trajectories formed after eliminating the time stamp deviation of multiple vehicle trajectories;

[0081] The principle of dynamic time warping refers to a method of matching time series data of different lengths through nonlinear path mapping.

[0082] In the embodiment of the present application, the phase offset of the time axis of different vehicle trajectories is first identified; secondly, a nonlinear time calibration method is used to flexibly match the trajectory points; then, the sampling time difference of multiple vehicle trajectories is eliminated by time axis expansion and contraction; finally, a synchronized trajectory sequence with a unified time base is output to ensure that all vehicle behaviors are in the same time reference frame.

[0083] Continuing with the above example, for vehicles traveling on a curve, the system detects that their trajectory timestamps are asynchronous. Through dynamic calibration, the sudden braking moments of different vehicles are aligned to the same time node, forming a synchronized trajectory sequence that can be compared and analyzed.

[0084] 103. Dynamically adjust the light projection direction of the target vehicle at the curve using an intelligent guidance beacon, extract the geometric curvature of the tunnel curve, and use the geometric curvature of the tunnel curve to perform spatial coordinate transformation on the synchronized trajectory sequence to generate curve adaptation trajectory data;

[0085] In this step, the curve adaptation trajectory data refers to the vehicle trajectory after the spatial coordinate transformation to eliminate the curvature distortion;

[0086] The geometric curvature of a tunnel curve refers to the geometric characteristic parameters composed of the centerline turning radius and the longitudinal slope.

[0087] In an embodiment of the present application, the rotation mechanism of the guide beacon is first controlled to project the light along the tangent direction of the center line of the curve; secondly, the center line curvature parameters are extracted from the tunnel structure database; then a local space coordinate system with the geometric center of the curve as the origin is constructed; finally, the geodetic coordinates of the synchronized trajectory are mapped to the local coordinate system to eliminate the trajectory distortion caused by the curvature of the curve.

[0088] Continuing with the above example, in a curve section where the curvature changes significantly, the induction marker automatically adjusts the projection direction to the extension direction of the curve, and at the same time converts the vehicle trajectory coordinates to the local space coordinate system of the curve to correct the measurement position deviation caused by the centrifugal phenomenon.

[0089] 104. Integrate the curve adaptation trajectory data with the spatial position distribution characteristics of the target vehicle acquired in real time by the accident monitoring scanning device. When the abnormal position concentration of the target vehicle exceeds a preset dynamic threshold, generate an accident upstream blind spot coverage instruction and linkage light warning parameters.

[0090] In this step, the abnormal location aggregation refers to the abnormal value of the spatial distribution density of vehicles in the local area;

[0091] Accident monitoring scanning equipment refers to a tunnel top-mounted three-dimensional laser scanning device.

[0092] In an embodiment of the present application, the spatial distribution topology of the vehicle group is first obtained through a three-dimensional scanning device; secondly, the curve area is divided into equal-area analysis units; then the core unit where the speed mutation point is located is located; then the vehicle density weighted value of the area surrounding the core unit is calculated; finally, the judgment threshold is dynamically adjusted according to the real-time traffic density, and when the weighted density exceeds the threshold, an accident instruction and lighting parameters are generated.

[0093] Continuing with the above example, when a vehicle suddenly slows down on the inside of a curve, the system detects an abnormally high density of vehicles in the area where it is located. Combined with the current traffic flow, it determines that this is an accident and generates a coverage instruction pointing to the blind spot upstream of the accident and red high-frequency flashing warning parameters.

[0094] 105. Responding to the accident upstream blind spot coverage instruction, dynamically adjusting the physical projection posture of the smart guidance beacon through the magnetorheological damper built into the smart guidance beacon to generate real-time projection posture parameters;

[0095] In this step, the real-time projection posture parameter refers to the digitally expressed induced target space angle state;

[0096] A magnetorheological damper is an actuator that outputs damping force by controlling fluid viscosity through a magnetic field.

[0097] In an embodiment of the present application, the spatial orientation parameters in the instruction are first parsed and converted into the target angle; secondly, the vector difference between the target angle and the current angle of the induction mark is calculated; then the angle difference is linearly mapped to a current intensity value; then, current is applied to the magnetorheological damper to form a chain arrangement of magnetic particles; finally, the fluid resistance is converted into a rotation angle through a four-bar linkage mechanism, and the digital posture parameters are output.

[0098] Continuing with the above example, the system calculates the pitch and yaw angles that need to be adjusted for the beacon based on the upstream direction of the accident. The magnetorheological damper drives the connecting rod mechanism to rotate under the action of electric current, so that the optical component of the beacon is accurately turned to the target area.

[0099] 106. Generate a real-time matching lighting control signal based on the linkage lighting warning parameter and the real-time projection posture parameter.

[0100] In this step, the real-time matching lighting control signal refers to a light instruction set linked to the physical gesture, including color frequency and brightness parameters.

[0101] In the embodiment of the present application, the light flashing mode code in the linkage parameter is first read; secondly, the spatial angle value in the real-time posture parameter is obtained; then, a spatial mapping relationship between the light color frequency and the projection angle is established; finally, a pulse control signal is generated to drive the light source chip to ensure that the light beam shape is completely matched with the physical posture.

[0102] Continuing with the above example, the control module combines the red high-frequency flashing command with the real-time angle parameter to generate a current pulse with a specific duty cycle, driving the LED array to project a red warning beam to the corrected spatial orientation.

[0103] In summary, steps 101 to 106 eliminate measurement time differences through spatial and temporal trajectory synchronization, correct position distortion using curve curvature mapping, dynamically determine the accident location based on vehicle density, and utilize a magnetorheological mechanism to achieve millisecond-level attitude control of the guidance beacon, ultimately achieving precise spatial alignment between the warning light and the accident blind spot. This entire process overcomes the coverage limitations of static guidance devices, overcomes environmental perception blind spots, and establishes an efficient closed-loop system from accident identification to proactive intervention, significantly improving the safety factor of driving on tunnel curves.

[0104] In order to solve the problem of delay and insufficient accuracy in attitude adjustment of the guidance beacon in tunnel curve accident warning, in some embodiments, in step 105, responding to the blind spot coverage instruction upstream of the accident, and dynamically adjusting the physical projection attitude of the smart guidance beacon through the magnetorheological damper built into the smart guidance beacon to generate real-time projection attitude parameters include:

[0105] 201. Analyze the spatial orientation information contained in the accident upstream blind spot coverage instruction and convert it into attitude adjustment requirement parameters, wherein the attitude adjustment requirement parameters include a target pitch angle value and a target yaw angle value;

[0106] In step 201, the spatial orientation information refers to the three-dimensional positioning data of the blind spot upstream of the accident in the tunnel coordinate system;

[0107] The attitude adjustment requirement parameter refers to the target angle set that quantifies the induction beacon's spatial pointing direction, including the target pitch angle value, i.e., the beam elevation angle in the vertical plane;

[0108] The target yaw angle value refers to the beam deflection angle in the horizontal plane.

