Road intelligent traffic test platform and test method based on human-vehicle-road-cloud cooperation

Abstract

The present application relates to a kind of road intelligent transportation test platform and test method, in particular to a kind of road intelligent transportation test platform and test method based on man-car road cloud cooperation.The test platform includes environmental situation mirror image generation system, trajectory decision cooperation system and elastic road right redistribution system.The method includes based on the transmission path and influence range of low adhesion area in the process of vehicle flow movement based on road adhesion coefficient gradient graph, output dynamic risk heat map;According to the conflict resolution result of trajectory conflict point, dynamically generate vehicle group traffic sequence, verify global feasibility through cloud platform, and issue as executable traffic instruction;Traffic demand prediction calculated by cloud is superimposed with dynamic risk heat map, and the time-space elastic capacity of key bottleneck area in road network pressure situation map is identified, and the lane function boundary is redivided.The present application provides a full-dimensional verification environment covering individual vehicle control, group cooperative decision-making and road network resource allocation for intelligent networked vehicles.

Description

Technical Field

[0001] This invention relates to a road intelligent transportation testing platform and testing method, and particularly to a road intelligent transportation testing platform and testing method based on human-vehicle-road-cloud collaboration. Background Technology

[0002] With the deep integration of intelligent connected vehicles and intelligent transportation systems, traditional single-dimensional testing methods can no longer meet the verification needs of complex traffic scenarios. The industry currently faces three core challenges: real-world road testing is high-risk and high-cost, with an extreme scenario reproducibility rate of less than 0.3%; laboratory simulations lack multi-element dynamic coupling, making it difficult to simulate the emergence of swarm intelligence in vehicle wireless communication technology environments; and existing platforms generally suffer from data silos, with data latency between roadside equipment, vehicle terminals, and cloud systems often exceeding 200ms. This severely restricts the reliability verification of Level 4 and above autonomous driving systems.

[0003] In the prior art, for example, Chinese patent application (application number 201910100449.8) discloses an emergency method and device for autonomous vehicles to ensure the safety of the vehicle and driver during autonomous vehicle testing. This method involves real-time detection of the output data of each pre-defined functional module in the autonomous vehicle; if the output data of at least one functional module is abnormal, a preset emergency strategy is executed. While executing the preset emergency strategy ensures the safety of the vehicle and driver, it focuses on single-vehicle fault detection and is a post-event remedial mechanism; the emergency strategy only involves single-vehicle control, such as emergency braking or pulling over, and does not involve dynamic allocation of road resources; in complex traffic environments, it cannot achieve closed-loop decision-making from micro-level vehicle control to macro-level road network optimization through the collaboration of all elements of people, vehicles, roads, and the cloud. Therefore, this invention aims to provide an intelligent transportation testing platform and method based on human-vehicle-road-cloud collaboration. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a road intelligent transportation test platform based on human-vehicle-road-cloud collaboration, including an environmental situation mirror generation system, a trajectory decision collaboration system, and a flexible right-of-way redistribution system.

[0005] The environmental situation mirror generation system is configured to: extract unstructured features from dispersed vehicle movement trajectories, microscopic road surface conditions monitored by roadside facilities, and macroscopic meteorological data from the cloud; map the extracted unstructured features to a unified spatiotemporal coordinate system, and generate a road surface adhesion coefficient gradient map through a cross-entity association algorithm; based on the road surface adhesion coefficient gradient map, deduce the transmission path and impact range of low adhesion areas during the movement of traffic flow, and output a dynamic risk heat map.

[0006] The trajectory decision-making coordination system is configured to: receive real-time control status and passenger travel preferences of vehicle steering angle change rate and drive torque gradient, and parse them into quantifiable dynamic behavior vectors; fuse multi-vehicle dynamic behavior vectors and dynamic risk heat map output by the environmental situation mirror generation system, construct a virtual decision sandbox at roadside edge nodes, and simulate trajectory conflict points under different coordination strategies; dynamically generate vehicle group passage sequences based on the conflict resolution results of trajectory conflict points, and issue executable passage instructions after verifying global feasibility through a cloud platform;

[0007] The elastic right-of-way reallocation system is configured to: quantify the throughput efficiency and latency threshold of each path node based on the executable passage instructions generated by the trajectory decision-making collaboration system, and construct a road network pressure situation field; overlay the traffic demand forecast calculated in the cloud with the dynamic risk heat map output by the environmental situation mirror generation system to identify the spatiotemporal elastic capacity of key bottleneck areas in the road network pressure situation field; and redefine the lane functional boundaries based on the elastic capacity.

[0008] Furthermore, the environmental situation mirror generation system includes: a multi-source spatiotemporal fragment feature anchoring subsystem, a heterogeneous field strength coupling mapping subsystem, and a risk manifold topology inference subsystem;

[0009] The multi-source spatiotemporal fragment feature anchoring subsystem is configured to extract the speed change points in the vehicle trajectory, the abnormal tire noise spectrum sections in the micro road surface conditions collected from the roadside, and the time-varying curve of precipitation intensity in the cloud-based macro meteorological data as spatiotemporally discrete unstructured features.

[0010] The heterogeneous field strong coupling mapping subsystem is configured to: align the unstructured features output by the multi-source spatiotemporal fragment feature anchoring subsystem to the road network digital base through spatiotemporal projection, so that the speed change point coincides with the abnormal section of the tire noise spectrum in space, automatically trigger the cloud macroscopic erosion factor weighting, and generate a dynamic adhesion attenuation field with road grid as the unit, i.e., the road surface adhesion coefficient gradient map.

[0011] The risk manifold topology deduction subsystem is configured as follows: based on the spatial distribution of low-intensity regions in the road surface adhesion coefficient gradient map output by the heterogeneous field strong coupling mapping subsystem, a real-time traffic flow density vector is injected, a risk transmission channel is established with the traffic flow direction as the axis, the inertial disturbance factor is calculated based on the vehicle mass distribution in the channel, and the adhesion attenuation gradient is superimposed to generate a risk propagation vector field; finally, a dynamic risk heat map with spatiotemporal evolution attributes is output.

[0012] Furthermore, the trajectory decision-making collaborative system includes: a manipulation-preference coupling vectorization subsystem, a multi-body risk interference field construction subsystem, and a flexible passage contract generation subsystem;

[0013] The control-preference coupled vectorization subsystem is configured to: generate trajectory curvature potential energy by time integral of the vehicle steering angle change rate; transform the driving torque gradient into a dynamic response spectrum through frequency domain transformation; compress passenger travel preferences into a spatiotemporal constraint domain; and output the dynamic behavior vector of each vehicle through tensor synthesis.

[0014] The multi-body risk interference field construction subsystem is configured as follows: at the roadside edge node, the dynamic behavior vectors of multiple vehicles output by the control-preference coupling vectorization subsystem are projected into motion state and superimposed with the hazard contour lines of the dynamic risk heat map output by the environmental state mirror generation system. The trajectory conflict focus cloud map is generated through the field strength interference principle to mark the potential collision domains under different cooperative strategies.

[0015] The flexible passage contract generation subsystem is configured to: perform energy level attenuation operation on the trajectory conflict focus cloud map generated by the multi-body risk interference field construction subsystem, apply passage phase delay to high conflict areas, and inject priority passage factors into low-risk channels; generate vehicle group passage wave sequence with spatiotemporal offset, and after the cloud platform verifies the global manifold continuity of the wave sequence, issue it as an executable passage command bundle.

[0016] Furthermore, the elastic right-of-way reallocation system includes: a node performance matrix construction subsystem, a pressure field generation subsystem, a capacity elasticity assessment subsystem, and a boundary reconstruction subsystem;

[0017] The node performance matrix construction subsystem is configured to: discretize the vehicle group passage sequence data in the executable passage instructions output by the trajectory decision coordination system in the time dimension, extract the maximum passage volume per unit time of each path node as the throughput efficiency base, and synchronously record the instruction transmission delay between adjacent nodes as the time threshold parameter.

[0018] The pressure field generation subsystem is configured to: perform tensor product operation on the throughput efficiency base of the node performance matrix construction subsystem and the time threshold parameter to generate a composite index matrix containing spatial load intensity and temporal congestion degree, and form a continuous road network pressure situation field through three-dimensional interpolation.

[0019] The capacity elasticity assessment subsystem is configured as follows: after feature alignment between cloud-based traffic demand forecast data and dynamic risk heat map output by environmental situation mirror generation system, the region exceeding the critical value is located in the road network pressure situation field generated by pressure field generation subsystem; the compressible time margin of each bottleneck area is calculated based on the gradient change rate of dynamic risk heat map, and the spatial reorganization potential is derived by combining the traffic fluctuation characteristics of demand forecast to obtain spatiotemporal elastic capacity.

[0020] The boundary reconstruction subsystem is configured to convert the spatiotemporal elastic capacity assessment results into lane function adjustment parameters: the time margin corresponds to the adjustable range of the signal phase, the spatial potential determines the variable range of the lane markings, and finally outputs a set of lane function boundaries with dynamic constraints.

[0021] Preferably, the manipulation-preference coupled vectorization subsystem includes:

[0022] The motion eigenvalue reconstruction module is configured as follows: the rate of change of the vehicle steering angle is integrated over time to generate a continuous curvature manifold; the continuous curvature manifold is used to extract the principal curvature ridges through differential topological transformation; the ridge length and the curvature integral value together constitute a scalar set of trajectory curvature potential energy.

[0023] The dynamic-constraint tensor field generation module is configured to: convert the driving torque gradient into a frequency domain band distribution via fast spectral remapping; compress passenger travel preferences into a spatiotemporal constraint envelope with boundary rigidity; and project the frequency domain band distribution into the spatiotemporal constraint envelope to form a dynamic response spectrum tensor field.

[0024] The behavior vector condensation output module is configured to: inject the trajectory curvature potential energy scalar set generated by the motion eigenbase reconstruction module as a basis vector into the dynamic response spectrum tensor field generated by the dynamic-constraint tensor field generation module; stretch the tensor field along the principal curvature ridge direction through curvature-driven deformation; extract the maximum band density region at the constraint envelope boundary; and output a dynamic behavior vector containing spatial trend and intensity attributes.

[0025] Preferably, the multibody risk interference field construction subsystem includes:

[0026] The dynamic behavior vector projection mapping module is configured to project the multi-vehicle dynamic behavior vectors output by the control-preference coupling vectorization subsystem onto the road network digital base. The spatial trend component of each vector determines the main motion axis, and the intensity component is converted into the field intensity distribution along the axis to form the individual vehicle motion field.

[0027] The risk contour field intensity control module is configured to: deconstruct the hazard contour lines of the dynamic risk heat map output by the environmental situation mirror generation system into a gradient sensitive layer, and perform spatial convolution with the motion field: when the main axis direction of the motion field is orthogonal to the risk gradient direction, interference attenuation is generated; when they are in the same direction, field strength superposition is formed to generate a composite field.

[0028] The interference fringe feature extraction module is configured as follows: In the composite field modulated by the risk contour field intensity modulation module, when the phase difference between the motion field intensity and the risk gradient exceeds the critical threshold, interference fringe bands are generated, and the intersection area of ​​the fringe bands is marked as the first-level conflict focus; the field intensity superposition saturation area is marked as the second-level conflict focus.

[0029] The cooperative strategy imprinting injection module is configured such that the travel phase delay coefficients corresponding to different cooperative strategies are applied to the interference fringe bands generated by the interference fringe feature extraction module, causing the first-level conflict focus to shift spatially and the second-level conflict focus to split into energy levels, ultimately outputting a trajectory conflict focus cloud map carrying the strategy imprint.

