An accident detection and early warning method and system based on intelligent induction marks

Abstract

The application provides an accident detection and early warning method and system based on intelligent induction signs. In the method, track space-time characteristic data is collected, dynamic time warping principle is used to align multi-vehicle track time sequence to generate synchronized sequence; at the same time, intelligent induction signs are driven to dynamically adjust the projection direction of the curve light, the tunnel design geometric curvature is extracted, and the track sequence is adapted to the spatial structure of the curve; the curve adapted track and the spatial distribution characteristics of the vehicle are fused, when the abnormal position aggregation degree exceeds the dynamic threshold, the accident upstream blind area coverage instruction and the linked light parameters are generated; after responding to the instruction, the real-time parameters are generated by dynamically adjusting the physical projection posture; based on the real-time matching of the light parameters and the posture parameters, the control signal is generated, and millisecond level cooperative early warning of the tunnel curve accident is realized. Through the cross-layer cooperation of track curvature adaptation, spatial aggregation degree determination and magnetorheological posture regulation, the dynamic light accurate coverage of the upstream blind area of the tunnel curve accident is realized.

Description

TECHNICAL FIELD

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

[0002] Due to the special spatial structure and driving environment of tunnel curves, there are multiple risks such as visual obstruction, vehicle centrifugal deviation and insufficient braking distance, and a real-time and accurate accident perception and dynamic induction mechanism needs to be established. The scene requires that the technical solution must break through three core bottlenecks: first, overcome the monitoring blind area caused by the curvature of the curve, realize no dead angle accident capture; second, dynamically adjust the warning coverage range according to the accident location, accurately guide the upstream vehicles; third, fuse multi-dimensional data such as vehicle behavior, road geometry and environmental state, complete the closed-loop response from accident identification to active intervention in a very short time, so as to block the secondary accident chain reaction.

[0003] The current mainstream technology adopts an event detection system based on computer vision to link a static induction device, through deploying intelligent cameras and millimeter wave radar fusion perception equipment on the side wall of the tunnel, using deep learning algorithm to identify vehicle abnormal behavior characteristics, and triggering the fixed angle induction sign at the preset position to perform light warning. This scheme realizes localized analysis and decision through edge computing equipment, avoiding the delay of the central platform forwarding, and achieving preliminary results in some straight tunnel sections.

[0004] However, the existing scheme still has fundamental defects in the curve accident scene: first, the fixedly installed induction sign cannot dynamically adjust the projection angle, resulting in a serious mismatch between the light coverage range and the spatial distribution of the curve blind area, and there are a large number of warning dead angles in the section with small curvature radius; second, the visual perception is interfered by the wall obstruction and light refraction in the curve, and the low-speed accident target close to the inner lane has a significantly increased missed detection rate, and the false triggering frequently occurs in rainy and foggy weather; finally, the system response link still relies on multi-stage software decision, and there is a delay window that is difficult to eliminate from event confirmation to hardware action, which cannot meet the time efficiency demand of three-second-level safety stop in the curve scene. These defects together cause the existing scheme to often have problems such as warning lag, coverage deviation and response inaccuracy in real curve accident disposal. SUMMARY

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

[0006] In a first aspect, the embodiments of the present application provide an accident detection and early warning method based on intelligent induction signs, comprising:

[0007] Obtaining trajectory spatiotemporal feature data of multiple target vehicles, the trajectory spatiotemporal feature data containing position, speed and acceleration information;

[0008] align the trajectory spatio-temporal feature data in time sequence by using the dynamic time warping principle to generate a synchronized trajectory sequence;

[0009] extract the tunnel curve geometric curvature by dynamically adjusting the light projection direction of the target vehicle at the curve by using the intelligent guidepost, and perform spatial coordinate transformation on the synchronized trajectory sequence by using the tunnel curve geometric curvature to generate curve-adapted trajectory data;

[0010] fuse the curve-adapted trajectory data with the spatial position distribution features of the target vehicle acquired in real time by the accident monitoring and scanning device, and generate an accident upstream blind area coverage instruction and a linked light warning parameter when the abnormal position aggregation degree of the target vehicle exceeds a preset dynamic threshold;

[0011] respond to the accident upstream blind area coverage instruction and dynamically adjust the physical projection posture of the intelligent guidepost by using the magneto-rheological damper built-in the intelligent guidepost to generate real-time projection posture parameters;

[0012] generate real-time matched light control signals based on the linked light warning parameters and the real-time projection posture parameters.

[0013] Optionally, responding to the accident upstream blind area coverage instruction and dynamically adjusting the physical projection posture of the intelligent guidepost by using the magneto-rheological damper built-in the intelligent guidepost to generate real-time projection posture parameters, comprises:

[0014] analyzing the spatial orientation information contained in the accident upstream blind area coverage instruction to convert it into posture adjustment requirement parameters, wherein the posture adjustment requirement parameters include target pitch angle value and target yaw angle value;

[0015] determining the input current intensity of the magneto-rheological damper according to the difference between the target pitch angle value and the current pitch angle value of the posture adjustment requirement parameters, and the difference between the target yaw angle value and the current yaw angle value;

[0016] applying the input current intensity to the magneto-rheological damper to make the magnetic particles inside the magneto-rheological damper form a chain structure under the action of the magnetic field, generating a controllable damping force positively correlated with the current intensity;

[0017] driving the support linkage mechanism connected to the base of the intelligent guidepost by the controllable damping force to make the support linkage mechanism produce angular displacement changes around the rotation axis, obtaining an angular displacement change amount;

[0018] convert the angular displacement change amount into real-time projection posture parameters in digital form, wherein the real-time projection posture parameters include actual pitch angle value and actual yaw angle value.

[0019] Optionally, the input current intensity is applied to the magneto-rheological damper, so that the magnetic particles inside the magneto-rheological damper form a chain structure under the action of the magnetic field, and a controllable damping force positively related to the current intensity is generated, comprising:

[0020] The input current intensity is transmitted to the multi-layer electromagnetic coil tightly wound on the surface of the annular inner cavity of the magneto-rheological damper through a wire;

[0021] The electromagnetic coil converts electrical energy into magnetic energy to generate a spatially uniform axial magnetic field in the annular inner cavity of the magneto-rheological damper;

[0022] Under the action of the axial magnetic field, the magnetic particles suspended in the magneto-rheological fluid overcome the Brownian motion force and migrate in the direction of the magnetic force line, and when the magnetic particles migrate to the region with the maximum magnetic flux density, the magnetic dipole moment attractive force between adjacent magnetic particles is generated, so that a plurality of particles are connected head to tail to form a columnar chain structure penetrating the fluid gap;

[0023] The opposite sides of the inner cavity of the magneto-rheological damper are bridged by the columnar chain structure, which mechanically hinders the fluid flow perpendicular to the direction of the chain structure;

[0024] Based on the viscous resistance principle of fluid mechanics, the mechanical hindering effect is converted into a controllable damping force linearly corresponding to the current intensity.

[0025] Optionally, the direction of light projection of the target vehicle at the curve is dynamically adjusted using the intelligent induction marker, the geometric curvature of the tunnel curve is extracted, and the spatial coordinate transformation is performed on the synchronized trajectory sequence using the geometric curvature of the tunnel curve to generate curve-adapted trajectory data, comprising:

[0026] According to the tunnel design drawings, the center line curvature radius and the pitch slope parameters are obtained to generate the geometric curvature characteristics of the tunnel curve;

[0027] Based on the geometric curvature characteristics, a curve spatial coordinate system is established, wherein the origin of the curve spatial coordinate system is located at the entry point of the curve, the horizontal axis is perpendicular to the road surface, and the vertical axis is along the tangent direction of the center line;

[0028] The vehicle position data in the synchronized trajectory sequence is mapped from the preset geodetic coordinate system to the curve spatial coordinate system to generate intermediate conversion positions;

[0029] According to the lateral offset law of the actual driving trajectory of the tunnel curve and the theoretical center line of the tunnel curve, the intermediate conversion positions are curvature-compensated and corrected to generate curve-adapted trajectory data.