[0109] In an embodiment of the present application, the spatial coordinate data in the blind spot coverage instruction is first read; secondly, an azimuth-angle mapping relationship is established based on the tunnel centerline curvature model; then, the coordinates of the blind spot center point are converted into target values ​​of the pitch angle and yaw angle; finally, the attitude adjustment requirement parameters containing the dual angle values ​​are output.

[0110] 202. Determine the input current intensity of the magnetorheological damper according to the difference between the target pitch angle value and the current pitch angle value of the attitude adjustment requirement parameter, and the difference between the target yaw angle value and the current yaw angle value;

[0111] In step 202 , the input current intensity refers to the amplitude of the electrical signal driving the magnetorheological damper to operate, and is linearly proportional to the angle difference.

[0112] In an embodiment of the present application, the target angle value in the attitude adjustment requirement parameter is first obtained; secondly, the current actual angle value fed back by the angle sensor is read; then, the algebraic difference between the target value and the current value of the pitch angle and yaw angle is calculated respectively; finally, the angle difference is mapped to a current intensity value according to a preset proportional coefficient.

[0113] 203. Applying the input current intensity to the magnetorheological damper so that the magnetic particles inside the magnetorheological damper form a chain structure under the action of the magnetic field, thereby generating a controllable damping force that is positively correlated with the current intensity;

[0114] In step 203, the chain structure refers to a columnar arrangement of magnetic particles connected end to end along magnetic lines of force under the action of a magnetic field;

[0115] Controllable damping force refers to the mechanical resistance output that can be linearly adjusted by changing the current intensity.

[0116] In an embodiment of the present application, first, the current intensity is input into the electromagnetic coil wrapped around the inner wall of the damper; secondly, the coil is energized to generate an axial uniform magnetic field; then the magnetic field drives the ferromagnetic particles suspended in the magnetorheological fluid to overcome the fluid resistance and migrate in a directional manner; then, the particles are connected end to end under the attraction of the magnetic dipole moment to form a chain-like column that penetrates the fluid gap; finally, the chain structure enhances the shear strength of the fluid and outputs a controllable damping force that is positively correlated with the current intensity.

[0117] 204. Drive a supporting link mechanism connected to the base of the intelligent induction marker by using the controllable damping force, so that the supporting link mechanism generates an angular displacement change around the rotation axis, and obtain an angular displacement change amount;

[0118] In step 204 , the angular displacement variation refers to the angular variation of the supporting link mechanism rotating around the fixed axis, which is generated by converting the linear damping force through the lever mechanism.

[0119] In an embodiment of the present application, the controllable damping force is first transmitted to the hinge point of the active arm supporting the linkage mechanism; secondly, the damping force forms a lever torque around the rotation axis at the end of the active arm; then the torque drives the driven arm to slide along the arc guide rail; then the guide rail path is constrained to make the end of the driven arm move along a precise circular arc trajectory; finally, the displacement on the circular arc path is recorded through the displacement marking point.

[0120] 205. Convert the angular displacement variation into real-time projection attitude parameters in digital form, wherein the real-time projection attitude parameters include an actual pitch angle value and an actual yaw angle value.

[0121] In step 205, the real-time projection posture parameter refers to a data set that digitally represents the actual spatial angle of the guide beacon optical component.

[0122] In an embodiment of the present application, the displacement arc length of the end of the supporting link is first measured by a laser ranging device; secondly, the rotation angle value is calculated based on the fixed radius of the arc guide rail; then, the mechanical rotation angle is decomposed into pitch and yaw components; finally, digital parameters containing the actual pitch angle value and the actual yaw angle value are output.

[0123] Here's a specific example:

[0124] An accident occurred on a bend in a tunnel, and the system generated an instruction to cover the upstream blind spot. Step 201 parses the spatial coordinates of the instruction and converts them into the required parameters of a pitch angle of 5 degrees and a yaw angle of 25 degrees. Step 202 calculates the current attitude of 3 degrees pitch and 20 degrees yaw, determines that 2 degrees of pitch and 5 degrees of yaw need to be increased, and maps them to a specific current value. Step 203 energizes the magnetorheological damper, and the magnetic field drives the particles to form a chain structure to output the corresponding damping force. In step 204, the damping force drives the connecting rod mechanism, causing the driven arm to rotate along the guide rail to produce an angle change. Step 205 captures the displacement arc length through laser ranging, and converts it into the actual pitch of 4.8 degrees and yaw of 24.9 degrees based on the guide rail radius, and outputs the digital attitude parameters.

[0125] In summary, steps 201 to 205 achieve precise and rapid control of the physical posture of the induction beacon through a five-step closed loop consisting of command parsing, current mapping, magnetorheological response, mechanical conversion, and digital measurement. The magnetic field-driven chain structure formation mechanism overcomes the inertial delay bottleneck of traditional actuators. The mechanical design of the supporting connecting rod and curved guide rail ensures zero backlash in angle conversion, and laser non-contact measurement eliminates mechanical wear errors. The entire process dynamically matches the spatial distribution of the blind spot on the curve with millisecond-level response speeds, fundamentally addressing the coverage inaccuracies and response lags of static induction devices and significantly improving the effectiveness of secondary accident prevention and control.

[0126] In order to overcome the bottlenecks of response delay and precision attenuation of traditional actuators in tunnel vibration environments, in some embodiments, in step 203, the input current intensity is applied to the magnetorheological damper, so that the magnetic particles inside the magnetorheological damper form a chain structure under the action of the magnetic field, thereby generating a controllable damping force positively correlated with the current intensity, including:

[0127] 301. Transmitting the input current intensity to a multi-layer electromagnetic coil tightly wound around the surface of the annular inner cavity of the magnetorheological damper through a wire;

[0128] In step 301, the wire refers to a metal conductor that carries current;

[0129] A multi-layer electromagnetic coil refers to a multi-turn conductive winding tightly wound in a spiral manner on the surface of a toroidal cavity to establish a closed magnetic circuit.

[0130] In an embodiment of the present application, the input current intensity is first introduced into the terminal block via an insulated copper wire; secondly, the current is distributed to the starting turns of the multi-layer coil through the welding point; then the current flows through each layer of the winding along a spiral path; finally, a closed-loop current path is formed that passes through the annular inner cavity.