[0030] Preferably, the flexible access contract generation subsystem includes:

[0031] The conflict energy level spectrum deconstruction module is configured to decompose the trajectory conflict focus cloud map generated by the multibody risk interferometric field construction subsystem into discrete energy level units. High conflict regions are represented as focus clusters with energy level density exceeding the threshold, while low risk channels are represented as interference fringe bands with uniform energy level spacing.

[0032] The dynamic attenuation operator injection module is configured to: apply anisotropic attenuation to the focal clusters generated by the conflict energy level spectrum deconstruction module, apply an exponential phase delay along the motion axis, and use step-type energy level suppression in the vertical direction; and implant a pulse-type passage factor into the interference fringe region, the amplitude of which is inversely proportional to the energy level interval.

[0033] The wave sequence primitive generation module is configured as follows: the dynamic attenuation operator injection module generates a deformation field from the trajectory conflict focus cloud map after attenuation, which drives the vehicle motion parameters to redistribute along the attenuation gradient; the vehicle motion parameters in the high conflict area form a time delay differential, and the parameters in the low risk area generate a lead integral, which together constitute the wave sequence primitive with offset.

[0034] The manifold continuity verification module is configured as follows: the cloud platform maps the wave sequence primitives generated by the wave sequence primitive generation module to the road network topology space and detects the parameter transitions between adjacent wave sequence primitives; when the axial offset derivative and the radial energy level gradient satisfy the bilinear constraint, it is determined that the global manifold is continuous; the wave sequence primitives that pass the manifold verification are compiled into a three-dimensional instruction structure: the time dimension carries the phase delay, the spatial dimension records the passage factor distribution, and the energy dimension encodes the energy level attenuation coefficient, finally forming an executable passage instruction bundle.

[0035] Preferably, the wave sequence primitive generation module includes:

[0036] The deformation field gradient extraction submodule is configured as follows: the spatial deformation field is generated by the trajectory conflict focus cloud map after the dynamic decay operator injection module decays, the phase delay gradient component is extracted along the motion axis direction, and the energy level suppression gradient component is obtained in the vertical direction.

[0037] The parameter differential field construction submodule is configured as follows: the energy level density gradient of the focal cluster in the high conflict region drives the construction of the time delay differential field, establishes the first-order derivative relationship in the time dimension along the phase delay gradient direction, and forms the second-order constraint condition in the energy level suppression gradient direction.

[0038] The parameter integration field generation submodule is configured as follows: the gradient of the energy level interval of the interference fringe in the low-risk region generates the advanced integration field, the amplitude of the pulsed passage factor is used as the integration coefficient, and the reciprocal of the energy level interval determines the integration step size.

[0039] The offset synthesis submodule is configured as follows: the differential field and the integral field are convolved with tensors in the road network topology space, the time delay differential component and the lead integral component are superimposed along the motion axis, and the energy level gradient components are kept orthogonally constrained; the synthesized offset field is regularized, the axial offset is normalized to the phase angle, and the radial energy level difference is normalized to the weight coefficient; finally, a wave sequence primitive structure with spatiotemporal-energy three-dimensional characteristics is formed.

[0040] Preferably, the capacity resilience assessment subsystem includes:

[0041] The risk gradient extraction module is configured to: differentiate the rate of change of the values ​​of each pixel in the dynamic risk heat map output by the environmental situation mirror generation system along the road axis, obtain the gradient intensity distribution in the risk propagation direction, and form a risk gradient field.

[0042] The time margin calculation module is configured to: in the critical region marked in the road network pressure situation field generated by the pressure field generation subsystem, the point corresponding to the maximum slope of the risk gradient field is determined as the compression start position, and the risk decay curve is integrated along the gradient decrease direction, and the length of the integration interval is converted into a compressible time margin.

[0043] The spatial potential mapping module is configured to: project the difference between the peak and valley values ​​of traffic fluctuations in cloud-based traffic demand forecasting into the risk gradient field to form a spatial reorganization coefficient matrix; perform Hadamard product operation between the spatial reorganization coefficient matrix and the spatial load intensity component of the road network pressure situation field to generate a spatial reorganization potential distribution map.

[0044] The elastic capacity synthesis module is configured to: expand the compressible time margin output by the time margin calculation module into a time elastic vector along the road network topology; extract the main variation direction of the spatial recombination potential distribution map generated by the spatial potential mapping module through principal component analysis to form a spatial elastic tensor; and synthesize the spatiotemporal elastic capacity by combining the time elastic vector and the spatial elastic tensor through the Kronecker product.

[0045] Preferably, the elastic capacity synthesis module includes:

[0046] The vector processing submodule is configured to: expand the compressible time margin output by the time margin calculation module along the road network topology, and the time adjustment parameters corresponding to each topology node constitute a time elastic vector; the spatial reorganization potential distribution map generated by the spatial potential mapping module is transformed by principal component transformation, and the extracted variation direction is orthogonally constrained to the propagation direction of the original risk gradient field, and the resulting spatial elastic tensor implies the nonlinear coupling relationship between the peak difference of flow fluctuation and the road network pressure situation field.

[0047] The cross-combination submodule is configured as follows: the network topology constraint carried by the time elastic vector is used as the first operand, and the mutation direction contained in the spatial elastic tensor is used as the second operand. The dimensional expansion of the first operand and the second operand follows the angle relationship between the gradient field propagation direction and the main mutation direction. The spatiotemporal elastic capacity is generated by cross-combining the expanded dimensions, wherein each capacity unit simultaneously encodes the decay characteristics of the time integral interval and the load characteristics of the spatial recombination coefficient.

[0048] The product operation submodule is configured as follows: in the final synthesized spatiotemporal elastic capacity matrix, the row vectors inherit the road node sequence of the time elastic vector, the column vectors continue the main variation dimension of the spatial elastic tensor, the matrix element values ​​are determined by the product of the corresponding row and column parameters, the product operation reflects the dynamic balance relationship between the slope of the risk decay curve and the spatial recombination coefficient; the diagonal elements of the matrix retain the original gradient intensity information at the compression start position.

[0049] Preferably, the cross-combination submodule includes:

[0050] The first operand preprocessing unit is configured as follows: a time elastic vector based on the road network topology expansion, whose node sorting direction is aligned with the propagation direction of the risk gradient field; the time adjustment parameters corresponding to each node are serialized along the propagation direction to form a one-dimensional vector structure with directional constraints, carrying the gradient intensity information of the compression start position, and retaining the integral characteristics of the risk decay curve.

[0051] The second operand direction adaptation unit is configured such that: the main variation direction contained in the spatial elastic tensor is adjusted by orthogonal projection to form a fixed angle with the gradient field propagation direction; the angle is determined by the covariant relationship between the directional derivative of the original risk gradient field and the peak difference of the flow fluctuation; the adjusted tensor maintains the nonlinear coupling characteristics with the road network pressure situation field, while ensuring that its dimensional expansion reference is compatible with the direction of the time elastic vector.

[0052] The dimension expansion rule building unit is configured as follows: the expansion dimension of the time elastic vector is linearly extended along the gradient field propagation direction, and the expansion dimension of the spatial elastic tensor is nonlinearly expanded along the main variation direction; the expansion ratio is adjusted by the sinusoidal component of the included angle: when the included angle approaches orthogonality, the expansion magnitude of the spatial dimension increases; when the included angle decreases, the expansion weight of the time dimension increases; the expanded time vector is upgraded to a strip matrix, and the spatial tensor is upgraded to a block matrix.

[0053] The capacity unit is generated by cross-combination and is configured as follows: the strip matrix and block matrix after the dimension expansion rule construction unit are combined according to the following rules;

[0054] The row index of the strip matrix corresponds to the road node sequence, and the column index of the block matrix corresponds to the main mutation dimension level. The value of each capacity unit is determined by the temporal control parameters of the intersecting rows and columns, where the time component contributes to the decay rate and the spatial component contributes to the load fluctuation intensity. The generation function of the unit value is forced to satisfy: the rate of change of capacity in the direction of gradient field propagation is equal to the rate of change of recombination coefficient in the direction of main mutation, forming a dynamic equilibrium constraint.

[0055] The diagonal characteristic preservation unit is configured to: mark the matrix row and column intersection points corresponding to the compression start position during the cross-combination process of generating capacity units; and write the original gradient intensity values ​​corresponding to the matrix row and column intersection points into the matrix diagonal.

[0056] Preferably, the dimension expansion rule construction unit includes:

[0057] The directional constraint sequence expansion subunit is configured as follows: a time elastic vector carrying gradient intensity information is axially replicated and extended along the propagation direction of the risk gradient field. The original time adjustment parameter of each node is used as the core value. The parameters of adjacent nodes generate a decay margin band according to the integral characteristics of the risk decay curve. The extension width is determined by the gradient intensity decay rate.

[0058] The strip-shaped structure forming sub-unit is configured such that: the parameter sequence after axial replication and extension is fixed at the position of the core value as the main ridge of the matrix through the direction locking mechanism, and the attenuation margin band is symmetrically expanded on both sides along the propagation direction to form a strip matrix with directional bandwidth gradient characteristics.

[0059] The variant axial fission subunit is configured as follows: the spatial elastic tensor triggers dimensional fission along the main mutation direction, the original tensor dimension serves as the basic lattice, and each lattice point generates a curvature multiplication factor based on the coupling strength with the road network pressure situation field. The curvature multiplication factor drives the lattice to expand nonlinearly along the main direction.

[0060] The blocky topological reconstruction subunit is configured such that the lattice after fission is reorganized through angular constraints, the expansion amplitude is modulated by the sinusoidal component of the orthogonal projection angle, the high curvature region produces a dimensional folding effect to form a dense blocky core, the low curvature region maintains a sparse connection framework, and the whole is constructed into a blocky matrix with a heterogeneous density distribution.

[0061] This invention establishes a closed-loop testing logic encompassing environmental risk perception, group decision-making simulation, and infrastructure response, providing a comprehensive verification environment for intelligent connected vehicles that covers individual vehicle control, group collaborative decision-making, and road network resource allocation.

[0062] On the other hand, the present invention also provides an intelligent transportation testing method based on human-vehicle-road-cloud collaboration, and an intelligent transportation testing platform based on human-vehicle-road-cloud collaboration provided by the present invention, comprising the following steps:

[0063] Step S1: Extract unstructured features from dispersed vehicle movement trajectories, microscopic road surface conditions monitored by roadside facilities, and macroscopic meteorological data from the cloud; map the extracted unstructured features to a unified spatiotemporal coordinate system, and generate a road surface adhesion coefficient gradient map through a cross-entity association algorithm; based on the road surface adhesion coefficient gradient map, deduce the transmission path and impact range of low adhesion areas during the movement of traffic flow, and output a dynamic risk heat map.

[0064] Further, in step S101, the speed change points in the vehicle trajectory, the abnormal tire noise spectrum sections in the micro road surface conditions collected from the roadside, and the time-varying curve of precipitation intensity in the cloud-based macro meteorological data are extracted as spatiotemporally discrete unstructured features.

[0065] Step S102: Align unstructured features to the road network digital base through spatiotemporal projection, so that speed change points and abnormal tire noise spectrum sections coincide in space, automatically trigger cloud-based macroscopic erosion factor weighting, and generate a dynamic adhesion attenuation field with road grid as the unit, i.e., road surface adhesion coefficient gradient map.

[0066] Step S103: Based on the spatial distribution of low-intensity areas in the road surface adhesion coefficient gradient map, inject the real-time traffic flow density vector, establish a risk transmission channel with the traffic flow direction as the axis, calculate the inertial disturbance factor based on the vehicle mass distribution in the channel, and superimpose the adhesion attenuation gradient to generate a risk propagation vector field; finally, output a dynamic risk heat map with spatiotemporal evolution attributes.