[0030] Optionally, according to the actual driving track of the tunnel curve and the lateral offset rule of the theoretical center line of the tunnel curve, the intermediate conversion position is subjected to curvature compensation correction to generate a curve adaptive track data, including:

[0031] The vehicle track point set passing through the target curve in a continuous preset period is called from the tunnel curve monitoring system as the position data of the historical natural driving path;

[0032] The distance value perpendicular to the normal plane of the theoretical center line is calculated for each track point in the vehicle track point set to separate the pure lateral position offset component;

[0033] The tunnel curve is divided into statistical sections according to the equal interval of the arc length, and the pure lateral position offset component is aggregated in each section to extract the lateral offset distribution statistical characteristics under different curvature radii;

[0034] A curvature radius and compensation amount mapping table is constructed based on the lateral offset distribution statistical characteristics and stored in the local storage of the intelligent induction marker;

[0035] According to the curve arc length coordinate where the intermediate conversion position is located, the curvature radius index value of the corresponding statistical section is matched;

[0036] The target lateral compensation amount is obtained by querying the curvature radius and compensation amount mapping table through the curvature radius index value;

[0037] The target lateral compensation amount is superimposed on the lateral coordinate value of the intermediate conversion position, the longitudinal and elevation coordinates remain unchanged, and the curve adaptive track data after curvature compensation correction is generated.

[0038] Optionally, the support linkage mechanism connected to the base of the intelligent induction marker is driven by the controllable damping force to produce 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 end hinge joint of the driving arm of the support linkage mechanism to form a lever moment around the rotation axis at the end of the driving arm;

[0040] The driven arm connected to the driving arm is driven by the lever moment to slide along the arc-shaped guide rail fixed to the induction marker base;

[0041] The center of the arc-shaped guide rail is constrained to coincide with the rotation axis to ensure that the movement track of the end of the driven arm is an accurate circular arc path with constant radius, and a displacement marker point is arranged at the end of the driven arm to capture the arc length change data of the displacement marker point relative to the initial position in real time through a laser ranging sensor;

[0042] According to the geometric proportional relationship between the arc length change data and the radius of the arc-shaped guide rail, an angular displacement change amount of the support linkage mechanism is calculated.

[0043] Optionally, the curved track adaptive trajectory data is fused with the spatial position distribution characteristics of the target vehicle acquired by the accident monitoring scanning device in real time, and when the abnormal position aggregation degree of the target vehicle exceeds a preset dynamic threshold, an accident upstream blind area coverage instruction and a linked light warning parameter are generated, including:

[0044] The spatial partition units of the tunnel curve are divided, wherein the inner side area of the tunnel curve adopts a high-density unit division mode, and the outer side area of the tunnel curve adopts a low-density unit division mode;

[0045] The number of vehicles existing in the spatial partition units is scanned in real time, and a position aggregation degree basic quantity bound to the position of the spatial partition unit is generated;

[0046] The speed mutation abnormal point position identified in the curved track adaptive trajectory data is mapped to the corresponding spatial partition unit, and is marked as an accident risk core unit;

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

[0048] According to the traffic flow level fed back by the preset tunnel entrance vehicle flow detection device in real time, the judgment reference value of the comprehensive position aggregation degree index is self-adjusted, and when the comprehensive position aggregation degree index continuously exceeds the dynamically adjusted judgment reference value for a preset time window, an accident upstream blind area coverage instruction and a linked light warning parameter are generated synchronously.

[0049] In a second aspect, an accident detection and warning system based on intelligent induction signs is provided, including:

[0050] A data acquisition module acquires trajectory spatiotemporal characteristic data of multiple target vehicles, and the trajectory spatiotemporal characteristic data contains position, speed and acceleration information;

[0051] A time sequence synchronization module performs time sequence alignment processing on the trajectory spatiotemporal characteristic data by using a dynamic time warping principle to generate a synchronized trajectory sequence;

[0052] A curvature adaptive module adjusts the light projection direction of the target vehicle at a curve by using an intelligent induction sign, extracts the geometric curvature of a tunnel curve, and performs spatial coordinate transformation on the synchronized trajectory sequence by using the geometric curvature of the tunnel curve to generate curved track adaptive trajectory data;

[0053] a decision generation module, which fuses the bend-adaptive trajectory data and spatial position distribution features of the target vehicle acquired by the accident monitoring and scanning device in real time, and generates an accident upstream blind area coverage instruction and a linkage light warning parameter when the abnormal position aggregation degree of the target vehicle exceeds a preset dynamic threshold;

[0054] a posture control module, which responds to the accident upstream blind area coverage instruction, dynamically adjusts the physical projection posture of the intelligent induction marker through a magneto-rheological damper built in the intelligent induction marker, and generates real-time projection posture parameters;

[0055] a light signal linkage control module, which generates real-time matched light control signals based on the linkage light warning parameter and the real-time projection posture parameter.

[0056] In a third aspect, an embodiment of the present application provides a computing device, including 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, so as to realize the accident detection and warning method based on the intelligent induction marker as described in the first aspect.

[0057] In a fourth aspect, an embodiment of the present application provides a computer storage medium, which stores a computer program; when the computer program is executed by a computer, the accident detection and warning method based on the intelligent induction marker as described in the first aspect is realized.

[0058] In the embodiment of the present application, the trajectory spatiotemporal feature data of multiple target vehicles is acquired, and the trajectory spatiotemporal feature data contains position, speed and acceleration information; the dynamic time warping principle is used to perform time series alignment processing on the trajectory spatiotemporal feature data, to generate a synchronized trajectory sequence; the intelligent induction marker is used to dynamically adjust the light projection direction of the target vehicle at the bend, to extract the tunnel bend geometric curvature, and to perform spatial coordinate transformation on the synchronized trajectory sequence by using the tunnel bend geometric curvature, to generate bend-adaptive trajectory data; the bend-adaptive trajectory data is fused with the spatial position distribution features of the target vehicle acquired by the accident monitoring and scanning device in real time, and an accident upstream blind area coverage instruction and a linkage light warning parameter are generated when the abnormal position aggregation degree of the target vehicle exceeds a preset dynamic threshold; the accident upstream blind area coverage instruction is responded to, and the physical projection posture of the intelligent induction marker is dynamically adjusted through a magneto-rheological damper built in the intelligent induction marker, to generate real-time projection posture parameters; and real-time matched light control signals are generated based on the linkage light warning parameter and the real-time projection posture parameter.

[0059] The technical scheme of the application has the following beneficial effects: the problem of original behavior data source of multi-target vehicles is solved, and key feature inputs such as speed mutation are provided for accident judgment. Phase deviation caused by sampling time difference of multi-vehicle trajectories is eliminated, and time-synchronized trajectory sequences are generated to improve accident positioning accuracy. The trajectory data adapted to the physical environment is generated by dynamically adjusting the light projection direction to match the curve direction and simultaneously using the tunnel geometric curvature to correct the trajectory space distortion. The accident authenticity is verified in combination with the position distribution characteristics, and the generation instruction and parameters are dynamically decided based on the aggregation degree threshold, thereby reducing the false alarm rate. Millisecond-level dynamic adjustment of the physical projection posture of the guide sign is realized, and the hardware limitation of fixed angle is broken through. The light signal is ensured to be matched with the physical posture in real time, and the collaborative response of blind area coverage and warning enhancement is achieved.

[0060] Further, the target posture parameters are generated by analyzing the blind area instruction, the current intensity of the magnetorheological damper is calculated according to the angle difference, the magnetic particles are driven to form a chain structure to generate controllable damping force, the force acts on the support connecting rod mechanism to convert into rotational angular displacement, and finally the digital real-time projection posture parameters are output.