[0131] 302. Convert electrical energy into magnetic energy through the electromagnetic coil to generate a spatially uniformly distributed axial magnetic field in the annular inner cavity of the magnetorheological damper;

[0132] In step 302, the axial magnetic field refers to the magnetic induction field whose magnetic force lines are parallel to the central axis of the damper;

[0133] Uniform spatial distribution means that the gradient of magnetic field intensity on the annular cross section approaches zero.

[0134] In the embodiment of the present application, first, the energized coil generates an initial magnetic field in the annular inner cavity; secondly, the magnetic field distribution is adjusted by optimizing the pitch of the multi-layer coil turns; then, the soft magnetic material cavity wall is used to constrain the direction of the magnetic lines of force; finally, a stable magnetic field environment is formed that is uniformly distributed along the axial direction.

[0135] 303. Under the action of the axial magnetic field, the magnetic particles suspended in the magnetorheological fluid are driven to overcome the Brownian motion force and migrate in a directional manner along the magnetic lines of force. When the magnetic particles migrate to the area with the maximum magnetic flux density, magnetic dipole moments of attraction are generated between adjacent magnetic particles, so that multiple particles are connected end to end to form a columnar chain structure that penetrates the fluid gap.

[0136] In step 303, the Brownian motion force refers to the random impact force of the thermal motion of fluid molecules on the particles;

[0137] Magnetic dipole moment attraction refers to the mutual attraction between the north and south poles of magnetized particles.

[0138] In the embodiment of the present application, a magnetizing force is first applied to the ferromagnetic particles in the magnetorheological fluid in an axial uniform magnetic field environment; secondly, this force overcomes the random collision interference caused by the thermal motion of the fluid molecules; then, the particles are driven by the magnetic field gradient to perform directional displacement and gather toward the area with maximum magnetic field intensity; then, adjacent particles are attracted to each other due to the north and south magnetic poles generated by magnetization; finally, the particles are connected end to end along the direction of the magnetic field lines to form a columnar chain structure that runs through the fluid gap.

[0139] 304. Bridging opposite sides of the inner cavity of the magnetorheological damper by the columnar chain structure to produce a mechanical obstruction to the fluid flow perpendicular to the direction of the chain structure;

[0140] In step 304, the mechanical obstruction refers to the resistance effect caused by the chain structure physically blocking the fluid flow path;

[0141] Bridging refers to the mechanical state in which both ends of the chain structure are anchored to the inner wall of the cavity.

[0142] In the embodiment of the present application, first, a columnar chain structure spans the inner cavity of the damper; secondly, both ends of the chain body are embedded in the preset micro-grooves of the cavity wall; then the middle part of the chain body is suspended in the magnetorheological fluid; then when the fluid flows perpendicular to the direction of the chain body; finally, the chain structure forms a physical barrier to prevent the fluid from passing through.

[0143] 305. Based on the principle of viscous resistance in fluid mechanics, the mechanical obstruction is converted into a controllable damping force that corresponds linearly to the current intensity.

[0144] In step 305, fluid mechanics viscous resistance refers to the internal friction force generated when the fluid is sheared and deformed;

[0145] Linear correspondence refers to the direct proportional relationship between damping force and current intensity.

[0146] In the embodiment of the present application, first, the fluid flows through the gap of the chain structure to produce shear deformation; secondly, the viscous fluid forms a velocity gradient on the surface of the chain; then resistance is generated according to Newton's law of viscosity; finally, the resistance is proportional to the magnetic field strength, that is, linearly corresponding to the current intensity.

[0147] Here's a specific example:

[0148] When the system requires increased damping force, a specific current is fed through a multilayer electromagnetic coil via a wire. The coil establishes a uniform axial magnetic field within the annular cavity. This field drives ferromagnetic particles to the center, where they connect end to end under the attraction of the magnetic poles, forming a columnar chain. The chain's ends are anchored to the cavity walls, while a suspended segment in the middle hinders vertical fluid flow. Shear deformation of the fluid generates viscous resistance. This resistance increases linearly with increasing input current, ultimately delivering the desired controllable damping force.

[0149] In summary, steps 301 to 305 achieve precise and controllable regulation of the damping force through a hierarchical, progressive mechanism involving current conduction, magnetic field construction, particle self-assembly, structural bridging, and viscous-resistance conversion. The axially uniform magnetic field ensures the consistency of particle migration direction, magnetic dipole moment attraction enables the self-organization of the chain structure, the bridging design enhances mechanical resistance stability, and viscous resistance conversion ensures output linearity. This entire process overcomes the temperature sensitivity and mechanical hysteresis shortcomings of traditional hydraulic dampers, maintaining millisecond-level response characteristics even in tunnel vibration environments, providing highly reliable execution power for induction beacon attitude control.

[0150] To overcome the problems of vehicle trajectory measurement distortion and inaccurate induced control in a curved environment, in some embodiments, in step 103, an intelligent induced beacon is used to dynamically adjust the light projection direction of the target vehicle at the curve, extract the geometric curvature of the tunnel curve, and use the geometric curvature of the tunnel curve to perform spatial coordinate transformation on the synchronized trajectory sequence to generate curve-adapted trajectory data, including:

[0151] 401. Obtain the centerline curvature radius and pitch slope parameters according to the tunnel design drawings, and combine them to generate the geometric curvature characteristics of the tunnel curve;

[0152] In step 401, the centerline curvature radius refers to the turning radius value of the tunnel centerline arc segment;

[0153] The pitch slope parameter refers to the tangent value of the longitudinal inclination of the road;

[0154] Geometric curvature characteristics refer to the three-dimensional spatial curvature measurement constructed by combining curvature radius and slope.

[0155] In this embodiment, the digitized design drawings of the target curve are first retrieved through the tunnel engineering database interface. Next, the centerline geometry parameter annotation layer is located in the drawing structure tree. The curvature radius design value and slope percentage data within this layer are then parsed. Finally, these data are packaged into a geometric curvature feature data package that characterizes the three-dimensional curvature characteristics of the curve. This process requires verifying parameter validity and discarding historical versions prior to construction changes.

[0156] 402. Establish a curve spatial coordinate system based on the geometric curvature characteristics, wherein the origin of the curve spatial coordinate system is located at the curve entrance tangent point, the horizontal axis is perpendicular to the road surface, and the vertical axis is along the tangent direction of the centerline;

[0157] In step 402, the curve spatial coordinate system refers to a rectangular coordinate system based on the local geometric features of the curve;

[0158] The entry tangent point refers to the point where the starting end of the curve meets the straight segment;

[0159] The centerline tangent direction refers to the instantaneous extension direction of the centerline at that point.