[0067] Step S2: Receive the real-time control status of vehicle steering angle change rate and drive torque gradient, along with passenger travel preferences, and parse them into quantifiable dynamic behavior vectors; fuse multi-vehicle dynamic behavior vectors and dynamic risk heatmaps, construct a virtual decision sandbox at roadside edge nodes, and simulate trajectory conflict points under different collaborative strategies; dynamically generate vehicle group passage sequences based on the conflict resolution results of trajectory conflict points, and issue executable passage instructions after verifying global feasibility through the cloud platform;

[0068] Further, in step S201, the trajectory curvature potential energy is generated by the time integral of the vehicle steering angle change rate, the driving torque gradient is transformed in the frequency domain to form the dynamic response spectrum, and the passenger travel preference is compressed into the spatiotemporal constraint domain; the dynamic behavior vector of each vehicle is output through tensor synthesis.

[0069] Step S202: At the roadside edge node, project the motion state of the multi-vehicle dynamic behavior vectors and overlay them with the hazard contour lines of the dynamic risk heat map. Generate a trajectory conflict focus cloud map through the field strength interference principle and mark the potential collision domains under different cooperative strategies.

[0070] Step S203: Perform energy level attenuation operation on the trajectory conflict focus cloud map, apply passage phase delay to high conflict areas, and inject priority passage factor into low-risk channels; generate vehicle group passage wave sequence with spatiotemporal offset, and after the cloud platform verifies the global manifold continuity of the wave sequence, issue it as an executable passage command bundle.

[0071] Step S3: Based on the generated executable passage instructions, quantify the throughput efficiency and latency threshold of each path node to construct a road network pressure situation field; overlay the traffic demand forecast from the cloud with the dynamic risk heat map to identify the spatiotemporal elastic capacity of key bottleneck areas in the road network pressure situation field; and redefine lane functional boundaries based on elastic capacity.

[0072] Further, in step S301, the vehicle group passage sequence data in the executable passage instructions is discretized in the time dimension, the maximum passage volume per unit time of each path node is extracted as the throughput efficiency base, and the instruction transmission delay between adjacent nodes is recorded synchronously as the time threshold parameter.

[0073] Step S302: Perform tensor product operation on the throughput efficiency base and the time threshold parameter to generate a composite index matrix containing spatial load intensity and temporal congestion degree, and form a continuous road network pressure situation field through three-dimensional interpolation.

[0074] Step S303: After feature alignment between cloud-based traffic demand forecast data and dynamic risk heat map, locate areas exceeding critical values ​​in the road network pressure situation field; calculate the compressible time margin of each bottleneck area based on the gradient change rate of the dynamic risk heat map, and deduce the spatial reorganization potential by combining the traffic fluctuation characteristics of demand forecast to obtain the spatiotemporal elastic capacity.

[0075] Step S304: Convert the spatiotemporal elastic capacity assessment results into lane function adjustment parameters: the time margin corresponds to the adjustable range of the signal phase, the spatial potential determines the variable range of lane markings, and finally outputs a set of lane function boundaries with dynamic constraints.

[0076] The intelligent transportation testing method based on human-vehicle-road-cloud collaboration described in this invention achieves integrated verification of multi-source heterogeneous data fusion, group collaborative decision optimization, and dynamic right-of-way allocation, realizing the following technical features: Full-element environmental situation modeling and risk propagation simulation: Through unstructured feature extraction and unified spatiotemporal mapping in step S1, vehicle trajectories, micro-road conditions, and macro-meteorological data are transformed into road surface adhesion coefficient gradient maps, thereby generating dynamic risk heat maps. This solves the problem of missing risk transmission path modeling in low-adhesion areas caused by fragmented environmental data in traditional testing methods. Dynamic conflict prediction and resolution in multi-vehicle collaborative decision-making: Step S2 generates dynamic behavior vectors by analyzing individual vehicle control states and passenger preferences. Combined with the risk heat map, a virtual decision sandbox is constructed at edge nodes to pre-simulate multi-vehicle trajectory conflict points and generate traffic sequences. This achieves vertical decision verification from individual vehicle control to group collaboration, avoiding chain congestion caused by delayed conflict response in actual road tests. The road network resource elastic adaptation and global efficiency optimization, step S3 quantifies the node throughput efficiency and latency threshold based on the passage instructions, superimposes traffic demand forecasting and risk heat map to identify the elastic capacity of bottleneck areas, dynamically reconstructs the lane functional boundaries, and forms a closed-loop verification chain of risk perception-decision generation-resource allocation, which improves the accuracy of the test platform's assessment of the road network's adaptive capability under extreme conditions.

[0077] In summary, this invention constructs a complete testing framework covering the environmental perception layer, collaborative decision-making layer, and resource scheduling layer, providing a full-process verification method for intelligent connected transportation systems that supports dynamic risk simulation, multi-agent game simulation, and infrastructure resilient response.

[0078] The present invention also provides an electronic device, comprising: a processor and a memory; the memory is used to store at least one executable instruction, the executable instruction causing the processor to perform the functions of various structures in the test platform of the present invention, or causing the processor to perform the test method of the present invention.

[0079] The present invention also provides a computer storage medium storing a computer program thereon, which, when executed by a processor, implements the functions of the various structures in the test platform described in the present invention, or implements the test method described in the present invention.

[0080] The beneficial effects of this invention are:

[0081] This invention integrates discrete vehicle motion data, microscopic road surface conditions, and macroscopic meteorological information through an environmental situation mirror generation system to establish a road surface adhesion coefficient gradient map and a dynamic risk heat map, achieving a quantitative representation from static environmental parameters to dynamic risk propagation paths; it solves the problem of insufficient modeling of chain effects in low-adhesion areas in traditional testing platforms. The trajectory decision-making collaboration system transforms single-vehicle control states and passenger preferences into dynamic behavior vectors, and combines the risk heat map to pre-simulate multi-vehicle collaborative strategies at roadside edge nodes; it identifies trajectory conflict points in advance and generates passage sequences through a virtual decision sandbox, reducing the frequency of emergency avoidance in real-vehicle testing while ensuring global feasibility in the cloud. The elastic right-of-way redistribution system dynamically adjusts lane function boundaries by quantifying node throughput efficiency and latency thresholds, overlaying traffic demand forecasts and dynamic risk heat maps. This mechanism allows the testing platform to verify dynamic lane function reconstruction strategies under special conditions (such as icy and snowy weather + peak traffic), improving bottleneck area traffic efficiency compared to fixed right-of-way allocation schemes. Attached Figure Description

[0082] Figure 1 This is a logic block diagram of the test platform provided in Embodiment 1 of the present invention;

[0083] Figure 2 This is a logic block diagram of the environmental situation mirror generation system provided in Embodiment 1 of the present invention;

[0084] Figure 3 This is a logical block diagram of the trajectory decision-making collaborative system provided in Embodiment 1 of the present invention;

[0085] Figure 4 This is a logical block diagram of the flexible right-of-way redistribution system provided in Embodiment 1 of the present invention;

[0086] Figure 5 This is a schematic diagram of the test method provided in Embodiment 7 of the present invention. Detailed Implementation

[0087] Example 1

[0088] like Figure 1-4 As shown, this embodiment provides an intelligent transportation test platform based on human-vehicle-road-cloud collaboration, including an environmental situation mirror generation system, a trajectory decision collaboration system, and a flexible right-of-way redistribution system;

[0089] The environmental situation mirror generation system is configured to: extract unstructured features from dispersed vehicle movement trajectories, microscopic road surface conditions monitored by roadside facilities, and macroscopic meteorological data from the cloud; map the extracted unstructured features to a unified spatiotemporal coordinate system, and generate a road surface adhesion coefficient gradient map through a cross-entity association algorithm; based on the road surface adhesion coefficient gradient map, deduce the transmission path and impact range of low adhesion areas during the movement of traffic flow, and output a dynamic risk heat map.

[0090] The trajectory decision-making coordination system is configured to: receive real-time control status and passenger travel preferences of vehicle steering angle change rate and drive torque gradient, and parse them into quantifiable dynamic behavior vectors; fuse multi-vehicle dynamic behavior vectors and dynamic risk heat map output by the environmental situation mirror generation system, construct a virtual decision sandbox at roadside edge nodes, and simulate trajectory conflict points under different coordination strategies; dynamically generate vehicle group passage sequences based on the conflict resolution results of trajectory conflict points, and issue executable passage instructions after verifying global feasibility through a cloud platform;

[0091] The elastic right-of-way reallocation system is configured to: quantify the throughput efficiency and latency threshold of each path node based on the executable passage instructions generated by the trajectory decision-making collaboration system, and construct a road network pressure situation field; overlay the traffic demand forecast calculated in the cloud with the dynamic risk heat map output by the environmental situation mirror generation system to identify the spatiotemporal elastic capacity of key bottleneck areas in the road network pressure situation field; and redefine the lane functional boundaries based on the elastic capacity.

[0092] In this embodiment, an environmental situation mirror generation system fuses discrete vehicle motion data, microscopic road surface conditions, and macroscopic meteorological information to establish a road surface adhesion coefficient gradient map and a dynamic risk heat map, achieving a quantitative representation from static environmental parameters to dynamic risk propagation paths; this solves the problem of insufficient modeling of chain effects in low-adhesion areas in traditional testing platforms. The trajectory decision-making collaboration system transforms single-vehicle control states and passenger preferences into dynamic behavior vectors, combining the risk heat map to pre-simulate multi-vehicle collaborative strategies at roadside edge nodes; by using a virtual decision sandbox to identify trajectory conflict points in advance and generate passage sequences, it reduces the frequency of emergency avoidance in real-vehicle testing while ensuring global feasibility in the cloud. The elastic right-of-way reallocation system dynamically adjusts lane function boundaries by quantifying node throughput efficiency and latency thresholds, overlaying traffic demand forecasts and risk heat maps. This mechanism allows the testing platform to verify dynamic lane function reconstruction strategies under special conditions (such as icy and snowy weather + peak traffic), improving bottleneck area traffic efficiency compared to fixed right-of-way allocation schemes.

[0093] In summary, this embodiment forms a closed-loop testing logic for environmental risk perception, group decision-making simulation, and infrastructure response, providing a full-dimensional verification environment for intelligent connected vehicles covering individual vehicle control, group collaborative decision-making, and road network resource allocation.

[0094] Furthermore, the environmental situation mirror generation system includes: a multi-source spatiotemporal fragment feature anchoring subsystem, a heterogeneous field strength coupling mapping subsystem, and a risk manifold topology inference subsystem;

[0095] The multi-source spatiotemporal fragment feature anchoring subsystem is configured to extract the speed change points in the vehicle trajectory, the abnormal tire noise spectrum sections in the micro road surface conditions collected from the roadside, and the time-varying curve of precipitation intensity in the cloud-based macro meteorological data as spatiotemporally discrete unstructured features.

[0096] The heterogeneous field strong coupling mapping subsystem is configured to: align the unstructured features output by the multi-source spatiotemporal fragment feature anchoring subsystem to the road network digital base through spatiotemporal projection, so that the speed change point coincides with the abnormal section of the tire noise spectrum in space, automatically trigger the cloud macroscopic erosion factor weighting, and generate a dynamic adhesion attenuation field with road grid as the unit, i.e., the road surface adhesion coefficient gradient map.

[0097] The risk manifold topology deduction subsystem is configured as follows: based on the spatial distribution of low-intensity regions in the road surface adhesion coefficient gradient map output by the heterogeneous field strong coupling mapping subsystem, a real-time traffic flow density vector is injected, a risk transmission channel is established with the traffic flow direction as the axis, the inertial disturbance factor is calculated based on the vehicle mass distribution in the channel, and the adhesion attenuation gradient is superimposed to generate a risk propagation vector field; finally, a dynamic risk heat map with spatiotemporal evolution attributes is output.