[0061] The magnetorheological effect is combined with the mechanical transmission mechanism to realize instantaneous response adjustment of the physical posture of the guide sign, and the delay bottleneck of the traditional actuator is completely broken through. The direct drive mechanism based on the current-angle difference mapping eliminates the closed-loop feedback link to achieve accurate angle output without cumulative error. The self-stabilizing damping field formed by the chain structure of the magnetic particles effectively suppresses the posture drift caused by vibration in the tunnel environment. The coordinated adjustment mechanism of the pitch angle and the yaw angle adaptively matches the upstream blind area spatial distribution characteristics of the curve with different curvatures.

[0062] These aspects or other aspects of the application will be more apparent in the following description of the embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0063] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0064] Figure 1 A flowchart of an accident detection and warning method based on an intelligent guide sign is shown;

[0065] Figure 2 A structural schematic diagram of an accident detection and warning system based on an intelligent guide sign is shown;

[0066] Figure 3 A structural schematic diagram of a computing device is shown. DETAILED DESCRIPTION

[0067] In order to better understand the scheme of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in combination with the accompanying drawings in the embodiments of the present application.

[0068] In some processes described in the specification and claims of the present application and the above-mentioned drawings, a plurality of operations appearing in a specific order are included, but it should be clearly understood that these operations can be executed or in parallel without the order in which they appear in this text, and the serial numbers of the operations such as 101, 102, etc. are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes can include more or fewer operations, and these operations can be executed in sequence or in parallel. It should be noted that the descriptions such as “first”, “second” and the like in this text are used to distinguish different messages, devices, modules, etc., and do not represent the sequence, nor limit the types of “first” and “second”.

[0069] Research finds that the tunnel curve accident warning technology is limited by static hardware architecture and centralized decision mechanism, and has fundamental application defects: the fixed-angle induction device cannot adapt to the dynamic blind area distribution caused by the change of curve curvature, resulting in a large number of dead angles in the upstream warning coverage of the accident; the video sensing means is affected by wall obstruction and light interference, and the missing detection rate of low-speed targets in the inner lane is high; the multi-level software decision link causes response delay, making the warning action lag behind the safety braking time window of the curve, and the secondary accident blocking efficiency is low. These defects are collectively caused by the lack of capabilities in the three dimensions of physical layer dynamic regulation, spatial layer accurate perception and timing layer real-time collaboration in the existing scheme, which is difficult to meet the closed-loop needs of accident disposal in the curve scene.

[0070] In view of the above problems, the present application proposes an accident detection and warning method based on intelligent induction sign, and specifically, the innovation core is: through the magnetic fluid damper to realize the millisecond-level dynamic adjustment of the physical projection posture of the induction sign, combined with the trajectory space transformation driven by the tunnel curvature and the vehicle aggregation degree fusion judgment mechanism, to build a “perception-decision-execution” integrated closed loop. Specifically, the geometric curvature of the curve is used to correct the distortion of the vehicle trajectory space, and the position distribution characteristics of the scanning device are synchronously fused to accurately position the accident; based on the positioning result, a blind area coverage instruction is generated to drive the magnetic fluid damper to adjust the pitch and yaw angle of the induction sign in real time; finally, through the linkage control of the posture parameters and the light signal, the warning range is realized. The adaptive coverage of the blind area upstream of the accident. This method fundamentally solves the coverage mismatch problem of static induction, eliminates system delay through physical layer direct drive control, and overcomes the perception blind area through curvature space mapping, forming a complete prevention and control closed loop for the tunnel curve accident scene.

[0071] With reference to the drawings and the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work are within the scope of protection of the present application.

[0072] Figure 1 A flow chart of an accident detection and early warning method based on intelligent induction marks is provided for the embodiments of the present application, as shown in Figure 1 The method comprises the following steps.

[0073] 101. Obtain trajectory space-time feature data of multiple target vehicles, wherein the trajectory space-time feature data comprises position, speed and acceleration information.

[0074] In this step, the trajectory space-time feature data refers to a dynamic data set comprising vehicle position information, i.e., spatial coordinates, speed information, i.e., instantaneous motion rate, and acceleration information, i.e., speed change rate.

[0075]

[0076] The data is collected by fusion of millimeter wave radars and video sensors arranged on the tunnel wall.

[0077] In the embodiments of the present application, first, the sensing device fixedly installed on the tunnel wall emits a detection signal to capture the real-time spatial coordinates of the vehicle; second, the continuous frame image analysis technology is used to extract the vehicle moving speed; third, the instantaneous acceleration is calculated by comprehensively considering the spatial displacement and time interval; and finally, the position, speed and acceleration information is integrated according to the vehicle target identifier to form a trajectory space-time feature data set.

[0078] In an actual case, at the entrance of a tunnel curve, the sensing device detects the spatial coordinate changes of multiple vehicles, determines the motion speed and acceleration trend by combining video analysis, and forms a feature data packet comprising three-dimensional position, motion speed and acceleration direction.

[0079] 102. Perform time sequence alignment processing on the trajectory space-time feature data by using the dynamic time warping principle to generate a synchronized trajectory sequence.

[0080] In this step, the synchronized trajectory sequence refers to a time sequence alignment trajectory set formed after eliminating the time stamp deviation of the multi-vehicle trajectories.

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

[0082] ​In the embodiment of the present application, first, the phase offset of different vehicle trajectory time axes is identified; second, a nonlinear time calibration method is used to perform elastic matching on the trajectory points; then, the sampling time difference of the multi-vehicle trajectory is eliminated through time axis stretching and shrinking; finally, a synchronized trajectory sequence with unified time reference is output, ensuring that all vehicle behaviors are in the same time reference system.

[0083] Continuing the above case, for the vehicle driving in the curve, the system detects that the trajectory timestamp of the vehicle is asynchronous, and through dynamic calibration, the sudden braking time of different vehicles is aligned to the same time node, forming a synchronized trajectory sequence that can be compared and analyzed.

[0084] 103. The target vehicle's light projection direction at the curve is dynamically adjusted by the intelligent induction sign, the tunnel curve geometry curvature is extracted, and the synchronized trajectory sequence is subjected to spatial coordinate transformation using the tunnel curve geometry curvature, to generate curve-adapted trajectory data;

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

[0086] The tunnel curve geometry curvature refers to a geometric feature parameter composed of the center line turning radius and the longitudinal slope.

[0087] In the embodiment of the present application, first, the induction sign rotating mechanism is controlled to project light along the tangent direction of the curve center line; second, the center line curvature parameter is extracted from the tunnel structure database; third, a local spatial coordinate system with the curve geometric center as the origin is constructed; finally, the geodetic coordinates of the synchronized trajectory are mapped to the local coordinate system, eliminating the trajectory distortion caused by the curve curvature.

[0088] Continuing the above case, in the curve section with significant curvature change, the induction sign automatically adjusts the projection direction to the curve extension direction, and at the same time, the vehicle trajectory coordinates are converted to the curve local spatial coordinate system, correcting the measurement position deviation caused by the centrifugal phenomenon.

[0089] 104. The curve-adapted trajectory data is fused with the spatial position distribution characteristics of the target vehicle obtained by the accident monitoring scanning device in real time, and when the abnormal position aggregation degree of the target vehicle exceeds the preset dynamic threshold, an accident upstream blind area coverage instruction and a linked light warning parameter are generated;

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

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

[0092] In the embodiment of the present application, first, the spatial distribution topology of the vehicle group is obtained by a three-dimensional scanning device; second, the curved area is divided into equal-area analysis units; third, the core unit where the speed mutation point is located is positioned; then, the vehicle density weighted value of the area around the core unit is calculated; finally, the real-time vehicle flow density is dynamically adjusted to determine the threshold value, and when the weighted density exceeds the threshold value, an accident instruction and a red high-frequency flashing warning parameter are generated.

[0093] Continuing the above case, when a vehicle suddenly reduces speed inside the curve, the system detects that the vehicle distribution density in the area is abnormally high, and determines that it is an accident in combination with the current vehicle flow state, generates a coverage instruction pointing to the blind area upstream of the accident and a red high-frequency flashing warning parameter.