[0160] In this embodiment, the spatial coordinates of the mathematical tangent point between a line segment and a circular arc segment are first calculated based on the geometric curvature characteristics. Next, the coordinate system is initialized with this tangent point as the origin. The road surface normal at this point is then obtained as the reference direction for the vertical axis. Finally, the positive longitudinal axis direction is determined by advancing the tangent direction along the centerline. Once the coordinate system is constructed, the origin coordinates and axial unit vectors are stored for subsequent use.

[0161] 403. Mapping the vehicle position data in the synchronized trajectory sequence from a preset earth coordinate system to the curve space coordinate system to generate an intermediate conversion position;

[0162] In step 403, the geodetic coordinate system refers to a global coordinate system based on the earth ellipsoid;

[0163] The intermediate conversion position refers to the transition coordinate value during the coordinate system conversion process.

[0164] In the embodiment of the present application, the vehicle's geodetic coordinates (longitude, latitude, and elevation) in the synchronized trajectory sequence are first read. A rotation transformation matrix, comprising an origin offset and an axial rotation angle, is then calculated based on the spatial position relationship of the dual coordinate systems. The vehicle's geodetic coordinates are then substituted into the matrix operation formula. Finally, a two-dimensional plane coordinate set, i.e., the intermediate transformation position, is output in the local coordinate system of the curve.

[0165] 404. Perform curvature compensation correction on the intermediate conversion position according to the actual driving trajectory of the tunnel curve and the lateral offset law of the theoretical center line of the tunnel curve to generate curve adaptation trajectory data.

[0166] In step 404, the lateral deviation law refers to the normal distance distribution characteristics of the actual trajectory of the vehicle relative to the theoretical center line;

[0167] Curvature compensation correction refers to the reverse calibration of coordinate values ​​based on the offset law.

[0168] In an embodiment of the present application, the historical vehicle trajectory dataset of the curve is first retrieved from the traffic data center; secondly, the vertical distance between each trajectory point and the theoretical center line, that is, the lateral offset, is calculated; then, the statistical distribution law of the offset in different curvature sections is analyzed to establish a mapping model; finally, the compensation value is obtained based on the real-time curvature retrieval mapping model of the current curve, and the lateral coordinates of the intermediate conversion position are algebraically superimposed and corrected to generate the final adapted trajectory.

[0169] Here's a specific example:

[0170] In a specific right-hand curve accident warning, the system retrieves parameters for a curvature radius of 600 meters and a slope of 3 degrees from the design drawing. A coordinate system is established with the curve's entrance point as the origin, with the horizontal axis perpendicular to the road surface and the vertical axis pointing in the direction of the curve. The GPS coordinates of the vehicle in the synchronized trajectory are converted to this coordinate system to generate intermediate position data. Based on historical data statistics, it is found that vehicles at this curvature have an average outward offset of 1.2 meters. Based on this, the horizontal coordinate value of the intermediate position is increased by 1.2 meters to compensate for this. The final output is the corrected curve-adapted trajectory data.

[0171] In summary, steps 401 to 404 achieve precise matching of the vehicle trajectory with the physical structure of the curve through the coordinated efforts of design parameter extraction, local coordinate system construction, spatial coordinate mapping, and behavioral compensation. Curvature characteristics drive the establishment of the coordinate system to ensure spatial datum consistency, while historical offset compensation eliminates measurement deviations caused by centrifugal effects. This entire process overcomes the limitations of the geodetic coordinate system in curved scenarios, providing a trajectory data foundation with geometric distortion correction for accident location, fundamentally resolving the spatial adaptation distortion problem of traditional straight-line trajectory models in curved environments.

[0172] To address the problem of misjudgment of accidents caused by measurement distortion of vehicle trajectories due to centrifugal force in a tunnel curve environment, in some embodiments, step 404 performs curvature compensation correction on the intermediate transition position based on the actual driving trajectory in the tunnel curve and the lateral offset law of the theoretical centerline of the tunnel curve to generate curve adaptation trajectory data, including:

[0173] 501. Retrieving a set of vehicle trajectory points that have passed through a target curve within a continuous preset period from a monitoring system of a tunnel curve as position data of a historical natural driving path;

[0174] In step 501, the location data of the historical natural driving path refers to a set of continuous trajectory points of the real vehicle without human intervention;

[0175] The preset period refers to the complete time span covering the peak and trough of traffic flow.

[0176] In an embodiment of the present application, an encrypted data request is first initiated to the tunnel central monitoring system; secondly, the system screens the trajectory point set of the target curve in consecutive complete natural days; then, the positioning drift points and abnormal stationary points are filtered; finally, the vehicle identity identifier is classified, packaged and transmitted to the induction mark edge computing unit.

[0177] 502. Calculate the distance value of the normal plane perpendicular to the theoretical center line for each trajectory point in the vehicle trajectory point set, and separate it to obtain a pure lateral position offset component;

[0178] In step 502, the normal plane refers to a plane passing through a point on the theoretical centerline and perpendicular to the tangent line of the point;

[0179] The pure lateral position offset component refers to the vertical distance value from the trajectory point to the normal plane.

[0180] In an embodiment of the present application, first, the vertical projection point of each trajectory point on the theoretical center line is calculated; secondly, the tangent direction vector of the center line at the projection point is obtained; then, a spatial normal plane perpendicular to the tangent is constructed; finally, the shortest Euclidean distance from the trajectory point to the normal plane is measured, which is the pure lateral position offset component.

[0181] 503. Divide the tunnel into statistical segments at equal intervals according to the arc length of the tunnel curve, aggregate the pure lateral position offset components in each segment, and extract statistical characteristics of lateral offset distribution under different curvature radii;

[0182] In step 503 , the statistical characteristics of the lateral offset distribution refer to the central tendency and dispersion of the offsets in different curvature segments.

[0183] In an embodiment of the present application, the arc length of the center line of the curve is first divided into continuous segments of fixed length; secondly, each trajectory point is assigned to the corresponding segment according to the arc length coordinate; then the arithmetic mean of all offset components in each segment is calculated; finally, a mapping relationship library between the curvature radius of the segment center point and the average offset is established.

[0184] 504. Constructing a mapping table of curvature radius and compensation amount based on the statistical characteristics of the lateral offset distribution, and storing the mapping table in a local memory of the intelligent guidance target;

[0185] In step 504 , the curvature radius and compensation amount mapping table refers to a data structure storing curvature values ​​and corresponding lateral compensation amounts.