[0098] The environmental situation mirror generation system, through the collaborative operation of multiple modules, constructs a dynamic traffic risk modeling system from micro to macro levels. Through a multi-source spatiotemporal fragment feature anchoring subsystem, it achieves three-dimensional data coupling of speed abrupt changes in vehicle motion characteristics, tire noise spectrum anomalies in road surface characteristics, and time-varying precipitation intensity in meteorological characteristics, forming a comprehensive feature extraction capability covering vehicles, roads, and the environment. The heterogeneous field strength coupling mapping subsystem transforms discrete features into continuous field strength distributions. Through the construction of a dynamic adhesion attenuation field, it quantifies the correlation between risk factors of different dimensions, such as speed abrupt changes and road slippage, within a unified physical model. The risk manifold topology deduction subsystem introduces traffic flow density vectors and inertial disturbance factors, establishing a risk transmission model considering mass distribution and motion direction, transforming the static adhesion coefficient gradient into a vector field with spatiotemporal propagation characteristics. The overall environmental situation mirror generation system forms a closed-loop processing chain of feature extraction, field strength mapping, and propagation deduction, providing a computable dynamic risk situation basis for subsequent decision-making systems.

[0099] Furthermore, the trajectory decision-making collaborative system includes: a manipulation-preference coupling vectorization subsystem, a multi-body risk interference field construction subsystem, and a flexible passage contract generation subsystem;

[0100] The control-preference coupled vectorization subsystem is configured to: generate trajectory curvature potential energy by time integral of the vehicle steering angle change rate; transform the driving torque gradient into a dynamic response spectrum through frequency domain transformation; compress passenger travel preferences into a spatiotemporal constraint domain; and output the dynamic behavior vector of each vehicle through tensor synthesis.

[0101] The multi-body risk interference field construction subsystem is configured as follows: at the roadside edge node, the dynamic behavior vectors of multiple vehicles output by the control-preference coupling vectorization subsystem are projected into motion state and superimposed with the hazard contour lines of the dynamic risk heat map output by the environmental state mirror generation system. The trajectory conflict focus cloud map is generated through the field strength interference principle to mark the potential collision domains under different cooperative strategies.

[0102] The flexible passage contract generation subsystem is configured to: perform energy level attenuation operation on the trajectory conflict focus cloud map generated by the multi-body risk interference field construction subsystem, apply passage phase delay to high conflict areas, and inject priority passage factors into low-risk channels; generate vehicle group passage wave sequence with spatiotemporal offset, and after the cloud platform verifies the global manifold continuity of the wave sequence, issue it as an executable passage command bundle.

[0103] The trajectory decision-making collaboration system achieves quantitative collaborative control of multi-vehicle dynamic behavior through a modular architecture. Its technical significance is specifically manifested in the following aspects:

[0104] The vehicle behavior vectorization modeling and the control-preference coupled vectorization subsystem integrate the steering angle change rate and driving torque gradient of the vehicle control characteristics with the spatiotemporal constraints of passenger preferences into a dynamic behavior vector, providing a computable vehicle motion representation for subsequent collaborative decision-making.

[0105] The multi-vehicle risk interference prediction and multi-body risk interference field construction subsystem quantifies the conflict probability between multiple vehicle trajectories by overlaying motion state projection and risk heat map, and generates a focal cloud map containing potential collision domains to achieve visualization and calibration of the risk space.

[0106] The dynamic traffic contract optimization and flexible traffic contract generation subsystem adjusts traffic priorities based on the conflict energy level distribution and constructs the traffic wave sequence of vehicle groups through spatiotemporal offsets. It achieves a balance between conflict resolution and traffic efficiency while ensuring manifold continuity, providing a collaborative control framework that takes into account both safety and efficiency for mixed traffic flows.

[0107] Furthermore, the elastic right-of-way reallocation system includes: a node performance matrix construction subsystem, a pressure field generation subsystem, a capacity elasticity assessment subsystem, and a boundary reconstruction subsystem;

[0108] The node performance matrix construction subsystem is configured to: discretize the vehicle group passage sequence data in the executable passage instructions output by the trajectory decision coordination system in the time dimension, extract the maximum passage volume per unit time of each path node as the throughput efficiency base, and synchronously record the instruction transmission delay between adjacent nodes as the time threshold parameter.

[0109] The pressure field generation subsystem is configured to: perform tensor product operation on the throughput efficiency base of the node performance matrix construction subsystem and the time threshold parameter to generate a composite index matrix containing spatial load intensity and temporal congestion degree, and form a continuous road network pressure situation field through three-dimensional interpolation.

[0110] The capacity elasticity assessment subsystem is configured as follows: after feature alignment between cloud-based traffic demand forecast data and dynamic risk heat map output by environmental situation mirror generation system, the region exceeding the critical value is located in the road network pressure situation field generated by pressure field generation subsystem; the compressible time margin of each bottleneck area is calculated based on the gradient change rate of dynamic risk heat map, and the spatial reorganization potential is derived by combining the traffic fluctuation characteristics of demand forecast to obtain spatiotemporal elastic capacity.

[0111] The boundary reconstruction subsystem is configured to convert the spatiotemporal elastic capacity assessment results into lane function adjustment parameters: the time margin corresponds to the adjustable range of the signal phase, the spatial potential determines the variable range of the lane markings, and finally outputs a set of lane function boundaries with dynamic constraints.

[0112] The aforementioned flexible right-of-way reallocation system achieves precise control of traffic resources through multi-dimensional parameter coupling and dynamic constraint transformation, providing a quantitative decision-making basis for the dynamic optimization of right-of-way allocation in complex traffic environments; its technical significance is specifically manifested in the following ways:

[0113] The node performance quantification and node performance matrix construction subsystem transforms the communication instructions into computable spatiotemporal parameter pairs, establishing a basic evaluation framework with throughput efficiency as the spatial dimension and transmission delay as the temporal dimension.

[0114] The road network status visualization and pressure field generation subsystem fuse discrete node parameters into a continuous field through tensor operations, realizing the coupled expression of spatial load and temporal congestion, and forming an analytical road network status topology.

[0115] The bottleneck resilience analysis and capacity resilience assessment subsystem locates key constraint areas through multi-source data fusion, quantifies the adjustable margin in the spatiotemporal dimensions, and provides precise adjustment boundaries for resource reorganization.

[0116] The dynamic resource allocation and boundary reconfiguration subsystem converts elastic capacity into operable traffic control parameters, establishes a mapping relationship between lane functions and spatiotemporal constraints, and forms an infrastructure adjustment scheme that meets dynamic needs.

[0117] Example 2

[0118] This embodiment provides an intelligent transportation testing platform based on human-vehicle-road-cloud collaboration. Building upon Embodiment 1, the control-preference coupled vectorization subsystem further includes:

[0119] The motion eigenvalue reconstruction module is configured as follows: the rate of change of the vehicle steering angle is integrated over time to generate a continuous curvature manifold; the continuous curvature manifold is used to extract the principal curvature ridges through differential topological transformation; the ridge length and the curvature integral value together constitute a scalar set of trajectory curvature potential energy.

[0120] The dynamic-constraint tensor field generation module is configured to: convert the driving torque gradient into a frequency domain band distribution via fast spectral remapping; compress passenger travel preferences into a spatiotemporal constraint envelope with boundary rigidity; and project the frequency domain band distribution into the spatiotemporal constraint envelope to form a dynamic response spectrum tensor field.

[0121] The behavior vector condensation output module is configured to: inject the trajectory curvature potential energy scalar set generated by the motion eigenbase reconstruction module as a basis vector into the dynamic response spectrum tensor field generated by the dynamic-constraint tensor field generation module; stretch the tensor field along the principal curvature ridge direction through curvature-driven deformation; extract the maximum band density region at the constraint envelope boundary; and output a dynamic behavior vector containing spatial trend and intensity attributes.

[0122] In this embodiment, the control-preference coupled vectorization subsystem achieves deep fusion of vehicle motion characteristics and passenger preferences through hierarchical processing, enabling the computational transformation of driving behavior from physical signals to decision-making vectors. Its technical significance is specifically manifested in the following ways:

[0123] The trajectory geometry characteristics are quantified. The motion eigenvalue reconstruction module transforms discrete turning operations into quantifiable indices of curvature potential with continuous characteristics through curvature manifold transformation and ridge extraction, thus establishing a mathematical representation of the trajectory geometry characteristics.

[0124] Dynamic-constraint co-modeling: The dynamic-constraint tensor field generation module realizes the coupled expression of driving torque gradient and passenger preference in a unified mathematical space by fusing the frequency domain energy band distribution with the projection of the spatiotemporal constraint envelope.

[0125] The behavior vector structured output module generates a composite vector that simultaneously contains the principal curvature ridge direction and energy band density by deforming the curvature-driven tensor field and extracting the band density. This provides the upper-level system with standardized behavior input that combines geometric features and dynamic constraints.

[0126] The multibody risk interference field construction subsystem further includes:

[0127] The dynamic behavior vector projection mapping module is configured to project the multi-vehicle dynamic behavior vectors output by the control-preference coupling vectorization subsystem onto the road network digital base. The spatial trend component of each vector determines the main motion axis, and the intensity component is converted into the field intensity distribution along the axis to form the individual vehicle motion field.

[0128] The risk contour field intensity control module is configured to: deconstruct the hazard contour lines of the dynamic risk heat map output by the environmental situation mirror generation system into a gradient sensitive layer, and perform spatial convolution with the motion field: when the main axis direction of the motion field is orthogonal to the risk gradient direction, interference attenuation is generated; when they are in the same direction, field strength superposition is formed to generate a composite field.

[0129] The interference fringe feature extraction module is configured as follows: In the composite field modulated by the risk contour field intensity modulation module, when the phase difference between the motion field intensity and the risk gradient exceeds the critical threshold, interference fringe bands are generated, and the intersection area of ​​the fringe bands is marked as the first-level conflict focus; the field intensity superposition saturation area is marked as the second-level conflict focus.

[0130] The cooperative strategy imprinting injection module is configured such that the travel phase delay coefficients corresponding to different cooperative strategies are applied to the interference fringe bands generated by the interference fringe feature extraction module, causing the first-level conflict focus to shift spatially and the second-level conflict focus to split into energy levels, ultimately outputting a trajectory conflict focus cloud map carrying the strategy imprint.

[0131] In this embodiment, the multi-body risk interference field construction subsystem achieves refined modeling of traffic conflicts through multi-level field transformation, providing the upper-level system with a conflict space model that combines physical accuracy and strategic extensibility. Its technical significance is specifically manifested in the following ways:

[0132] The motion-risk coupled field generation module transforms discrete vehicle behavior vectors into a continuous motion field distribution. The risk contour field emphasis control module establishes a mathematical coupling relationship between mechanical motion characteristics and environmental risk characteristics through spatial convolution of the gradient sensitive layer with the motion field.

[0133] The conflict feature quantum identification module uses the phase difference threshold to discretize the conflict, deconstructing the traditional continuous risk field into first- or second-order conflict foci with clear energy level differences, forming a computable risk quantum unit.

[0134] The strategy-responsive field reconstruction and collaborative strategy imprinting injection module adjusts the interference fringe stripes parametrically to generate strategy-related dynamic deformations at the static conflict focus, thereby achieving mathematical isomorphism between the physical conflict domain and the decision parameters.

[0135] The aforementioned flexible access contract generation subsystem further includes:

[0136] The conflict energy level spectrum deconstruction module is configured to decompose the trajectory conflict focus cloud map generated by the multibody risk interferometric field construction subsystem into discrete energy level units. High conflict regions are represented as focus clusters with energy level density exceeding the threshold, while low risk channels are represented as interference fringe bands with uniform energy level spacing.

[0137] The dynamic attenuation operator injection module is configured to: apply anisotropic attenuation to the focal clusters generated by the conflict energy level spectrum deconstruction module, apply an exponential phase delay along the motion axis, and use step-type energy level suppression in the vertical direction; and implant a pulse-type passage factor into the interference fringe region, the amplitude of which is inversely proportional to the energy level interval.