[0094] 105, in response to the accident upstream blind area coverage instruction, and dynamically adjusting the physical projection posture of the intelligent induction mark through the magneto-rheological damper built-in the intelligent induction mark, generating real-time projection posture parameters;

[0095] In this step, the real-time projection posture parameter refers to the digital expression of the spatial angle state of the induction mark;

[0096] The magneto-rheological damper refers to an actuator that controls the viscosity of the fluid to output damping force through a magnetic field.

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

[0098] Continuing the above case, the system calculates the pitch and yaw angles that the induction mark needs to adjust according to the position upstream of the accident, and the magneto-rheological damper drives the linkage mechanism to rotate under the action of the current, so that the optical components of the induction mark are accurately turned to the target area.

[0099] 106, based on the linkage light warning parameter and the real-time projection posture parameter, a real-time matching light control signal is generated.

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

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

[0102] Continuing the above case, the control module fuses the red high-frequency flickering instruction with the real-time angle parameter to generate a current pulse with a specific duty cycle, which drives the LED array to project a red warning light beam to the corrected spatial orientation.

[0103] In summary, steps 101 to 106 eliminate the measurement time difference through spatiotemporal trajectory synchronization, correct the position distortion using the curve curvature mapping, dynamically determine the accident location in combination with the vehicle distribution density, and achieve millisecond-level posture regulation of the induction marker by means of the magnetorheological mechanism, ultimately achieving precise spatial matching of the light warning and the accident blind area. The whole process breaks through the coverage limitations of static induction devices, overcomes the environmental perception blind area, establishes an efficient closed loop from accident identification to active intervention, and significantly improves the tunnel curve driving safety factor.

[0104] To solve the problems of posture adjustment delay and insufficient accuracy of the induction marker in tunnel curve accident warning, in some embodiments, in step 105, in response to the blind area coverage instruction upstream of the accident, the physical projection posture of the intelligent induction marker is dynamically adjusted through the magnetorheological damper built-in the intelligent induction marker, and real-time projection posture parameters are generated, including:

[0105] 201, analyze the spatial orientation information contained in the blind area coverage instruction upstream of the accident, and convert it into posture adjustment demand parameters, wherein the posture adjustment demand parameters include target pitch angle value and target yaw angle value;

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

[0107] The posture adjustment demand parameter refers to the target angle set that quantifies the spatial pointing of the induction marker, including the target pitch angle value, i.e., the elevation angle of the light beam in the vertical plane;

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

[0109] In the embodiments of the present application, first, the spatial coordinate data in the blind area coverage instruction is read; second, the orientation-angle mapping relationship is established according to the tunnel center line curvature model; third, the blind area center point coordinates are converted into target values of pitch angle and yaw angle; and finally, the posture adjustment demand parameters containing the double-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 posture adjustment demand 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 that drives the magnetorheological damper to work, which is in a linear proportional relationship with the angle difference.

[0112] In the embodiment of the present application, first, the target angle value in the posture adjustment requirement parameter is acquired; second, the current actual angle value fed back by the angle sensor is read; then, the algebraic difference values of the target value and the current value of the pitch angle and the yaw angle are respectively calculated; finally, the angle difference value is mapped to the current intensity value according to the preset proportion coefficient.

[0113] 203. Apply the input current intensity to the magneto-rheological damper, so that the magnetic particles inside the magneto-rheological damper form a chain structure under the action of the magnetic field, and a controllable damping force that is positively correlated with the current intensity is generated;

[0114] In step 203, the chain structure refers to the columnar arrangement formed by the magnetic particles under the action of the magnetic field and connected head to tail along the magnetic force line;

[0115] The controllable damping force refers to the mechanical resistance output linearly adjusted by changing the current intensity.

[0116] In the embodiment of the present application, first, the current intensity is input to the electromagnetic coil wound on the inner wall of the damper; second, the energized coil generates an axial uniform magnetic field; then the magnetic field drives the ferromagnetic particles suspended in the magneto-rheological fluid to overcome the fluid resistance and migrate directionally; then the particles are connected head to tail under the attraction of the magnetic dipole moment to form a chain 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 the support linkage mechanism connected to the intelligent induction mark base through the controllable damping force, so that the support linkage mechanism generates an angular displacement change around the rotation axis, and an angular displacement change amount is obtained;

[0118] In step 204, the angular displacement change amount refers to the angular change value of the support linkage mechanism rotating around the fixed axis, which is generated by converting the linear damping force through the lever mechanism.

[0119] In the embodiment of the present application, first, the controllable damping force is transmitted to the driven arm hinge point of the support linkage mechanism; second, the damping force forms a lever moment around the rotation axis at the end of the driving arm; then the moment drives the driven arm to slide along the arc-shaped 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 amount on the circular arc path is recorded through the displacement marker point.

[0120] 205. Convert the angular displacement change amount into a digital form of real-time projection posture parameter, wherein the real-time projection posture parameter includes 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 space angle of the induction mark optical assembly.

[0122] In the embodiments of the present application, first, the displacement arc length of the support connecting rod end is measured by the laser ranging device; second, the rotation angle value is calculated according to the fixed radius of the arc guide rail; then, the mechanical rotation angle is decomposed into pitch and yaw components; finally, the digital parameters including the actual pitch angle value and the actual yaw angle value are output.

[0123] The following is a specific example:

[0124] An accident occurs in a certain tunnel curve, and the system generates instructions covering the upstream blind area. Step 201 analyzes the instruction space coordinates 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 pitch of 3 degrees and yaw of 20 degrees, determines that 2 degrees of pitch and 5 degrees of yaw need to be added, and maps them into specific current values. Step 203 applies current to the magnetorheological damper, and the magnetic field drives the particles to form a chain structure to output corresponding damping force. In step 204, the damping force drives the connecting rod mechanism, causing the driven arm to rotate along the guide rail and produce an angle change. In step 205, the displacement arc length is captured by the laser ranging, and the actual pitch of 4.8 degrees and the actual yaw of 24.9 degrees are calculated by combining the guide rail radius, and the digital attitude parameters are output.

[0125] In summary, steps 201 to 205 realize the precise and rapid regulation and control of the physical attitude of the induction marker through five steps of instruction analysis, current mapping, magnetorheological response, mechanical conversion, and digital measurement. The chain structure formation mechanism driven by the magnetic field breaks through the inertia delay bottleneck of traditional actuators, the mechanical design of the support connecting rod and the arc guide rail guarantees zero backlash in angle conversion, and the non-contact measurement of the laser eliminates mechanical wear errors. The whole process dynamically matches the spatial distribution of the blind area of the curve at a response speed of milliseconds, fundamentally solving the defects of inaccurate coverage and response lag of static induction devices, and significantly improving the secondary accident control efficiency.

[0126] In order to break through the bottleneck of response delay and precision decay of traditional actuators in a tunnel vibration environment, in some embodiments, in step 203, the input current intensity is applied to the magnetorheological damper, causing the magnetic particles inside the magnetorheological damper to form a chain structure under the action of the magnetic field, generating a controllable damping force that is positively related to the current intensity, including:

[0127] 301. transmitting the input current intensity through a wire to a plurality of electromagnetic coils tightly wound on the surface of the annular cavity of the magnetorheological damper;

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

[0129] The plurality of electromagnetic coils refers to a plurality of turns of conductive windings tightly wound in a spiral manner on the surface of the annular cavity, used to establish a closed magnetic circuit.

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

[0131] 302. convert the 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 field line direction is parallel to the central axis of the damper;

[0133] Spatially uniform distribution means that the gradient change of the 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 turn pitch of the multi-layer coil; then the soft magnetic material cavity wall is used to constrain the magnetic field line direction; finally, a stable magnetic field environment uniformly distributed along the axial direction is formed.