[0186] In an embodiment of the present application, first, the changing pattern of the average offset of each section is analyzed; secondly, the compensation direction (positive for external deviation and negative for internal deviation) is determined according to the geometric characteristics of the curve; then, the corresponding relationship between the curvature radius index key and the compensation amount is established; finally, the mapping table is burned to the specified sector of the induction mark local flash memory chip.

[0187] 505. Match the curvature radius index value of the corresponding statistical segment according to the arc length coordinate of the curve where the intermediate conversion position is located;

[0188] In step 505, the curvature radius index value refers to the curvature identification code in the mapping table that matches the current arc length coordinate.

[0189] In the embodiment of the present application, the arc length coordinate value of the intermediate conversion position is first parsed; secondly, the pre-stored arc length-curvature correspondence tree is queried; then the curvature radius value of the segment where the current arc length is located is located; and finally, it is converted into a mapping table storage address offset.

[0190] 506. Query the curvature radius and compensation amount mapping table using the curvature radius index value to obtain a target lateral compensation amount.

[0191] In step 506 , the target lateral compensation value refers to a specific correction value to be added to the lateral coordinate.

[0192] In an embodiment of the present application, the original value of the compensation amount in the flash memory chip is first read according to the address offset; secondly, the value is checked to see if it is within a preset reasonable range; then the temperature compensation module is activated to correct the influence of the ambient temperature difference; finally, the target lateral compensation amount adapted to the environment is output.

[0193] 507. Add the target lateral compensation amount to the lateral coordinate value of the intermediate conversion position, keep the longitudinal and elevation coordinates unchanged, and generate curve adaptation trajectory data after curvature compensation correction.

[0194] In step 507 , curvature compensation correction refers to an operation of eliminating systematic errors through algebraic operations on coordinate values.

[0195] In an embodiment of the present application, the longitudinal coordinates and elevation coordinates of the intermediate conversion position are first copied; secondly, the sign and amplitude of the target lateral compensation amount are read; then an algebraic superposition operation is performed on the original lateral coordinates; finally, a new lateral coordinate value is generated and reorganized into a three-dimensional space coordinate.

[0196] Here's a specific example:

[0197] In tunnel curves with variable curvature, the system retrieves thirty days of vehicle trajectory data. For each trajectory point, its lateral offset relative to the theoretical centerline is calculated. The curve is divided into 20-meter segments based on arc length, and the average offset of each segment is calculated. Based on the statistical results, a mapping table is constructed and stored in the induction benchmark. For example, a curvature radius of 500 meters corresponds to a 0.8-meter offset compensation. When processing an intermediate transition position, its arc length coordinates are matched to the current curvature radius of 500 meters, and the mapping table is queried to obtain a 0.8-meter compensation. This compensation is superimposed on the lateral coordinates, ultimately generating the corrected adaptive trajectory data.

[0198] In summary, steps 501 to 507 achieve adaptive calibration of vehicle trajectory curvature for curves through a seven-step closed loop consisting of historical trajectory analysis, geometric projection calculation, statistical feature extraction, local mapping table construction, real-time curvature matching, compensation query, and coordinate correction. Normal plane distance calculation eliminates residual errors from coordinate transformation, arc length partitioning statistics capture the laws of centrifugal effects, and the localized mapping table ensures millisecond-level compensation response. This entire process overcomes the physical limitations of satellite positioning in curved scenarios, providing a data foundation that matches real-world vehicle trajectories for accident detection and fundamentally addressing the positioning distortion caused by traditional trajectory models that ignore driver behavior.

[0199] In order to solve the problem of mechanical angle measurement being easily worn and inaccurate in a tunnel vibration environment, in some embodiments, in step 204, the controllable damping force is used to drive a supporting link mechanism connected to the base of the intelligent induction marker, so that the supporting link mechanism generates an angular displacement change around the rotation axis, and the angular displacement change amount is obtained, including:

[0200] 601. The controllable damping force is transmitted to the front hinge point of the active arm of the supporting linkage mechanism, thereby forming a lever torque around the rotation axis at the end of the active arm;

[0201] In step 601 , the lever torque refers to the torque acting on the rotation axis, and its value is the product of the damping force and the lever arm length.

[0202] In an embodiment of the present application, the controllable damping force is first transmitted to the hinge hole at the front end of the active arm supporting the connecting rod mechanism through the ball joint; secondly, the damping force is transmitted to the end along the rigid rod of the active arm; then a force arm is formed between the end of the active arm and the rotation axis; finally, a torque value is generated according to the lever principle to drive the driven arm to rotate.

[0203] 602. Using the lever torque to drive the driven arm connected to the active arm, the driven arm slides along the arc-shaped guide rail fixed to the induction beacon base;

[0204] In step 602 , the arc guide rail refers to a groove-shaped track with a central angle less than 180 degrees, which constrains the moving path of the end of the driven arm.

[0205] In the embodiment of the present application, first, the lever torque acts on the base of the follower arm; secondly, the torque overcomes the static friction of the guide rail to start sliding; then the roller at the end of the follower arm is embedded in the guide rail groove; finally, the roller moves in a circular motion along the groove track, driving the follower arm as a whole to rotate around the axis.

[0206] 603. Constrain the center of the arc guide rail to coincide with the center of the rotation axis, ensuring that the movement trajectory of the end of the driven arm is a precise arc path with a constant radius, and set a displacement mark point at the end of the driven arm. Use a laser ranging sensor to capture the arc length change data of the displacement mark point relative to the initial position in real time;

[0207] In step 603, the arc length change data refers to the cumulative value of the curve distance of the displacement mark point moving along the arc trajectory.

[0208] In an embodiment of the present application, the arc guide rail is first finely machined to ensure that its center coincides with the axis of rotation; secondly, a high-reflectivity marker ball is installed at the end of the slave arm; then a laser ranging sensor emits an infrared beam to the marker ball; finally, the reflected light signal is received and the round-trip time difference of the light beam is calculated, which is converted into an arc length increment of the displacement marker point relative to the initial position.

[0209] 604. Calculate the angular displacement change of the support linkage mechanism according to the geometric proportional relationship between the arc length change data and the radius of the arc guide rail.

[0210] In step 604 , the geometric proportional relationship refers to a linear conversion rule between the arc length and the corresponding central angle.

[0211] In an embodiment of the present application, the fixed radius value of the arc guide rail is first obtained from the structural drawing; secondly, the real-time arc length data measured by the laser sensor is read; then the arc length value is divided by the guide rail radius; and finally, the calculation result is the angular displacement arc value of the supporting connecting rod mechanism rotating around the axis.