[0138] The wave sequence primitive generation module is configured as follows: the dynamic attenuation operator injection module generates a deformation field from the trajectory conflict focus cloud map after attenuation, which drives the vehicle motion parameters to redistribute along the attenuation gradient; the vehicle motion parameters in the high conflict area form a time delay differential, and the parameters in the low risk area generate a lead integral, which together constitute the wave sequence primitive with offset.

[0139] The manifold continuity verification module is configured as follows: the cloud platform maps the wave sequence primitives generated by the wave sequence primitive generation module to the road network topology space and detects the parameter transitions between adjacent wave sequence primitives; when the axial offset derivative and the radial energy level gradient satisfy the bilinear constraint, it is determined that the global manifold is continuous; the wave sequence primitives that pass the manifold verification are compiled into a three-dimensional instruction structure: the time dimension carries the phase delay, the spatial dimension records the passage factor distribution, and the energy dimension encodes the energy level attenuation coefficient, finally forming an executable passage instruction bundle.

[0140] In this embodiment, the flexible traffic contract generation subsystem achieves precise control of traffic decisions through multi-level energy field reconstruction, providing a distributed solution that satisfies energy optimization and topological constraints for complex traffic scenarios. Its technical significance is specifically manifested in the following ways:

[0141] The conflict energy dimension decoupling module discretizes the continuous conflict field into standardized energy level units, establishes a quantifiable conflict energy coordinate system, and achieves dimensional separation of risk characteristics.

[0142] The dynamic field control mechanism, through the dynamic decay operator injection module, achieves selective dissipation of conflict energy while maintaining motion continuity by employing an anisotropic operation strategy, thus forming an energy control channel with directional characteristics.

[0143] Spatiotemporal parameter coupling reconstruction: The sequence primitive generation module transforms energy field changes into differential-integral transformations of motion parameters, establishes a mathematical mapping relationship between energy gradient and spatiotemporal offset, and realizes the transformation from physical field to decision parameters.

[0144] The topological constraint verification system and the manifold continuity verification module ensure that the global motion pattern after local parameter adjustment still satisfies the topological conservation law through bilinear constraint conditions, thus preventing structural contradictions in decision parameters.

[0145] Example 3

[0146] This embodiment provides an intelligent transportation testing platform based on human-vehicle-road-cloud collaboration. Building upon Embodiment 2, the wave sequence primitive generation module further includes:

[0147] The deformation field gradient extraction submodule is configured as follows: the spatial deformation field is generated by the trajectory conflict focus cloud map after the dynamic decay operator injection module decays, the phase delay gradient component is extracted along the motion axis direction, and the energy level suppression gradient component is obtained in the vertical direction.

[0148] The parameter differential field construction submodule is configured as follows: the energy level density gradient of the focal cluster in the high conflict region drives the construction of the time delay differential field, establishes the first-order derivative relationship in the time dimension along the phase delay gradient direction, and forms the second-order constraint condition in the energy level suppression gradient direction.

[0149] The parameter integration field generation submodule is configured as follows: the gradient of the energy level interval of the interference fringe in the low-risk region generates the advanced integration field, the amplitude of the pulsed passage factor is used as the integration coefficient, and the reciprocal of the energy level interval determines the integration step size.

[0150] The offset synthesis submodule is configured as follows: the differential field and the integral field are convolved with tensors in the road network topology space, the time delay differential component and the lead integral component are superimposed along the motion axis, and the energy level gradient components are kept orthogonally constrained; the synthesized offset field is regularized, the axial offset is normalized to the phase angle, and the radial energy level difference is normalized to the weight coefficient; finally, a wave sequence primitive structure with spatiotemporal-energy three-dimensional characteristics is formed.

[0151] In this embodiment, the wave sequence primitive generation module achieves accurate reconstruction of traffic motion parameters through multi-dimensional field transformation, providing a motion parameter benchmark that meets spatiotemporal constraints and energy balance for traffic cooperative control. Its technical significance is specifically manifested in the following ways:

[0152] The spatial gradient analysis and deformation field gradient extraction submodule realizes the conversion from the conflict energy field to the gradient of the motion parameters, establishes the mapping relationship between energy distribution and spatial derivative, and provides the basic field quantity for subsequent parameter reconstruction.

[0153] The conflict parameter modeling and parameter differential field construction submodule transforms the high-energy-density region into a time-delay differential constraint. Through the first-order derivative relationship and the second-order constraint conditions, a motion delay control model for the conflict region is formed.

[0154] The passage parameter optimization submodule utilizes the characteristics of low-risk areas to construct an advanced integration mechanism, and achieves a quantitative improvement in passage efficiency by dynamically adjusting the integration coefficient and step size.

[0155] The global parameter fusion and offset synthesis submodule couple the differential field and integral field in the topological space through tensor convolution operation, maintaining orthogonal constraints while completing parameter normalization to form a unified control primitive.

[0156] Example 4

[0157] This embodiment provides an intelligent transportation testing platform based on human-vehicle-road-cloud collaboration. Building upon Embodiment 1, the capacity resilience assessment subsystem further includes:

[0158] The risk gradient extraction module is configured to: differentiate the rate of change of the values ​​of each pixel in the dynamic risk heat map output by the environmental situation mirror generation system along the road axis, obtain the gradient intensity distribution in the risk propagation direction, and form a risk gradient field.

[0159] The time margin calculation module is configured to: in the critical region marked in the road network pressure situation field generated by the pressure field generation subsystem, the point corresponding to the maximum slope of the risk gradient field is determined as the compression start position, and the risk decay curve is integrated along the gradient decrease direction, and the length of the integration interval is converted into a compressible time margin.

[0160] The spatial potential mapping module is configured to: project the difference between the peak and valley values ​​of traffic fluctuations in cloud-based traffic demand forecasting into the risk gradient field to form a spatial reorganization coefficient matrix; perform Hadamard product operation between the spatial reorganization coefficient matrix and the spatial load intensity component of the road network pressure situation field to generate a spatial reorganization potential distribution map.

[0161] The elastic capacity synthesis module is configured to: expand the compressible time margin output by the time margin calculation module into a time elastic vector along the road network topology; extract the main variation direction of the spatial recombination potential distribution map generated by the spatial potential mapping module through principal component analysis to form a spatial elastic tensor; and synthesize the spatiotemporal elastic capacity by combining the time elastic vector and the spatial elastic tensor through the Kronecker product.

[0162] In this embodiment, the overall significance of the capacity elasticity assessment subsystem lies in achieving dynamic quantitative assessment of the carrying capacity of the transportation network through multi-dimensional collaborative calculation, providing a complete decision-making basis for the elastic optimization of the transportation system. All calculation processes strictly follow the principles of vector field operation and tensor analysis, maintaining the rigor and interpretability of the mathematical model. The modules form the following technical closed loop: The risk gradient extraction module transforms the static risk heatmap into a vector field with directional attributes through spatial differential operations, providing basic field data for subsequent spatiotemporal analysis; the derivative calculation of the road axis direction ensures the consistency between the gradient direction and the traffic flow direction. The time margin calculation module establishes a mapping relationship between the pressure situation and the time dimension, transforming the spatial risk attenuation characteristics into quantifiable time buffer parameters through gradient field slope extreme point location and curve integration; this parameter reflects the system's adjustment potential in the time dimension. The spatial potential mapping module realizes the coupled calculation of demand forecasting and current load, maintaining matrix element-level operation characteristics through Hadamard product, ensuring the nonlinear superposition of traffic fluctuation characteristics and spatial load intensity, and generating physically meaningful spatial reorganization parameters. The elastic capacity synthesis module completes the tensor integration of spatiotemporal parameters: the temporal elastic vector preserves the temporal characteristics under network topology constraints, the spatial elastic tensor retains the main variation patterns through principal component analysis, and finally the dimensional expansion of spatiotemporal parameters is achieved through Kronecker product to form a complete spatiotemporal elasticity evaluation matrix.

[0163] Example 5

[0164] This embodiment provides an intelligent transportation testing platform based on human-vehicle-road-cloud collaboration. Building upon embodiment 4, the elastic capacity synthesis module further includes:

[0165] The vector processing submodule is configured to: expand the compressible time margin output by the time margin calculation module along the road network topology, and the time adjustment parameters corresponding to each topology node constitute a time elastic vector; the spatial reorganization potential distribution map generated by the spatial potential mapping module is transformed by principal component transformation, and the extracted variation direction is orthogonally constrained to the propagation direction of the original risk gradient field, and the resulting spatial elastic tensor implies the nonlinear coupling relationship between the peak difference of flow fluctuation and the road network pressure situation field.

[0166] The cross-combination submodule is configured as follows: the network topology constraint carried by the time elastic vector is used as the first operand, and the mutation direction contained in the spatial elastic tensor is used as the second operand. The dimensional expansion of the first operand and the second operand follows the angle relationship between the gradient field propagation direction and the main mutation direction. The spatiotemporal elastic capacity is generated by cross-combining the expanded dimensions, wherein each capacity unit simultaneously encodes the decay characteristics of the time integral interval and the load characteristics of the spatial recombination coefficient.

[0167] The product operation submodule is configured as follows: in the final synthesized spatiotemporal elastic capacity matrix, the row vectors inherit the road node sequence of the time elastic vector, the column vectors continue the main variation dimension of the spatial elastic tensor, the matrix element values ​​are determined by the product of the corresponding row and column parameters, the product operation reflects the dynamic balance relationship between the slope of the risk decay curve and the spatial recombination coefficient; the diagonal elements of the matrix retain the original gradient intensity information at the compression start position.

[0168] The output spatiotemporal elastic capacity matrix has the following characteristics: the time dimension reflects the adjustable time margin of the road network during the risk attenuation process; the spatial dimension characterizes the spatial reorganization potential under the coupling effect of traffic fluctuation and load intensity; the spatiotemporal coupling quantifies the comprehensive elasticity of the network under the spatiotemporal synergy through the cross combination and product operation of matrix elements.

[0169] The cross-combination submodule further includes:

[0170] The first operand preprocessing unit is configured as follows: a time elastic vector based on the road network topology expansion, whose node sorting direction is aligned with the propagation direction of the risk gradient field; the time adjustment parameters corresponding to each node are serialized along the propagation direction to form a one-dimensional vector structure with directional constraints, carrying the gradient intensity information of the compression start position, and retaining the integral characteristics of the risk decay curve.

[0171] The second operand direction adaptation unit is configured such that: the main variation direction contained in the spatial elastic tensor is adjusted by orthogonal projection to form a fixed angle with the gradient field propagation direction; the angle is determined by the covariant relationship between the directional derivative of the original risk gradient field and the peak difference of the flow fluctuation; the adjusted tensor maintains the nonlinear coupling characteristics with the road network pressure situation field, while ensuring that its dimensional expansion reference is compatible with the direction of the time elastic vector.

[0172] The dimension expansion rule building unit is configured as follows: the expansion dimension of the time elastic vector is linearly extended along the gradient field propagation direction, and the expansion dimension of the spatial elastic tensor is nonlinearly expanded along the main variation direction; the expansion ratio is adjusted by the sinusoidal component of the included angle: when the included angle approaches orthogonality, the expansion magnitude of the spatial dimension increases; when the included angle decreases, the expansion weight of the time dimension increases; the expanded time vector is upgraded to a strip matrix, and the spatial tensor is upgraded to a block matrix.