[0135] 303. Under the action of the axial magnetic field, the magnetic particles suspended in the magnetorheological fluid overcome the Brownian motion force and migrate along the magnetic field line direction, and when the magnetic particles migrate to the region with the maximum magnetic flux density, the magnetic dipole attraction force between adjacent magnetic particles is generated, so that a plurality of particles are connected head to tail to form a columnar chain structure penetrating through the fluid gap;

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

[0137] The magnetic dipole attraction force refers to the mutual attraction between the north and south poles of the magnetized particles.

[0138] In the embodiment of the present application, first, the ferromagnetic particles in the magnetorheological fluid are subjected to magnetization force in the axial uniform magnetic field environment; secondly, the action force overcomes the random collision disturbance generated by the fluid molecular thermal motion; then the particles are driven by the magnetic field gradient to make directional displacement and gather in the region with the maximum magnetic field intensity; then the adjacent particles are attracted to each other due to the north and south magnetic poles generated by magnetization; finally, the particles are connected head to tail along the magnetic field line direction to form a columnar chain structure penetrating through the fluid gap.

[0139] 304. Bridge the opposite sides of the inner cavity of the magnetorheological damper through the columnar chain structure to produce a mechanical hindering effect on the fluid flow perpendicular to the direction of the chain structure;

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

[0141] Bridge refers to the mechanical state of anchoring the chain structure at both ends to the inner wall of the cavity.

[0142] In the embodiments of the present application, first, the columnar chain structure spans the inner cavity of the damper; second, the chain body is embedded in the preset micro groove of the cavity wall; then the middle part of the chain body is suspended in the magnetorheological fluid; then when the fluid flows vertically to the chain body; finally, the chain structure forms a physical barrier to hinder the fluid from passing through.

[0143] 305, based on the principle of fluid mechanics viscous resistance, the mechanical hindering effect is converted into a controllable damping force linearly corresponding to the current intensity.

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

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

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

[0147] The following is a specific example:

[0148] When the system needs to enhance the damping force, a specific current value is input through the wire to the multi-layer electromagnetic coil. The coil establishes an axially uniform magnetic field in the annular cavity, the magnetic field drives the ferromagnetic particles to migrate to the central region, and the particles are connected head to tail to form a columnar chain body under the attraction of the magnetic poles. The chain body is anchored at both ends to the cavity wall, and the suspended chain segment in the middle hinders the vertical flow of the fluid, and the viscous resistance is generated by the shear deformation of the fluid. This resistance increases linearly with the increase of the input current, and finally outputs the controllable damping force that meets the requirements.

[0149] In summary, steps 301 to 305 realize precise and controllable adjustment of damping force through the hierarchical progressive mechanism of current conduction, magnetic field construction, particle self-assembly, structure bridging and viscous resistance conversion. Among them, the axially uniform magnetic field ensures the consistency of the particle migration direction, the magnetic dipole moment attraction realizes the self-organization of the chain structure, the bridge design strengthens the stability of the mechanical hindering, and the viscous resistance conversion ensures the linearity of the output. The whole process breaks through the temperature sensitivity and mechanical hysteresis defects of traditional hydraulic dampers, and still maintains the millisecond level response characteristics in the tunnel vibration environment, providing high reliable execution power for the induction of attitude control.

[0150] In order to overcome the distortion of vehicle trajectory measurement and the inaccuracy of induced control in a curved environment, in some embodiments, the direction of light projection of the target vehicle at the curve is dynamically adjusted by using the intelligent induced marker in step 103, the geometric curvature of the tunnel curve is extracted, and the spatial coordinate transformation is performed on the synchronized trajectory sequence by using the geometric curvature of the tunnel curve, to generate the curve adaptive trajectory data, including:

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

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

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

[0154] The geometric curvature characteristic refers to the three-dimensional space bending measurement constructed by the curvature radius and the slope.

[0155] In the embodiments of the present application, first, the digital design drawings of the target curve segment are called through the tunnel engineering database interface; second, the center line geometric parameter annotation layer is positioned in the drawing structure tree; then the curvature radius design value and the slope percentage data in this layer are analyzed; finally, the two are packaged as geometric curvature characteristic data packets representing the three-dimensional space bending characteristics of the curve. The parameter validity needs to be verified and the historical version data before the construction change is excluded in the process.

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

[0157] In step 402, the curve space coordinate system refers to a rectangular coordinate system with local geometric characteristics of the curve as the reference;

[0158] The entry cut point refers to the junction of the starting end of the curve and the straight line segment;

[0159] The center line tangent direction refers to the instantaneous extension direction of the center line at this point.

[0160] In the embodiments of the present application, first, the mathematical cut point space coordinates of the straight line segment and the arc segment are calculated according to the geometric curvature characteristics; second, the coordinate system is initialized with the cut point as the origin; then the road surface normal vector at this point is obtained as the vertical axis reference direction; finally, the longitudinal axis positive direction is determined along the center line advancing tangent direction. After the coordinate system is constructed, the origin coordinates and the axial unit vector are stored for subsequent calling.

[0161] 403. Map the vehicle position data in the synchronized trajectory sequence from the preset geodetic coordinate system to the curve space coordinate system to generate intermediate conversion positions;

[0162] In step 403, the geodetic coordinate system refers to a global coordinate system taking the earth ellipsoid as a reference;

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

[0164] In the embodiment of the application, firstly, the vehicle geodetic coordinates (longitude, latitude, and elevation) in the synchronized trajectory sequence are read; secondly, a rotation transformation matrix is calculated according to the spatial position relationship of the double coordinate systems, the matrix containing the original point offset and the axial rotation angle; then the vehicle geodetic coordinates are substituted into the matrix operation formula; finally, a two-dimensional plane coordinate set under the local coordinate system of the curve is output, that is, the intermediate conversion position.

[0165] 404. According to the actual driving trajectory of the tunnel curve and the lateral offset rule of the theoretical center line of the tunnel curve, the intermediate conversion position is corrected for curvature compensation to generate a curve adaptive trajectory data.

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

[0167] The curvature compensation correction refers to the reverse calibration of the coordinate value according to the offset rule.

[0168] In the embodiment of the application, firstly, the historical vehicle trajectory data set of the curve is called from the traffic data center; secondly, the perpendicular distance of each trajectory point to the theoretical center line, that is, the lateral offset, is calculated; then the statistical distribution rule of the offset in different curvature sections is analyzed to establish a mapping model; finally, the compensation value is obtained by retrieving the mapping model according to the real-time curvature of the current curve, and the lateral coordinate of the intermediate conversion position is algebraically superimposed and corrected to generate the final adaptive trajectory.

[0169] The following is a specific example:

[0170] In a certain right-turn curve accident warning, the system obtains the curvature radius of 600 meters and the slope of 3 degrees from the design drawing. A coordinate system is established with the entry point of the curve as the origin, the horizontal axis perpendicular to the road surface, and the vertical axis pointing to the extension direction of the curve. The vehicle GPS coordinates in the synchronized trajectory are converted to the coordinate system to generate intermediate position data. Based on the statistical data of historical data, it is found that the vehicle is averagely offset outward by 1.2 meters under this curvature, and accordingly the lateral coordinate value of the intermediate position is increased by 1.2 meters of compensation, and finally the corrected curve adaptive trajectory data is output.

[0171] In summary, steps 401 to 404 achieve precise matching of vehicle trajectory and physical structure of the curve through the four-step cooperation of design parameter extraction, local coordinate system construction, spatial coordinate mapping, and behavior compensation correction. The curvature feature driven coordinate system establishment ensures spatial reference consistency, and the historical deviation rule compensation eliminates the measurement deviation caused by the centrifugal effect. The whole process breaks through the application limitations of the geodetic coordinate system in the curve scene, provides trajectory data basis for geometric distortion correction for accident positioning, and fundamentally solves the spatial adaptation distortion problem of the traditional straight trajectory model in the curve environment.