[0212] Here's a specific example:

[0213] When the magnetorheological damper outputs a specific damping force, this force is transmitted to the front end of the master arm via the ball joint. The lever torque generated at the end of the master arm drives the slave arm, whose roller slides along a curved guide rail. The precision machining of the guide rail ensures that the center of the circle coincides with the axis of rotation, resulting in a precise circular trajectory for the roller. A ceramic marker ball mounted on the roller reflects a laser beam, and a sensor measures the increase in arc length. Given a fixed guide rail radius, the arc length increment is divided by the radius to determine the change in angular displacement, completing the entire force-to-angle conversion process.

[0214] In summary, steps 601 to 604 achieve precise mapping of damping force to rotation angle through five levels of physical transformation: ball joint force transmission, lever torque amplification, guide rail constraint, laser ranging, and geometric conversion. The concentric design of the curved guide rail and the rotation axis eliminates motion trajectory deviation, and non-contact laser measurement overcomes the wear limitations of traditional potentiometers. This entire process constructs a purely mechanical solution, maintaining micro-radian measurement accuracy even in strong vibration environments. This provides stable and reliable angle feedback for the induced target's attitude control, completely resolving the control inaccuracy problem caused by clearance wear in traditional linkage mechanisms.

[0215] In order to solve the problem of high misjudgment rate caused by sparse vehicle distribution and environmental interference in tunnel curve accident detection, in some embodiments, step 104 integrates the curve adaptation trajectory data with the spatial position distribution characteristics of the target vehicle obtained in real time by the accident monitoring scanning device. When the abnormal position concentration of the target vehicle exceeds a preset dynamic threshold, an accident upstream blind spot coverage instruction and linkage light warning parameters are generated, including:

[0216] 701. Divide the tunnel curve into spatial partition units, wherein the inner area of ​​the tunnel curve adopts a high-density unit partition mode, and the outer area of ​​the tunnel curve adopts a low-density unit partition mode;

[0217] In step 701, the high-density unit division mode refers to a division mode in which the number of division units per unit area is significantly higher than the average value;

[0218] A low-density unit division pattern refers to a division pattern in which the number of division units per unit area is lower than the average.

[0219] In an embodiment of the present application, the partition density dividing line is first determined according to the gradient of the curvature change of the curve; secondly, small-area units are used to continuously cover the inner area of ​​the curve; then, large-area units are sparsely distributed in the outer area of ​​the curve; finally, a non-uniform grid topology structure is generated and solidified for storage.

[0220] 702. Scan the number of vehicles in the spatial partition unit in real time to generate a location aggregation basic quantity bound to the location of the spatial partition unit;

[0221] In step 702, the location concentration base refers to the real-time statistical value of the number of vehicles existing in a single spatial partition unit.

[0222] In an embodiment of the present application, the vehicle point cloud distribution is first obtained through a top-mounted three-dimensional laser scanning device; secondly, the point cloud coordinates are mapped to the spatial partition unit grid; then the number of point clouds in each unit grid is counted; finally, the quantity value is bound to the corresponding unit geographic location code.

[0223] 703. Mapping the locations of abnormal speed mutation points identified in the curve adaptation trajectory data to corresponding spatial partition units, marking them as accident risk core units;

[0224] In step 703, the accident risk core unit refers to a spatial partition unit containing a speed mutation abnormal point.

[0225] In an embodiment of the present application, the acceleration mutation characteristics in the curve adaptation trajectory data are first analyzed; secondly, the spatial coordinates of the negative extreme acceleration point are located; then the coordinates are mapped to the spatial partition unit grid; finally, the target unit is marked as the accident risk core unit.

[0226] 704. Taking the accident risk core unit as the geometric center, perform multi-level radiation expansion to the surrounding adjacent units to form a step detection area, perform distance decay weighted aggregation operation on the position concentration basic quantity within the step detection area, and generate a comprehensive position concentration index;

[0227] In step 704, the step detection area refers to a multi-layer concentric expansion area centered on the core unit;

[0228] Distance decay weighted aggregation refers to an accumulation algorithm in which the weight value decreases as the distance from the core unit increases.

[0229] In the embodiment of the present application, first, a three-level radiation circle is established with the accident risk core unit as the center; secondly, the weight coefficient of the core circle is set to be the highest, and the outer circles decrease step by step; then, the weighted sum of the basic quantities of unit position concentration in each circle is calculated; finally, the weighted sums of each level are accumulated to generate a comprehensive position concentration index.

[0230] 705. Based on the traffic flow level fed back in real time by the preset tunnel entrance traffic flow detection device, the judgment benchmark value of the comprehensive position concentration index is self-adjusted. When the comprehensive position concentration index continues to exceed the judgment benchmark value after dynamic adjustment for a preset time window, the accident upstream blind spot coverage instruction and linkage light warning parameters are synchronously generated.

[0231] In step 705, the self-adjusting determination reference value refers to a threshold parameter that dynamically fluctuates with traffic flow;

[0232] The preset time window refers to the minimum length of time for determining whether an indicator exceeds the limit continuously.

[0233] In an embodiment of the present application, the real-time traffic level signal sent by the vehicle inspection device at the tunnel entrance is first received; secondly, a threshold curve is pre-stored according to the traffic level index; then the corresponding judgment benchmark value is dynamically loaded; finally, the continuous exceeding time of the comprehensive indicator is monitored, and the instruction generation is triggered after reaching the preset window.

[0234] Here's a specific example:

[0235] At the right-hand bend in the tunnel, which has a variable curvature, the system divides the inner area into high-density cells of five meters squared and the outer area into low-density cells of ten meters squared. Laser scanning revealed the presence of three vehicles within a high-density cell, which was bound to the basic quantity of location aggregation. At the same time, it was detected that the speed of a vehicle within this cell dropped sharply from 80 to 5, marking it as a core accident risk cell. A three-level detection circle was established with this cell as the center, with a weight of 1 for the core circle, 0.6 for the middle circle, and 0.3 for the outer circle. The basic quantities of the surrounding cells were weighted and aggregated to obtain a comprehensive index value. Based on the corresponding threshold value of the level load in the entrance traffic flow, if the comprehensive index exceeds the limit for 10 seconds, a red high-frequency flashing instruction covering the upstream 150-meter blind spot is generated.