[0173] The capacity unit is generated by cross-combination and is configured as follows: the strip matrix and block matrix after the dimension expansion rule construction unit are combined according to the following rules;

[0174] The row index of the strip matrix corresponds to the road node sequence, and the column index of the block matrix corresponds to the main mutation dimension level. The value of each capacity unit is determined by the temporal control parameters of the intersecting rows and columns, where the time component contributes to the decay rate and the spatial component contributes to the load fluctuation intensity. The generation function of the unit value is forced to satisfy: the rate of change of capacity in the direction of gradient field propagation is equal to the rate of change of recombination coefficient in the direction of main mutation, forming a dynamic equilibrium constraint.

[0175] The diagonal characteristic preservation unit is configured to: mark the matrix row and column intersection points corresponding to the compression start position during the cross-combination process of generating capacity units; and write the original gradient intensity values ​​corresponding to the matrix row and column intersection points into the matrix diagonal.

[0176] In this embodiment, the overall significance of the elastic capacity synthesis module lies in constructing a quantitative model of traffic network carrying capacity with spatiotemporal coupling characteristics. Through multi-level operations, it achieves the following technical functions: The vector processing submodule transforms the compressible margin in the time dimension and the reorganization potential in the spatial dimension into structured parameters; the temporal elastic vector retains the positional characteristics of the compression starting point in the risk gradient field and expands along the road network topology, forming a distributed representation of time-adjustable capacity; the spatial elastic tensor extracts key variation directions through principal component transformation, implicitly revealing the nonlinear correlation between traffic fluctuations and load intensity, ensuring that the directionality of spatial reorganization potential is orthogonal to the gradient field propagation direction. The cross-combination submodule achieves dimensional adaptation and fusion of time and spatial parameters. The network topology constraints of the temporal elastic vector and the variation direction of the spatial elastic tensor are dimensionally expanded. The expansion rules are constrained by the angle between the gradient field propagation direction and the main variation direction, ensuring the geometric consistency of spatiotemporal parameters; the spatiotemporal elastic capacity matrix generated by the cross-combination simultaneously encodes time decay characteristics and spatial load characteristics in each unit, forming a spatiotemporally coupled carrying capacity assessment basis. The product operation submodule completes the fine calculation of matrix elements: the row vectors continue the road node sequence of the time elastic vector, the column vectors maintain the variation dimension of the spatial elastic tensor, and the product operation of matrix elements reflects the dynamic balance between risk decay and spatial reorganization; the diagonal elements specially retain the original gradient information of the compression start position to ensure the data traceability of key risk points.

[0177] The aforementioned cross-combination submodule constructs a precise quantization framework for spatiotemporal parameter fusion through multi-level collaborative computation. Its overall technical significance is reflected in the following aspects: The first operand preprocessing unit forcibly aligns the node sequence of the time elastic vector with the propagation direction of the risk gradient field, ensuring that the distribution pattern of the time adjustment parameters inherits the spatial characteristics of the original gradient field; the output directional constraint vector retains the key gradient information of the compression starting point, providing a physically consistent time reference for subsequent dimensional expansion. The second operand direction adaptation unit adjusts the main variation direction of the spatial elastic tensor through orthogonal projection, making it form a fixed angle with the time vector direction determined by the covariant relationship of flow fluctuations; while maintaining the nonlinear characteristics of the pressure situation field, it establishes directional compatibility between spatial recombination potential and time parameters, laying a geometric foundation for cross-dimensional computation. The dimensional expansion rule construction unit dynamically adjusts the expansion ratio of spatiotemporal parameters according to the sinusoidal component of the angle: the linear extension of the time vector along the gradient field direction preserves the continuity of risk decay; the nonlinear expansion of the spatial tensor strengthens the fluctuation characteristics of the main variation direction; the adaptive adjustment of the expansion ratio ensures that the upgraded strip matrix and block matrix meet the dimensional matching requirements of subsequent cross-combination. Cross-combination generates capacity units through a strict correspondence between row and column indices, precisely mapping the road node sequence to the main mutation level. The product operation of spatiotemporal control parameters forces a dynamic balance between the gradient change rate and the recombination coefficient. The generated capacity units simultaneously encode decay rates and load fluctuations, forming a spatiotemporal coupling index with clear physical meaning. Diagonal characteristic-preserving units embed the original gradient intensity information into the diagonal of the final matrix by compressing the marking and assignment of the row and column intersections corresponding to the starting point; ensuring the data integrity of key risk points, so that the evaluation results of spatiotemporal elastic capacity always include the initial risk propagation characteristics.

[0178] In summary, this embodiment establishes a strict correspondence between the time vector and the spatial tensor in terms of direction, dimension, and intensity; it achieves geometrically compatible transformation of spatiotemporal parameters through angle-driven expansion rules; the generated spatiotemporal resilience matrix simultaneously satisfies road network topology constraints, risk gradient propagation laws, and traffic fluctuation characteristics; and the diagonal preservation mechanism ensures data traceability of key nodes and maintains physical consistency in the assessment process. The final output matrix provides a quantitative benchmark for traffic network resilience assessment that possesses both temporal adjustability and spatial reconfigurability, with all computational steps based on the intrinsic correlation of core features such as gradient field direction, principal variation angle, and expansion ratio.

[0179] Example 6

[0180] This embodiment provides an intelligent transportation testing platform based on human-vehicle-road-cloud collaboration. Building upon embodiment 4, the dimension expansion rule construction unit further includes:

[0181] The directional constraint sequence expansion subunit is configured as follows: a time elastic vector carrying gradient intensity information is axially replicated and extended along the propagation direction of the risk gradient field. The original time adjustment parameter of each node is used as the core value. The parameters of adjacent nodes generate a decay margin band according to the integral characteristics of the risk decay curve. The extension width is determined by the gradient intensity decay rate.

[0182] The strip-shaped structure forming sub-unit is configured such that: the parameter sequence after axial replication and extension is fixed at the position of the core value as the main ridge of the matrix through the direction locking mechanism, and the attenuation margin band is symmetrically expanded on both sides along the propagation direction to form a strip matrix with directional bandwidth gradient characteristics.

[0183] The variant axial fission subunit is configured as follows: the spatial elastic tensor triggers dimensional fission along the main mutation direction, the original tensor dimension serves as the basic lattice, and each lattice point generates a curvature multiplication factor based on the coupling strength with the road network pressure situation field. The curvature multiplication factor drives the lattice to expand nonlinearly along the main direction.

[0184] The expression for the curvature proliferation factor is:

[0185]

[0186] In the formula, This represents the curvature multiplication factor, used to quantify the intensity of lattice point expansion; It represents the coupling strength between lattice points and the road network pressure situation field. It is a scalar value that reflects the interaction strength between lattice points and the road network pressure situation field. This represents the proliferation sensitivity coefficient, used to control the degree of nonlinearity of expansion; it is set by system parameters, and the larger the value, the more sensitive the coupling strength is to the influence of expansion. This represents an exponential function, used to ensure the nonlinear nature of the expansion, meaning that a small change in coupling strength can lead to a large change in the degree of expansion.

[0187] The blocky topological reconstruction subunit is configured such that the lattice after fission is reorganized through angular constraints, the expansion amplitude is modulated by the sinusoidal component of the orthogonal projection angle, the high curvature region produces a dimensional folding effect to form a dense blocky core, the low curvature region maintains a sparse connection framework, and the whole is constructed into a blocky matrix with a heterogeneous density distribution.

[0188] In this embodiment, the directional constraint sequence expansion subunit implements a risk-aware parameter propagation mechanism. It establishes an axial extension path in the risk gradient field using the time elastic vector of gradient intensity information. This subunit employs an attenuation margin band generation algorithm to ensure strict matching between the parameter extension process and risk attenuation characteristics. The strip structure forming subunit transforms the extension parameters into a matrix structure with directional characteristics. Through main ridge locking and a bilateral symmetrical expansion mechanism, a band matrix with gradually varying bandwidth is formed. This structure is particularly suitable for processing data streams with directional attenuation characteristics. The variable axial fission subunit introduces the dynamic expansion capability of the spatial elastic tensor. Based on the coupling analysis of the pressure situation field, nonlinear fission expansion of dimensions is achieved through a curvature multiplication factor, providing the system with an elastic space to cope with complex situational changes. The block topology reconstruction subunit completes the final dimensional reorganization. A sinusoidal modulation mechanism with included angle constraints is used to generate a heterogeneous density distribution structural feature: a dense block core is formed in the high curvature region to enhance local processing capabilities, while a sparse framework is maintained in the low curvature region to ensure system flexibility.

[0189] In summary, this embodiment achieves adaptive parameter extension in the risk gradient field, precise construction of the directional banded matrix, situational response fission of the elastic tensor, and intelligent reorganization of heterogeneous density block topology through the cascaded operation of four sub-units. It is particularly suitable for complex computational scenarios requiring simultaneous handling of directional propagation, dynamic dimensional expansion, and heterogeneous structure reconstruction, providing a complete solution framework for dynamic modeling of multidimensional data.

[0190] Example 7

[0191] like Figure 5 As shown, this embodiment provides an intelligent transportation testing method based on human-vehicle-road-cloud collaboration, which is based on the intelligent transportation testing platform based on human-vehicle-road-cloud collaboration provided in embodiments 1-6, and includes the following steps:

[0192] Step S1: Extract unstructured features from dispersed vehicle movement trajectories, microscopic road surface conditions monitored by roadside facilities, and macroscopic meteorological data from the cloud; map the extracted unstructured features to a unified spatiotemporal coordinate system, and generate a road surface adhesion coefficient gradient map through a cross-entity association algorithm; based on the road surface adhesion coefficient gradient map, deduce the transmission path and impact range of low adhesion areas during the movement of traffic flow, and output a dynamic risk heat map.

[0193] Further, in step S101, the speed change points in the vehicle trajectory, the abnormal tire noise spectrum sections in the micro road surface conditions collected from the roadside, and the time-varying curve of precipitation intensity in the cloud-based macro meteorological data are extracted as spatiotemporally discrete unstructured features.

[0194] Step S102: Align unstructured features to the road network digital base through spatiotemporal projection, so that speed change points and abnormal tire noise spectrum sections coincide in space, automatically trigger cloud-based macroscopic erosion factor weighting, and generate a dynamic adhesion attenuation field with road grid as the unit, i.e., road surface adhesion coefficient gradient map.

[0195] Step S103: Based on the spatial distribution of low-intensity areas in the road surface adhesion coefficient gradient map, inject the real-time traffic flow density vector, establish a risk transmission channel with the traffic flow direction as the axis, calculate the inertial disturbance factor based on the vehicle mass distribution in the channel, and superimpose the adhesion attenuation gradient to generate a risk propagation vector field; finally, output a dynamic risk heat map with spatiotemporal evolution attributes.

[0196] Step S2: Receive the real-time control status of vehicle steering angle change rate and drive torque gradient, along with passenger travel preferences, and parse them into quantifiable dynamic behavior vectors; fuse multi-vehicle dynamic behavior vectors and dynamic risk heatmaps, construct a virtual decision sandbox at roadside edge nodes, and simulate trajectory conflict points under different collaborative strategies; dynamically generate vehicle group passage sequences based on the conflict resolution results of trajectory conflict points, and issue executable passage instructions after verifying global feasibility through the cloud platform;

[0197] Further, in step S201, the trajectory curvature potential energy is generated by the time integral of the vehicle steering angle change rate, the driving torque gradient is transformed in the frequency domain to form the dynamic response spectrum, and the passenger travel preference is compressed into the spatiotemporal constraint domain; the dynamic behavior vector of each vehicle is output through tensor synthesis.

[0198] Step S202: At the roadside edge node, project the motion state of the multi-vehicle dynamic behavior vectors and overlay them with the hazard contour lines of the dynamic risk heat map. Generate a trajectory conflict focus cloud map through the field strength interference principle and mark the potential collision domains under different cooperative strategies.