[0172] To solve the problem of vehicle trajectory measurement distortion caused by centrifugal effect in the tunnel curve environment, in some embodiments, in step 404, the curvature compensation correction is performed on the intermediate conversion position according to the lateral deviation rule of the actual driving trajectory of the tunnel curve and the theoretical center line of the tunnel curve, and the curve adaptation trajectory data is generated, including:

[0173] 501. Retrieve the vehicle trajectory point set through the target curve in a continuous preset period from the tunnel curve monitoring system as the position data of the historical natural driving path;

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

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

[0176] In the embodiments of the present application, first, an encrypted data request is initiated to the tunnel central monitoring system; second, the system selects the trajectory point set of the target curve in a continuous complete natural day; third, the positioning drift points and abnormal stationary points are filtered; and finally, the vehicle identity identifiers are classified and packaged for transmission to the edge computing unit of the guide marker.

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

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

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

[0180] In the embodiments of the present application, first, the vertical projection point of each trajectory point on the theoretical center line is calculated; second, the tangent direction vector of the center line at the projection point is obtained; third, the spatial normal plane perpendicular to the tangent is constructed; and finally, the shortest Euclidean distance of the trajectory point to the normal plane is measured as the pure lateral position deviation component.

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

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

[0183] In this embodiment, the arc length of the curve centerline is first divided into continuous segments of fixed length; then, each trajectory point is assigned to the corresponding segment according to its arc length coordinates; next, 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. Construct a mapping table of curvature radius and compensation amount based on the statistical characteristics of the lateral offset distribution, and store it in the local memory of the intelligent guide target;

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

[0186] In this embodiment, the variation law of the average offset of each section is first analyzed; then the compensation direction is determined according to the geometric characteristics of the curve (positive for external offset / negative for internal offset); next, the correspondence between the curvature radius index key and the compensation amount is established; finally, the mapping table is burned to the designated sector of the local flash memory chip of the induction target.

[0187] 505. Based on the arc length coordinates of the curve where the intermediate transition position is located, match the curvature radius index value of the corresponding statistical segment;

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

[0189] In this embodiment, the arc length coordinates of the intermediate transformation position are first parsed; then the pre-stored arc length-curvature correspondence tree is queried; next, the curvature radius value of the segment where the current arc length is located is located; and finally, it is converted into the address offset of the mapping table storage.

[0190] 506. By querying the curvature radius and compensation amount mapping table through the curvature radius index value, the target lateral compensation amount is obtained;

[0191] In step 506, the target lateral compensation amount refers to the specific correction value that needs to be superimposed on the lateral coordinate.

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

[0193] 507. The target lateral compensation amount is superimposed on the lateral coordinate value of the intermediate transformation position, while keeping the longitudinal and elevation coordinates unchanged, to generate the curvature compensation corrected curve adaptation trajectory data.

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

[0195] In this embodiment, the longitudinal and elevation coordinates of the intermediate transformation position are first copied; then the sign and magnitude of the target lateral compensation amount are read; next, an algebraic superposition operation is performed on the original lateral coordinates; finally, new lateral coordinate values ​​are generated and recombined into three-dimensional spatial coordinates.

[0196] Here is a specific example:

[0197] In tunnel curves with varying curvature, the system retrieves 30 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. A mapping table is constructed based on the statistical results and stored in the guidance benchmark; for example, a curvature radius of 500 meters corresponds to a 0.8-meter external offset compensation. When processing an intermediate transition position, the current curvature radius of 500 meters is matched based on its arc length coordinates, and the 0.8-meter compensation is obtained by querying the mapping table. This compensation is then superimposed on the lateral coordinates to finally generate the corrected and adapted trajectory data.

[0198] In summary, steps 501 to 507, through a seven-step closed loop—historical trajectory analysis, geometric projection calculation, statistical feature extraction, local mapping table construction, real-time curvature matching, compensation amount query, and coordinate correction—achieve adaptive calibration of vehicle trajectory curvature in curves. Specifically, the normal plane distance calculation eliminates residual errors from coordinate transformation, arc length partitioning statistics capture the centrifugal effect, and the localized mapping table ensures millisecond-level compensation response. This entire process overcomes the physical limitations of satellite positioning in curve scenarios, providing a data foundation consistent with real-world driving trajectories for accident detection, and fundamentally solving the positioning distortion problem caused by traditional trajectory models neglecting driving behavior characteristics.

[0199] To address the issue of wear and inaccuracy in mechanical angle measurements under tunnel vibration environments, in some embodiments, step 204 involves driving a support linkage mechanism connected to the intelligent guide beacon base using the controllable damping force. This causes the support linkage mechanism to undergo angular displacement around its rotation axis, yielding the amount of angular displacement change, including:

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

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

[0202] In the embodiment of the application, firstly, the controllable damping force is transmitted to the driven arm front end hinge hole of the support linkage mechanism through the spherical hinge joint; secondly, the damping force is conducted to the end along the rigid rod of the driving arm; then the force arm is formed between the end of the driving arm and the rotation axis; finally, the lever principle is used to generate the moment value of driving the rotation of the driven arm.

[0203] 602, using the lever moment to drive the driven arm connected with the driving arm, so that the driven arm slides along the arc-shaped guide rail fixed to the induction mark base;

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

[0205] In the embodiment of the application, firstly, the lever moment acts on the driven arm base; secondly, the moment overcomes the static friction of the guide rail to start sliding; then the roller at the end of the driven arm is embedded in the guide rail groove; finally, the roller makes circular motion along the groove track, driving the whole driven arm to rotate around the shaft.

[0206] 603, the center of the arc-shaped guide rail is constrained to coincide with the rotation axis, ensuring that the movement track of the end of the driven arm is an accurate circular arc path with a constant radius, and a displacement marker point is arranged at the end of the driven arm, and the arc length change data of the displacement marker point relative to the initial position is captured in real time by a laser ranging sensor;

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

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

[0209] 604, according to the geometric proportional relationship between the arc length change data and the radius of the arc-shaped guide rail, the angular displacement change amount of the support linkage mechanism is calculated.

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

[0211] In the embodiment of the application, firstly, the fixed radius value of the arc-shaped guide rail is 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; finally, the calculation result is the angular displacement radian value of the support linkage mechanism rotating around the shaft.

[0212] Here is a specific example:

[0213] When the magnetorheological damper outputs a specific damping force, this force is transmitted to the front end of the drive arm via a ball joint. A lever torque is generated at the end of the drive arm to drive the driven arm, and a roller at the end of the driven arm slides along an arc-shaped guide rail. The guide rail is machined with precision to ensure that its center coincides with the axis of rotation, making the roller's trajectory a precise circular arc. A ceramic marker ball mounted on the roller reflects the laser beam, and a sensor measures the increase in arc length. Given a fixed guide rail radius, the change in angular displacement is obtained by dividing the arc length increment by the radius, completing the entire force-angle conversion process.

[0214] In summary, steps 601 to 604 achieve a 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 arc-shaped guide rail and the rotation axis eliminates motion trajectory deviation, and non-contact laser measurement overcomes the wear limitations of traditional potentiometers. The entire process constructs a purely mechanical solution that maintains micro-radian-level measurement accuracy even under strong vibration environments, providing stable and reliable angle feedback for guide beacon attitude control and completely resolving the control inaccuracy problem caused by clearance wear in traditional linkage mechanisms.

[0215] To address the high misjudgment rate in tunnel curve accident detection due to vehicle sparse distribution and environmental interference, some embodiments integrate the curve adaptation trajectory data with the spatial location distribution characteristics of the target vehicles acquired in real time by the accident monitoring scanning device in step 104. When the abnormal location clustering of the target vehicles exceeds a preset dynamic threshold, an upstream blind zone coverage command and linked lighting warning parameters are generated, including:

[0216] 701. Divide the spatial partitioning units of the tunnel curve, wherein the inner region of the tunnel curve adopts a high-density unit partitioning mode, and the outer region of the tunnel curve adopts a low-density unit partitioning mode.

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

[0218] Low-density unit partitioning refers to a partitioning method where the number of partitioned units per unit area is lower than the average.

[0219] In this embodiment, the partition density boundary line is first determined based on the curvature gradient of the curve; then, small-area cells are continuously covered in the inner region of the curve; next, large-area cells are sparsely distributed in the outer region of the curve; finally, a non-uniform mesh topology structure is generated and stored.