[0236] In summary, steps 701 to 705 achieve accurate and reliable identification of curve accidents through a five-step collaborative approach: non-uniform spatial partitioning, real-time vehicle statistics, risk core location, tiered weighted aggregation, and dynamic threshold determination. Curvature-driven density partitioning enhances inboard sensitivity, three-level radiation detection overcomes the limitations of localized field of view, and a self-regulating traffic flow mechanism prevents misjudgments of congestion. The entire process builds a fusion decision model of spatial distribution and motion characteristics, significantly improving accident detection confidence and fundamentally addressing the adaptability limitations of traditional single-point detection in complex curve scenarios.

[0237] Figure 2 The present invention provides a schematic diagram of a structure of an accident detection and early warning system based on intelligent guidance targets, as shown in FIG. Figure 2 As shown, the system includes:

[0238] The data acquisition module 21 acquires the spatiotemporal feature data of the trajectories of multiple target vehicles, wherein the spatiotemporal feature data of the trajectories include position, velocity and acceleration information;

[0239] The timing synchronization module 22 performs timing alignment processing on the temporal and spatial feature data of the trajectory using the principle of dynamic time warping to generate a synchronized trajectory sequence;

[0240] The curvature adaptation module 23 dynamically adjusts the light projection direction of the target vehicle at the curve using an intelligent induction beacon, extracts the geometric curvature of the tunnel curve, and uses the geometric curvature of the tunnel curve to perform spatial coordinate transformation on the synchronized trajectory sequence to generate curve adaptation trajectory data;

[0241] The decision generation module 24 integrates the curve adaptation trajectory data with the spatial position distribution characteristics of the target vehicle acquired in real time by the accident monitoring scanning device. When the abnormal position concentration of the target vehicle exceeds a preset dynamic threshold, it generates an accident upstream blind spot coverage instruction and linkage light warning parameters.

[0242] The posture control module 25 responds to the accident upstream blind spot coverage instruction and dynamically adjusts the physical projection posture of the smart guidance beacon through the magnetorheological damper built into the smart guidance beacon to generate real-time projection posture parameters;

[0243] The optical communication control module 26 generates a real-time matching light control signal based on the linkage light warning parameters and the real-time projection posture parameters.

[0244] Figure 2 The accident detection and warning system based on intelligent guidance mark can be executed Figure 1 The implementation principle and technical effects of the accident detection and early warning method based on intelligent guidance markers described in the illustrated embodiment will not be elaborated on here. The specific manner in which each module and unit performs operations in the accident detection and early warning system based on intelligent guidance markers in the above embodiment has been described in detail in the embodiment of the method and will not be elaborated on here.

[0245] In one possible design, Figure 2 The accident detection and warning system based on intelligent guidance mark of the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0246] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0247] The processing component 32 is used for the above Figure 1 The embodiment provides an accident detection and early warning method based on intelligent guidance markers.

[0248] The processing component 32 may include one or more processors to execute computer instructions to perform all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0249] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0250] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0251] The input / output interface provides an interface between the processing component and the peripheral interface module, which can be an output device, an input device, etc.

[0252] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0253] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0254] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The embodiment shown is an accident detection and early warning method based on intelligent guidance marks.

[0255] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0256] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0257] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0258] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An accident detection and early warning method based on intelligent guidance marks, characterized in that: include: Acquire trajectory spatiotemporal feature data of multiple target vehicles, wherein the trajectory spatiotemporal feature data includes position, velocity, and acceleration information; Performing time sequence alignment processing on the spatiotemporal feature data of the trajectory using the principle of dynamic time warping to generate a synchronized trajectory sequence; Dynamically adjust the light projection direction of the target vehicle at the curve using an intelligent induction beacon, extract the geometric curvature of the tunnel curve, and use the geometric curvature of the tunnel curve to perform spatial coordinate transformation on the synchronized trajectory sequence to generate curve adaptation trajectory data; By integrating the curve adaptation trajectory data with the spatial position distribution characteristics of the target vehicle acquired in real time by the accident monitoring scanning equipment, when the abnormal position concentration of the target vehicle exceeds a preset dynamic threshold, an accident upstream blind spot coverage instruction and linkage light warning parameters are generated; Responding to the accident upstream blind spot coverage instruction, and dynamically adjusting the physical projection posture of the smart guidance beacon through the magnetorheological damper built into the smart guidance beacon to generate real-time projection posture parameters; Based on the linkage light warning parameters and the real-time projection posture parameters, a real-time matching light control signal is generated.

2. The method according to claim 1, characterized in that In response to the accident upstream blind spot coverage instruction, the physical projection posture of the intelligent guidance beacon is dynamically adjusted through the magnetorheological damper built into the intelligent guidance beacon to generate real-time projection posture parameters, including: parsing the spatial orientation information contained in the accident upstream blind spot coverage instruction and converting it into attitude adjustment requirement parameters, wherein the attitude adjustment requirement parameters include a target pitch angle value and a target yaw angle value; determining an input current intensity of the magnetorheological damper according to a difference between a target pitch angle value and a current pitch angle value of the attitude adjustment requirement parameter, and a difference between a target yaw angle value and a current yaw angle value; Applying the input current intensity to the magnetorheological damper so that the magnetic particles inside the magnetorheological damper form a chain structure under the action of the magnetic field, thereby generating a controllable damping force that is positively correlated with the current intensity; The controllable damping force drives a supporting link mechanism connected to the base of the intelligent induction marker, so that the supporting link mechanism generates an angular displacement change around the rotation axis, and obtains an angular displacement change amount; The angular displacement variation is converted into real-time projection attitude parameters in digital form, wherein the real-time projection attitude parameters include an actual pitch angle value and an actual yaw angle value.

3. The method according to claim 2, characterized in that Applying the input current intensity to the magnetorheological damper so that the magnetic particles inside the magnetorheological damper form a chain structure under the action of the magnetic field and generate a controllable damping force positively correlated with the current intensity, including: Transmitting the input current intensity through a wire to a multi-layer electromagnetic coil tightly wound around the surface of the annular inner cavity of the magnetorheological damper; Converting electrical energy into magnetic energy through the electromagnetic coil to generate a spatially uniformly distributed axial magnetic field in the annular inner cavity of the magnetorheological damper; Under the action of the axial magnetic field, the magnetic particles suspended in the magnetorheological fluid are driven to overcome the Brownian motion force and migrate in a directional manner along the magnetic lines of force. When the magnetic particles migrate to the area with the maximum magnetic flux density, magnetic dipole moments of attraction are generated between adjacent magnetic particles, so that multiple particles are connected end to end to form a columnar chain structure that penetrates the fluid gap. The columnar chain structure bridges two opposite sides of the inner cavity of the magnetorheological damper, thereby generating a mechanical obstruction to the fluid flow perpendicular to the direction of the chain structure; Based on the principle of viscous resistance in fluid mechanics, the mechanical obstruction is converted into a controllable damping force that corresponds linearly to the current intensity.