[0199] Step S203: Perform energy level attenuation operation on the trajectory conflict focus cloud map, apply passage phase delay to high conflict areas, and inject priority passage factor into low-risk channels; generate vehicle group passage wave sequence with spatiotemporal offset, and after the cloud platform verifies the global manifold continuity of the wave sequence, issue it as an executable passage command bundle.

[0200] Step S3: Based on the generated executable passage instructions, quantify the throughput efficiency and latency threshold of each path node to construct a road network pressure situation field; overlay the traffic demand forecast from the cloud with the dynamic risk heat map to identify the spatiotemporal elastic capacity of key bottleneck areas in the road network pressure situation field; and redefine lane functional boundaries based on elastic capacity.

[0201] Further, in step S301, the vehicle group passage sequence data in the executable passage instructions is discretized in the time dimension, the maximum passage volume per unit time of each path node is extracted as the throughput efficiency base, and the instruction transmission delay between adjacent nodes is recorded synchronously as the time threshold parameter.

[0202] Step S302: Perform tensor product operation on the throughput efficiency base and the time threshold parameter to generate a composite index matrix containing spatial load intensity and temporal congestion degree, and form a continuous road network pressure situation field through three-dimensional interpolation.

[0203] Step S303: After feature alignment between cloud-based traffic demand forecast data and dynamic risk heat map, locate areas exceeding critical values ​​in the road network pressure situation field; calculate the compressible time margin of each bottleneck area based on the gradient change rate of the dynamic risk heat map, and deduce the spatial reorganization potential by combining the traffic fluctuation characteristics of demand forecast to obtain the spatiotemporal elastic capacity.

[0204] Step S304: Convert the spatiotemporal elastic capacity assessment results into lane function adjustment parameters: the time margin corresponds to the adjustable range of the signal phase, the spatial potential determines the variable range of lane markings, and finally outputs a set of lane function boundaries with dynamic constraints.

[0205] The intelligent transportation testing method based on human-vehicle-road-cloud collaboration described in this embodiment achieves integrated verification of multi-source heterogeneous data fusion, group collaborative decision optimization, and dynamic right-of-way allocation, realizing the following technical features: Full-element environmental situation modeling and risk propagation simulation: Through unstructured feature extraction and unified spatiotemporal mapping in step S1, vehicle trajectories, micro-road conditions, and macro-meteorological data are transformed into road surface adhesion coefficient gradient maps, thereby generating dynamic risk heat maps. This solves the problem of missing risk transmission path modeling in low-adhesion areas caused by fragmented environmental data in traditional testing methods. Dynamic conflict prediction and resolution in multi-vehicle collaborative decision-making: Step S2 generates dynamic behavior vectors by analyzing individual vehicle control states and passenger preferences. Combined with the risk heat map, a virtual decision sandbox is constructed at edge nodes to pre-simulate multi-vehicle trajectory conflict points and generate traffic sequences. This achieves vertical decision verification from individual vehicle control to group collaboration, avoiding chain congestion caused by delayed conflict response in actual road tests. The road network resource elastic adaptation and global efficiency optimization, step S3 quantifies the node throughput efficiency and latency threshold based on the passage instructions, superimposes traffic demand forecasting and risk heat map to identify the elastic capacity of bottleneck areas, dynamically reconstructs the lane functional boundaries, and forms a closed-loop verification chain of risk perception-decision generation-resource allocation, which improves the accuracy of the test platform's assessment of the road network's adaptive capability under extreme conditions.

[0206] In summary, this embodiment constructs a complete testing framework covering the environmental perception layer, collaborative decision-making layer, and resource scheduling layer, providing a full-process verification method for intelligent connected transportation systems that supports dynamic risk simulation, multi-agent game simulation, and infrastructure resilient response.

[0207] Example 8

[0208] This embodiment provides an electronic device, including: a processor and a memory; the memory is used to store at least one executable instruction, which causes the processor to perform the functions of the various structures in the test platform described in any one of embodiments 1-6, or causes the processor to perform the test method described in embodiment 7.

[0209] Example 9

[0210] This embodiment provides a computer storage medium storing a computer program. When the program is executed by a processor, it implements the functions of each structure in the test platform described in any one of embodiments 1-6, or implements the test method described in embodiment 7.

[0211] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system.

[0212] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A road intelligent transportation test platform based on human-vehicle-road-cloud collaboration, characterized in that: include: The environmental situation mirror generation system is configured to: extract unstructured features from dispersed vehicle movement trajectories, micro-road conditions monitored by roadside facilities, and macro-meteorological data from the cloud; map the extracted unstructured features to a unified spatiotemporal coordinate system, generate a road surface adhesion coefficient gradient map through a cross-entity association algorithm; and deduce the transmission path and impact range of low adhesion areas during the movement of traffic based on the road surface adhesion coefficient gradient map, and output a dynamic risk heat map. The trajectory decision-making coordination system is configured to receive real-time control status and passenger travel preferences of vehicle steering angle change rate and drive torque gradient, and parse them into quantifiable dynamic behavior vectors. By integrating dynamic behavior vectors of multiple vehicles and dynamic risk heatmaps, a virtual decision sandbox is constructed at roadside edge nodes to simulate trajectory conflict points under different collaborative strategies. Based on the conflict resolution results of trajectory conflict points, a vehicle group passage sequence is dynamically generated. After verifying the global feasibility through a cloud platform, it is issued as an executable passage instruction. The flexible right-of-way reallocation system is configured to: quantify the throughput efficiency and latency threshold of each path node based on executable passage instructions to construct a road network pressure situation field; overlay cloud-based traffic demand forecasts with dynamic risk heat maps to identify the spatiotemporal elastic capacity of key bottleneck areas in the road network pressure situation field; and redefine lane functional boundaries based on elastic capacity.

2. The intelligent road transportation test platform based on human-vehicle-road-cloud collaboration according to claim 1, characterized in that: The environmental situation mirror generation system includes: a multi-source spatiotemporal fragment feature anchoring subsystem, a heterogeneous field strength coupling mapping subsystem, and a risk manifold topology inference subsystem; The multi-source spatiotemporal fragment feature anchoring subsystem is configured to extract the speed change points in the vehicle trajectory, the abnormal tire noise spectrum sections in the micro road surface conditions collected from the roadside, and the time-varying curve of precipitation intensity in the cloud-based macro meteorological data as spatiotemporally discrete unstructured features. The heterogeneous field strong coupling mapping subsystem is configured to: align the unstructured features output by the multi-source spatiotemporal fragment feature anchoring subsystem to the road network digital base through spatiotemporal projection, so that the speed change point coincides with the abnormal section of the tire noise spectrum in space, automatically trigger the cloud macroscopic erosion factor weighting, and generate a dynamic adhesion attenuation field with road grid as the unit, i.e., the road surface adhesion coefficient gradient map. The risk manifold topology deduction subsystem is configured as follows: based on the spatial distribution of low-intensity regions in the road surface adhesion coefficient gradient map output by the heterogeneous field strong coupling mapping subsystem, a real-time traffic flow density vector is injected, a risk transmission channel is established with the traffic flow direction as the axis, the inertial disturbance factor is calculated based on the vehicle mass distribution in the channel, and the adhesion attenuation gradient is superimposed to generate a risk propagation vector field; finally, a dynamic risk heat map with spatiotemporal evolution attributes is output.

3. The intelligent road transportation test platform based on human-vehicle-road-cloud collaboration according to claim 1, characterized in that: The trajectory decision-making collaboration system includes: a manipulation-preference coupling vectorization subsystem, a multi-body risk interference field construction subsystem, and a flexible passage contract generation subsystem; The control-preference coupled vectorization subsystem is configured to: generate trajectory curvature potential energy by time integral of the vehicle steering angle change rate; transform the driving torque gradient into a dynamic response spectrum through frequency domain transformation; compress passenger travel preferences into a spatiotemporal constraint domain; and output the dynamic behavior vector of each vehicle through tensor synthesis. The multi-body risk interference field construction subsystem is configured as follows: at the roadside edge node, the dynamic behavior vectors of multiple vehicles output by the control-preference coupling vectorization subsystem are projected into motion state and superimposed with the hazard contour lines of the dynamic risk heat map output by the environmental state mirror generation system. The trajectory conflict focus cloud map is generated through the field strength interference principle to mark the potential collision domains under different cooperative strategies. The flexible passage contract generation subsystem is configured to: perform energy level attenuation operation on the trajectory conflict focus cloud map generated by the multi-body risk interference field construction subsystem, apply passage phase delay to high conflict areas, and inject priority passage factors into low-risk channels; generate vehicle group passage wave sequence with spatiotemporal offset, and after the cloud platform verifies the global manifold continuity of the wave sequence, issue it as an executable passage command bundle.

4. The intelligent road transportation test platform based on human-vehicle-road-cloud collaboration according to claim 1, characterized in that: The elastic right-of-way reallocation system includes: a node performance matrix construction subsystem, a pressure field generation subsystem, a capacity elasticity assessment subsystem, and a boundary reconstruction subsystem; The node performance matrix construction subsystem is configured to: discretize the vehicle group passage sequence data in the executable passage instructions output by the trajectory decision coordination system in the time dimension, extract the maximum passage volume per unit time of each path node as the throughput efficiency base, and synchronously record the instruction transmission delay between adjacent nodes as the time threshold parameter. The pressure field generation subsystem is configured to: perform tensor product operation on the throughput efficiency base of the node performance matrix construction subsystem and the time threshold parameter to generate a composite index matrix containing spatial load intensity and temporal congestion degree, and form a continuous road network pressure situation field through three-dimensional interpolation. The capacity elasticity assessment subsystem is configured as follows: after feature alignment between cloud-based traffic demand forecast data and dynamic risk heat map output by environmental situation mirror generation system, the region exceeding the critical value is located in the road network pressure situation field generated by pressure field generation subsystem; the compressible time margin of each bottleneck area is calculated based on the gradient change rate of dynamic risk heat map, and the spatial reorganization potential is derived by combining the traffic fluctuation characteristics of demand forecast to obtain spatiotemporal elastic capacity. The boundary reconstruction subsystem is configured to convert the spatiotemporal elastic capacity assessment results into lane function adjustment parameters: the time margin corresponds to the adjustable range of the signal phase, the spatial potential determines the variable range of the lane markings, and finally outputs a set of lane function boundaries with dynamic constraints.