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

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

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

[0223] 703. Map the locations of speed change anomaly points identified in the curve adaptation trajectory data to the corresponding spatial partition units and mark them as core accident risk units;

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

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

[0226] 704. Taking the core unit of accident risk as the geometric center, a multi-level radial expansion is carried out to the surrounding adjacent units to form a tiered detection area. The distance-attenuation type weighted aggregation operation is performed on the basic quantity of location clustering in the tiered detection area to generate a comprehensive location clustering index.

[0227] In step 704, the tiered detection area refers to a multi-layered 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 this embodiment, a three-level radiation layer is first established with the core unit of accident risk as the center; then, the core layer is set to have the highest weight coefficient, and the outer layers decrease step by step; next, the weighted sum of the basic quantity of unit location clustering in each layer is calculated; finally, the weighted sums of each level are accumulated to generate a comprehensive location clustering index.

[0230] 705. Based on the traffic flow level fed back in real time by the preset tunnel entrance vehicle flow detection device, the judgment benchmark value of the comprehensive location aggregation index is automatically adjusted. When the comprehensive location aggregation index continuously exceeds the dynamically adjusted judgment benchmark value for a preset time window, an upstream blind spot coverage instruction and linkage lighting warning parameters are generated simultaneously.

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

[0232] The preset time window refers to the minimum length of time during which the judgment index continuously exceeds the limit.

[0233] In the embodiments of the present application, first, the real-time traffic level signal sent by the tunnel entrance vehicle detector is received; second, the threshold curve corresponding to the traffic level index is pre-stored; then, the corresponding judgment reference value is dynamically loaded; finally, the continuous over-limit time of the comprehensive index is monitored, and the instruction is generated after the preset window is reached.

[0234] The following is a specific example:

[0235] In a right curve of a tunnel with variable curvature, the system divides the inner side area into high-density units of five square meters and the outer side into low-density units of ten square meters. Laser scanning shows that there are three vehicles in a certain high-density unit, which are bound as the position aggregation degree basic quantity. At the same time, it is detected that the vehicle speed in the unit decreases from eighty to five, which is marked as the accident risk core unit. A three-level detection circle is established around it, with the core circle weight being 1, the middle circle 0.6, and the outer circle 0.3. The comprehensive index value is obtained by weighted aggregation of the basic quantities of the surrounding units. According to the corresponding threshold loaded in the entrance traffic, when the comprehensive index continuously exceeds the limit for 10 seconds, a red high-frequency flashing instruction covering the 150-meter blind area upstream is generated.

[0236] In summary, steps 701 to 705 realize accurate and reliable identification of curve accidents through the five-level cooperation of non-uniform spatial partitioning, real-time vehicle statistics, risk core positioning, hierarchical weighted aggregation, and dynamic threshold judgment. Among them, the curvature-driven density partitioning enhances the sensitivity of the inner side, the three-level radiation detection breaks through the limitations of local vision, and the traffic self-adjusting mechanism avoids congestion misjudgment. The whole process builds a fusion decision-making model of spatial distribution and motion characteristics, significantly improves the accident detection confidence, and fundamentally solves the adaptability defects of traditional single-point detection in complex curve scenes.

[0237] Figure 2 A structural diagram of an accident detection and warning system based on intelligent induction signs is provided for the embodiments of the present application, as shown in Figure 2 The system comprises:

[0238] A data acquisition module 21 acquires trajectory spatiotemporal feature data of multiple target vehicles, wherein the trajectory spatiotemporal feature data includes position, speed, and acceleration information;

[0239] A time sequence synchronization module 22 performs time sequence alignment processing on the trajectory spatiotemporal feature data using the dynamic time warping principle 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 by the intelligent induction sign, extracts the geometric curvature of the tunnel curve, and performs spatial coordinate transformation on the synchronized trajectory sequence by the geometric curvature of the tunnel curve to generate curve adaptation trajectory data.

[0241] The decision generation module 24 fuses the curve adaptation trajectory data and the spatial position distribution characteristics of the target vehicle obtained by the accident monitoring scanning device in real time, generates an accident upstream blind area coverage instruction and a linkage light early warning parameter when the abnormal position aggregation degree of the target vehicle exceeds a preset dynamic threshold.

[0242] The posture control module 25 responds to the accident upstream blind area coverage instruction and dynamically adjusts the physical projection posture of the intelligent induction sign by the magneto-rheological damper built in the intelligent induction sign to generate real-time projection posture parameters.

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

[0244] Figure 2 The accident detection and early warning system based on the intelligent induction sign can perform Figure 1 The implementation principle and technical effects of the accident detection and early warning method based on the intelligent induction sign of the embodiment are not described again. The specific operation of each module and unit of the accident detection and early warning system based on the intelligent induction sign in the above embodiment has been described in detail in the embodiment related to the method, and will not be described in detail here.

[0245] In one possible design, Figure 2 The accident detection and early warning system based on the intelligent induction sign of the embodiment can be implemented as a computing device, such as Figure 3 As shown, the computing device can 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 configured to perform the above Figure 1 The accident detection and early warning method based on the intelligent induction sign of the embodiment.

[0248] The processing component 32 can include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component can also be 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, micro-controllers, microprocessors or other electronic components, configured to perform the methods described above.

[0249] The storage component 31 is configured to store various types of data to support the operation of the terminal. The storage component can be implemented by any type of volatile or non-volatile storage devices or their combinations, 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 storage, flash memory, magnetic disk or optical disk.

[0250] Of course, the computing device can also include other components, such as an input / output interface, a display component, a communication component, 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 wired or wireless communication between the computing device and other devices, etc.

[0253] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform, and the computing device can be a cloud server, and the processing component, the storage component, etc. can be a basic server resource rented or purchased from the cloud computing platform.

[0254] The embodiment of the present application also provides a computer storage medium, which stores a computer program, and the computer program can implement the above-mentioned Figure 1 An accident detection and early warning method based on intelligent induction signs is provided.

[0255] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, device and unit can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0256] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purposes of the embodiments according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0257] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and necessary general hardware platforms, and of course can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of software products, and the computer software products can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and include a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some 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, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate 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 induced markers, characterized in that, The method comprises the following steps: acquiring trajectory spatiotemporal feature data of a multi-target vehicle, the trajectory spatiotemporal feature data containing position, speed and acceleration information; performing time series alignment processing on the trajectory spatiotemporal feature data by using the dynamic time warping principle to generate a synchronized trajectory sequence; adjusting the light projection direction of the target vehicle at a curve by using an intelligent induction marker, extracting the geometric curvature of a tunnel curve, and performing spatial coordinate transformation on the synchronized trajectory sequence by using the geometric curvature of the tunnel curve to generate curve-adapted trajectory data; fusing the curve-adapted trajectory data and the spatial position distribution features of the target vehicle acquired by an accident monitoring and scanning device in real time, and generating an accident upstream blind area coverage instruction and a linked light warning parameter when the abnormal position aggregation degree of the target vehicle exceeds a preset dynamic threshold; responding to the accident upstream blind area coverage instruction and dynamically adjusting the physical projection posture of the intelligent induction marker by using a magneto-rheological damper built in the intelligent induction marker to generate real-time projection posture parameters; generating real-time matching light control signals based on the linked light warning parameter and the real-time projection posture parameter.

2. The method of claim 1, wherein, The method for dynamically adjusting the physical projection posture of the intelligent induction marker by using the magneto-rheological damper built in the intelligent induction marker to generate real-time projection posture parameters in response to the accident upstream blind area coverage instruction comprises the following steps: analyzing the spatial orientation information contained in the accident upstream blind area coverage instruction to convert it into posture adjustment requirement parameters, wherein the posture adjustment requirement parameters include a target pitch angle value and a target yaw angle value; determining the input current intensity of the magneto-rheological damper according to the difference between the target pitch angle value and the current pitch angle value of the posture adjustment requirement parameter and the difference between the target yaw angle value and the current yaw angle value; applying the input current intensity to the magneto-rheological damper to make the magnetic particles inside the magneto-rheological 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; driving a support linkage mechanism connected to the base of the intelligent induction marker by using the controllable damping force to make the support linkage mechanism produce angular displacement changes around the rotation axis, thereby obtaining an angular displacement change amount; converting the angular displacement change amount into real-time projection posture parameters in digital form, wherein the real-time projection posture parameters contain an actual pitch angle value and an actual yaw angle value.