4. The method according to claim 1, wherein Dynamically adjusting the light projection direction of the target vehicle at the curve using an intelligent induction beacon, extracting the geometric curvature of the tunnel curve, and using the geometric curvature of the tunnel curve to perform spatial coordinate transformation on the synchronized trajectory sequence to generate curve adaptation trajectory data, including: Obtain the centerline curvature radius and pitch slope parameters based on the tunnel design drawings, and combine them to generate the geometric curvature characteristics of the tunnel curve; Establishing a curve spatial coordinate system based on the geometric curvature characteristics, wherein the origin of the curve spatial coordinate system is located at the curve entrance tangent point, the horizontal axis is perpendicular to the road surface, and the vertical axis is along the tangent direction of the centerline; Mapping the vehicle position data in the synchronized trajectory sequence from a preset earth coordinate system to the curve space coordinate system to generate an intermediate conversion position; According to the actual driving trajectory of the tunnel curve and the lateral offset law of the theoretical center line of the tunnel curve, the curvature compensation correction is performed on the intermediate conversion position to generate curve adaptation trajectory data.

5. The method according to claim 4, characterized in that According to the actual driving trajectory of the tunnel curve and the lateral deviation law of the theoretical center line of the tunnel curve, the curvature compensation correction is performed on the intermediate conversion position to generate curve adaptation trajectory data, including: The vehicle trajectory point set that passes through the target curve within a continuous preset period is retrieved from the tunnel curve monitoring system as the location data of the historical natural driving path; Calculating a distance value of a normal plane perpendicular to a theoretical center line for each trajectory point in the vehicle trajectory point set, and separating the distance value to obtain a pure lateral position offset component; Dividing the statistical segments into equally spaced segments according to the arc length of the tunnel curve, aggregating the pure lateral position offset components in each segment, and extracting the statistical characteristics of the lateral offset distribution under different curvature radii; Constructing a mapping table of curvature radius and compensation amount based on the statistical characteristics of the lateral offset distribution, and storing the mapping table in the local memory of the intelligent guidance target; Matching the curvature radius index value of the corresponding statistical segment according to the arc length coordinate of the curve where the intermediate conversion position is located; Querying the curvature radius and compensation amount mapping table through the curvature radius index value to obtain a target lateral compensation amount; The target lateral compensation amount is added to the lateral coordinate value of the intermediate conversion position, and the longitudinal and elevation coordinates are kept unchanged to generate the curve adaptation trajectory data after curvature compensation correction.

6. The method according to claim 2, characterized in that The controllable damping force drives the supporting link mechanism connected to the base of the intelligent induction marker, so that the supporting link mechanism generates an angular displacement change around the rotation axis, and the angular displacement change amount is obtained, including: The controllable damping force is transmitted to the front hinge point of the active arm of the supporting linkage mechanism, forming a lever torque around the rotation axis at the end of the active arm; The lever torque is used to drive the driven arm connected to the active arm, so that the driven arm slides along the arc guide rail fixed to the induction mark base; The center of the arc guide rail is constrained to coincide with the center of the rotation axis, ensuring that the movement trajectory of the end of the driven arm is a precise arc path with a constant radius. A displacement mark point is set at the end of the driven arm, and the arc length change data of the displacement mark point relative to the initial position is captured in real time by a laser ranging sensor; The angular displacement variation of the support link mechanism is calculated based on the geometric proportional relationship between the arc length variation data and the radius of the arc guide rail.

7. The method according to claim 1, wherein By integrating the curve adaptation trajectory data with the spatial position distribution characteristics of the target vehicle acquired in real time by the accident monitoring scanning device, when the abnormal position concentration of the target vehicle exceeds a preset dynamic threshold, an accident upstream blind spot coverage instruction and linkage light warning parameters are generated, including: Dividing the tunnel curve into spatial partitioning units, wherein the inner area of ​​the tunnel curve adopts a high-density unit partitioning mode, and the outer area of ​​the tunnel curve adopts a low-density unit partitioning mode; Scanning the number of vehicles in the spatial partition unit in real time to generate a location aggregation base quantity bound to the location of the spatial partition unit; Mapping the locations of abnormal speed mutation points identified in the curve adaptation trajectory data to corresponding spatial partition units, marking them as accident risk core units; Taking the accident risk core unit as the geometric center, multi-level radiation expansion is performed to the surrounding adjacent units to form a step detection area, and a distance-decay weighted aggregation operation is performed on the position concentration basic quantity within the step detection area to generate a comprehensive position concentration index; According to the traffic flow level fed back in real time by the preset tunnel entrance traffic flow detection device, the judgment benchmark value of the comprehensive position concentration index is self-adjusted. When the comprehensive position concentration index continues to exceed the dynamically adjusted judgment benchmark value for a preset time window, the accident upstream blind spot coverage instruction and linkage light warning parameters are generated synchronously.

8. An accident detection and early warning system based on intelligent guidance marks, characterized in that: include: A data acquisition module is used to obtain the spatiotemporal feature data of the trajectories of multiple target vehicles, wherein the spatiotemporal feature data of the trajectories include position, velocity and acceleration information; A timing synchronization module uses the principle of dynamic time warping to perform timing alignment processing on the temporal and spatial feature data of the trajectory to generate a synchronized trajectory sequence; a curvature adaptation module, which uses an intelligent induction beacon to dynamically adjust the light projection direction of the target vehicle at the curve, extracts the geometric curvature of the tunnel curve, and uses the geometric curvature of the tunnel curve to perform spatial coordinate transformation on the synchronized trajectory sequence to generate curve adaptation trajectory data; A decision generation module, which integrates the curve adaptation trajectory data with the spatial position distribution characteristics of the target vehicle acquired in real time by the accident monitoring scanning equipment, and generates an accident upstream blind spot coverage instruction and linkage light warning parameters when the abnormal position concentration of the target vehicle exceeds a preset dynamic threshold; a posture control module, responding to the accident upstream blind spot coverage instruction, and dynamically adjusting the physical projection posture of the smart guidance beacon through the magnetorheological damper built into the smart guidance beacon to generate real-time projection posture parameters; The optical communication joint control module generates a real-time matching light control signal based on the linkage light warning parameters and the real-time projection posture parameters.

9. A computing device, characterized in that It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement an accident detection and early warning method based on intelligent induction targets as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, an accident detection and early warning method based on intelligent guidance marks as described in any one of claims 1 to 7 is implemented.

Images