3. The method of claim 2, wherein, The method for applying the input current intensity to the magneto-rheological damper to make the magnetic particles inside the magneto-rheological 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, comprises the following steps: transmitting the input current intensity to a plurality of layers of electromagnetic coils tightly wound around the surface of the annular inner cavity of the magneto-rheological damper through wires; converting electrical energy into magnetic energy by using the electromagnetic coils to generate a spatially uniformly distributed axial magnetic field in the annular inner cavity of the magneto-rheological damper; Under the action of the axial magnetic field, the magnetic particles suspended in the magnetorheological fluid are driven to migrate along the magnetic force line direction to overcome the Brownian motion force, and when the magnetic particles migrate to the region with the maximum magnetic flux density, the magnetic dipole moment attraction force between adjacent magnetic particles is generated to connect multiple particles head to tail to form a columnar chain structure penetrating the fluid gap; The opposite sides of the inner cavity of the magnetorheological damper are bridged by the columnar chain structure to generate a mechanical hindering effect on the fluid flow perpendicular to the chain structure direction; Based on the viscous resistance principle of fluid mechanics, the mechanical hindering effect is converted into a controllable damping force linearly corresponding to the current intensity.

4. The method of claim 1, wherein, The dynamic adjustment of the light projection direction of the target vehicle at the curve is realized by using the intelligent induction marker, the tunnel curve geometric curvature is extracted, and the spatial coordinate transformation of the synchronized trajectory sequence is realized by using the tunnel curve geometric curvature to generate the curve adaptive trajectory data, including: According to the tunnel design drawing, the center line curvature radius and the pitch slope parameters are obtained, and the geometric curvature characteristics of the tunnel curve are combined to generate; Based on the geometric curvature characteristics, a curve space coordinate system is established, wherein the origin of the curve space coordinate system is located at the entry point of the curve, the horizontal axis is perpendicular to the road surface, and the vertical axis is along the tangent direction of the center line; The vehicle position data in the synchronized trajectory sequence is mapped from the preset geodetic coordinate system to the curve space coordinate system to generate intermediate conversion positions; According to the lateral offset law of the actual driving trajectory of the tunnel curve and the theoretical center line of the tunnel curve, the intermediate conversion positions are curvature compensated and corrected to generate the curve adaptive trajectory data.

5. The method of claim 4, wherein, According to the lateral offset law of the actual driving trajectory of the tunnel curve and the theoretical center line of the tunnel curve, the intermediate conversion positions are curvature compensated and corrected to generate the curve adaptive trajectory data, including: The vehicle trajectory point set passing through the target curve in a continuous preset period is retrieved from the tunnel curve monitoring system as the position data of the historical natural driving path; The distance value perpendicular to the theoretical center line is calculated for each trajectory point in the vehicle trajectory point set to separate the pure lateral position offset component; According to the arc length of the tunnel curve, the statistical sections are equally spaced, and the pure lateral position offset components are aggregated in each section to extract the lateral offset distribution statistical characteristics under different curvature radii; Based on the lateral offset distribution statistical characteristics, a curvature radius and compensation amount mapping table is constructed and stored in the local storage of the intelligent induction marker; According to the curve arc length coordinate of the intermediate conversion position, the curvature radius index value of the corresponding statistical section is matched; The target lateral compensation amount is obtained by querying the curvature radius and compensation amount mapping table through the curvature radius index value; The target lateral compensation amount is added to the lateral coordinate value of the intermediate conversion position, and the longitudinal and elevation coordinates remain unchanged to generate the curvature compensated and corrected curve adaptive trajectory data.

6. The method as claimed in claim 2, wherein, The support link mechanism connected to the base of the intelligent induction marker is driven by the controllable damping force to produce angular displacement changes around the rotation axis to obtain the angular displacement change amount, including: The controllable damping force is transmitted to the front end hinge joint of the active arm of the support 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-shaped guide rail fixed to the induction mark base; The center of the arc-shaped guide rail is constrained to coincide with the rotation axis, ensuring that the movement trajectory of the end of the driven arm is a precise circular arc path with constant radius, and a displacement marker point is arranged at the end of the driven arm, and the arc length change data of the displacement marker point relative to the initial position is captured in real time by a laser ranging sensor; According to the geometric proportional relationship between the arc length change data and the radius of the arc-shaped guide rail, the angular displacement change of the support linkage mechanism is calculated.

7. The method of claim 1, wherein, Fusion of the curved track adaptive trajectory data and the spatial position distribution characteristics of the target vehicle obtained by the accident monitoring scanning device in real time, when the abnormal position aggregation degree of the target vehicle exceeds the preset dynamic threshold, the accident upstream blind area covering instruction and the linkage light warning parameter are generated, including: Divide the spatial partition unit of the tunnel curve, wherein the inner side area of the tunnel curve adopts a high-density unit division mode, and the outer side area of the tunnel curve adopts a low-density unit division mode; Real-time scanning of the number of vehicles existing in the spatial partition unit, generating a position aggregation degree basic quantity bound to the position of the spatial partition unit; Map the speed mutation abnormal point position identified in the curved track adaptive trajectory data to the corresponding spatial partition unit, and mark it as an accident risk core unit; Taking the accident risk core unit as the geometric center, the surrounding adjacent units are radiated and expanded in multiple levels to form a gradient detection area, and the position aggregation degree basic quantity in the gradient detection area is executed distance attenuation type weighted aggregation operation to generate a comprehensive position aggregation degree index; According to the traffic flow level fed back by the preset tunnel entrance vehicle flow detection device in real time, the judgment reference value of the comprehensive position aggregation degree index is self-adjusted, and when the comprehensive position aggregation degree index continuously exceeds the dynamically adjusted judgment reference value for a preset time window, the accident upstream blind area covering instruction and the linkage light warning parameter are generated synchronously.

8. An accident detection and early warning system based on intelligent induced markers, characterized in that, Including: A data acquisition module acquires trajectory spatiotemporal feature data of multiple target vehicles, and the trajectory spatiotemporal feature data contains position, speed and acceleration information; A time sequence synchronization module performs time sequence alignment processing on the trajectory spatiotemporal feature data using the dynamic time warping principle to generate a synchronized trajectory sequence; A curvature adaptive module uses an intelligent induction mark to dynamically adjust the light projection direction of the target vehicle at a curve, extracts the geometric curvature of a tunnel curve, and performs spatial coordinate transformation on the synchronized trajectory sequence using the geometric curvature of the tunnel curve to generate curved track adaptive trajectory data; A decision generation module fuses the curved track adaptive trajectory data and the spatial position distribution characteristics of the target vehicle obtained by the accident monitoring scanning device in real time, and generates an accident upstream blind area covering instruction and a linkage light warning parameter when the abnormal position aggregation degree of the target vehicle exceeds a preset dynamic threshold. The posture control module responds to the accident upstream blind area coverage instruction and dynamically adjusts a physical projection posture of the intelligent induction marker through a magneto-rheological damper built in the intelligent induction marker to generate real-time projection posture parameters; The light signal joint control module generates real-time matching light control signals based on the joint light early warning parameters and the real-time projection posture parameters.

9. A computing device, comprising: The method comprises 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 realize the method for accident detection and early warning based on the intelligent induction marker according to any one of claims 1-7.

10. A computer storage medium, characterized in that, The computer program is stored in the computer and is executed by the computer to realize the method for accident detection and early warning based on the intelligent induction marker according to any one of claims 1-7.

Images