5. A road intelligent transportation test platform based on human-vehicle-road-cloud collaboration according to claim 3, characterized in that: The aforementioned manipulation-preference coupled vectorization subsystem includes: The motion eigenvalue reconstruction module is configured as follows: the rate of change of the vehicle steering angle is integrated over time to generate a continuous curvature manifold; the continuous curvature manifold is used to extract the principal curvature ridges through differential topological transformation; the ridge length and the curvature integral value together constitute a scalar set of trajectory curvature potential energy. The dynamic-constraint tensor field generation module is configured to: convert the driving torque gradient into a frequency domain band distribution via fast spectral remapping; compress passenger travel preferences into a spatiotemporal constraint envelope with boundary rigidity; and project the frequency domain band distribution into the spatiotemporal constraint envelope to form a dynamic response spectrum tensor field. The behavior vector condensation output module is configured to: inject the trajectory curvature potential energy scalar set generated by the motion eigenbase reconstruction module as a basis vector into the dynamic response spectrum tensor field generated by the dynamic-constraint tensor field generation module; stretch the tensor field along the principal curvature ridge direction through curvature-driven deformation; extract the maximum band density region at the constraint envelope boundary; and output a dynamic behavior vector containing spatial trend and intensity attributes. The multi-body risk interferometry field construction subsystem includes: The dynamic behavior vector projection mapping module is configured to project the multi-vehicle dynamic behavior vectors output by the control-preference coupling vectorization subsystem onto the road network digital base. The spatial trend component of each vector determines the main motion axis, and the intensity component is converted into the field intensity distribution along the axis to form the individual vehicle motion field. The risk contour field intensity control module is configured to: deconstruct the hazard contour lines of the dynamic risk heat map output by the environmental situation mirror generation system into a gradient sensitive layer, and perform spatial convolution with the motion field: when the main axis direction of the motion field is orthogonal to the risk gradient direction, interference attenuation is generated; when they are in the same direction, field strength superposition is formed to generate a composite field. The interference fringe feature extraction module is configured as follows: In the composite field modulated by the risk contour field intensity modulation module, when the phase difference between the motion field intensity and the risk gradient exceeds the critical threshold, interference fringe bands are generated, and the intersection area of ​​the fringe bands is marked as the first-level conflict focus; the field intensity superposition saturation area is marked as the second-level conflict focus. The cooperative strategy imprinting injection module is configured to: the passage phase delay coefficient corresponding to different cooperative strategies is applied to the interference fringe stripe generated by the interference fringe feature extraction module, causing the first-level conflict focus to shift spatially, the second-level conflict focus to split energy level, and finally outputting a trajectory conflict focus cloud map carrying the strategy imprint; The aforementioned flexible access contract generation subsystem includes: The conflict energy level spectrum deconstruction module is configured to decompose the trajectory conflict focus cloud map generated by the multibody risk interferometric field construction subsystem into discrete energy level units. High conflict regions are represented as focus clusters with energy level density exceeding the threshold, while low risk channels are represented as interference fringe bands with uniform energy level spacing. The dynamic attenuation operator injection module is configured to: apply anisotropic attenuation to the focal clusters generated by the conflict energy level spectrum deconstruction module, apply an exponential phase delay along the motion axis, and use step-type energy level suppression in the vertical direction; and implant a pulse-type passage factor into the interference fringe region, the amplitude of which is inversely proportional to the energy level interval. The wave sequence primitive generation module is configured as follows: the dynamic attenuation operator injection module generates a deformation field from the trajectory conflict focus cloud map after attenuation, which drives the vehicle motion parameters to redistribute along the attenuation gradient; the vehicle motion parameters in the high conflict area form a time delay differential, and the parameters in the low risk area generate a lead integral, which together constitute the wave sequence primitive with offset. The manifold continuity verification module is configured as follows: the cloud platform maps the wave sequence primitives generated by the wave sequence primitive generation module to the road network topology space and detects the parameter transitions between adjacent wave sequence primitives; when the axial offset derivative and the radial energy level gradient satisfy the bilinear constraint, it is determined that the global manifold is continuous; the wave sequence primitives that pass the manifold verification are compiled into a three-dimensional instruction structure: the time dimension carries the phase delay, the spatial dimension records the passage factor distribution, and the energy dimension encodes the energy level attenuation coefficient, finally forming an executable passage instruction bundle.

6. A road intelligent transportation test platform based on human-vehicle-road-cloud collaboration according to claim 5, characterized in that: The wave sequence primitive generation module includes: The deformation field gradient extraction submodule is configured as follows: the spatial deformation field is generated by the trajectory conflict focus cloud map after the dynamic decay operator injection module decays, the phase delay gradient component is extracted along the motion axis direction, and the energy level suppression gradient component is obtained in the vertical direction. The parameter differential field construction submodule is configured as follows: the energy level density gradient of the focal cluster in the high conflict region drives the construction of the time delay differential field, establishes the first-order derivative relationship in the time dimension along the phase delay gradient direction, and forms the second-order constraint condition in the energy level suppression gradient direction. The parameter integration field generation submodule is configured as follows: the gradient of the energy level interval of the interference fringe in the low-risk region generates the advanced integration field, the amplitude of the pulsed passage factor is used as the integration coefficient, and the reciprocal of the energy level interval determines the integration step size. The offset synthesis submodule is configured as follows: the differential field and the integral field are convolved with tensors in the road network topology space, the time delay differential component and the lead integral component are superimposed along the motion axis, and the energy level gradient components are kept orthogonally constrained; the synthesized offset field is regularized, the axial offset is normalized to the phase angle, and the radial energy level difference is normalized to the weight coefficient; finally, a wave sequence primitive structure with spatiotemporal-energy three-dimensional characteristics is formed.

7. A road intelligent transportation test platform based on human-vehicle-road-cloud collaboration according to claim 4, characterized in that: The capacity resilience assessment subsystem includes: The risk gradient extraction module is configured to: differentiate the rate of change of the values ​​of each pixel in the dynamic risk heat map output by the environmental situation mirror generation system along the road axis, obtain the gradient intensity distribution in the risk propagation direction, and form a risk gradient field. The time margin calculation module is configured to: in the critical region marked in the road network pressure situation field generated by the pressure field generation subsystem, the point corresponding to the maximum slope of the risk gradient field is determined as the compression start position, and the risk decay curve is integrated along the gradient decrease direction, and the length of the integration interval is converted into a compressible time margin. The spatial potential mapping module is configured to: project the difference between the peak and valley values ​​of traffic fluctuations in cloud-based traffic demand forecasting into the risk gradient field to form a spatial reorganization coefficient matrix; perform Hadamard product operation between the spatial reorganization coefficient matrix and the spatial load intensity component of the road network pressure situation field to generate a spatial reorganization potential distribution map. The elastic capacity synthesis module is configured to: expand the compressible time margin output by the time margin calculation module into a time elastic vector along the road network topology; extract the main variation direction of the spatial reorganization potential distribution map generated by the spatial potential mapping module through principal component analysis to form a spatial elastic tensor; and synthesize the spatiotemporal elastic capacity by combining the time elastic vector and the spatial elastic tensor through the Kronecker product. The elastic capacity synthesis module includes: The vector processing submodule is configured to: expand the compressible time margin output by the time margin calculation module along the road network topology, and the time adjustment parameters corresponding to each topology node constitute a time elastic vector; the spatial reorganization potential distribution map generated by the spatial potential mapping module is transformed by principal component transformation, and the extracted variation direction is orthogonally constrained to the propagation direction of the original risk gradient field, and the resulting spatial elastic tensor implies the nonlinear coupling relationship between the peak difference of flow fluctuation and the road network pressure situation field. The cross-combination submodule is configured as follows: the network topology constraint carried by the time elastic vector is used as the first operand, and the mutation direction contained in the spatial elastic tensor is used as the second operand. The dimensional expansion of the first operand and the second operand follows the angle relationship between the gradient field propagation direction and the main mutation direction. The spatiotemporal elastic capacity is generated by cross-combining the expanded dimensions, wherein each capacity unit simultaneously encodes the decay characteristics of the time integral interval and the load characteristics of the spatial recombination coefficient. The product operation submodule is configured as follows: in the final synthesized spatiotemporal elastic capacity matrix, the row vectors inherit the road node sequence of the time elastic vector, the column vectors continue the main variation dimension of the spatial elastic tensor, the matrix element values ​​are determined by the product of the corresponding row and column parameters, and the product operation reflects the dynamic balance relationship between the slope of the risk decay curve and the spatial recombination coefficient. The diagonal elements of the matrix retain the original gradient strength information at the starting position of compression.

8. A road intelligent transportation test platform based on human-vehicle-road-cloud collaboration according to claim 7, characterized in that: The aforementioned cross-combination submodule includes: The first operand preprocessing unit is configured as follows: a time elastic vector based on the road network topology expansion, whose node sorting direction is aligned with the propagation direction of the risk gradient field; the time adjustment parameters corresponding to each node are serialized along the propagation direction to form a one-dimensional vector structure with directional constraints, carrying the gradient intensity information of the compression start position, and retaining the integral characteristics of the risk decay curve. The second operand direction adaptation unit is configured such that: the main variation direction contained in the spatial elastic tensor is adjusted by orthogonal projection to form a fixed angle with the gradient field propagation direction; the angle is determined by the covariant relationship between the directional derivative of the original risk gradient field and the peak difference of the flow fluctuation; the adjusted tensor maintains the nonlinear coupling characteristics with the road network pressure situation field, while ensuring that its dimensional expansion reference is compatible with the direction of the time elastic vector. The dimension expansion rule building unit is configured as follows: the expansion dimension of the time elastic vector is linearly extended along the gradient field propagation direction, and the expansion dimension of the spatial elastic tensor is nonlinearly expanded along the main variation direction; the expansion ratio is adjusted by the sinusoidal component of the included angle: when the included angle approaches orthogonality, the expansion magnitude of the spatial dimension increases; when the included angle decreases, the expansion weight of the time dimension increases; the expanded time vector is upgraded to a strip matrix, and the spatial tensor is upgraded to a block matrix. The capacity unit is generated by cross-combination and is configured as follows: the strip matrix and block matrix after the dimension expansion rule construction unit are combined according to the following rules; The row index of the strip matrix corresponds to the road node sequence, and the column index of the block matrix corresponds to the main mutation dimension level. The value of each capacity unit is determined by the temporal control parameters of the intersecting rows and columns, where the time component contributes to the decay rate and the spatial component contributes to the load fluctuation intensity. The generation function of the unit value is forced to satisfy: the rate of change of capacity in the direction of gradient field propagation is equal to the rate of change of recombination coefficient in the direction of main mutation, forming a dynamic equilibrium constraint. The diagonal characteristic preservation unit is configured to: mark the matrix row and column intersection points corresponding to the compression start position during the cross-combination process of generating capacity units; and write the original gradient intensity values ​​corresponding to the matrix row and column intersection points into the matrix diagonal. The aforementioned dimension expansion rule construction unit includes: The directional constraint sequence expansion subunit is configured as follows: a time elastic vector carrying gradient intensity information is axially replicated and extended along the propagation direction of the risk gradient field. The original time adjustment parameter of each node is used as the core value. The parameters of adjacent nodes generate a decay margin band according to the integral characteristics of the risk decay curve. The extension width is determined by the gradient intensity decay rate. The strip-shaped structure forming sub-unit is configured such that: the parameter sequence after axial replication and extension is fixed at the position of the core value as the main ridge of the matrix through the direction locking mechanism, and the attenuation margin band is symmetrically expanded on both sides along the propagation direction to form a strip matrix with directional bandwidth gradient characteristics. The variant axial fission subunit is configured as follows: the spatial elastic tensor triggers dimensional fission along the main mutation direction, the original tensor dimension serves as the basic lattice, and each lattice point generates a curvature multiplication factor based on the coupling strength with the road network pressure situation field. The curvature multiplication factor drives the lattice to expand nonlinearly along the main direction. The blocky topological reconstruction subunit is configured such that the lattice after fission is reorganized through angular constraints, the expansion amplitude is modulated by the sinusoidal component of the orthogonal projection angle, the high curvature region produces a dimensional folding effect to form a dense blocky core, the low curvature region maintains a sparse connection framework, and the whole is constructed into a blocky matrix with a heterogeneous density distribution.

9. A smart transportation testing method based on human-vehicle-road-cloud collaboration, characterized in that: The testing using the road intelligent transportation test platform based on human-vehicle-road-cloud collaboration as described in any one of claims 1-8 includes the following steps: Step S1: Extract unstructured features from dispersed vehicle movement trajectories, microscopic road surface conditions monitored by roadside facilities, and macroscopic meteorological data from the cloud; map the extracted unstructured features to a unified spatiotemporal coordinate system, and generate a road surface adhesion coefficient gradient map through a cross-entity association algorithm; Based on the road surface adhesion coefficient gradient map, the transmission path and impact range of low adhesion areas during the movement of traffic flow are deduced, and a dynamic risk heat map is output. Step S2: Receive the real-time control status and passenger travel preferences of the vehicle steering angle change rate and drive torque gradient, and parse them into quantifiable dynamic behavior vectors. By integrating dynamic behavior vectors of multiple vehicles and dynamic risk heatmaps, a virtual decision sandbox is constructed at roadside edge nodes to simulate trajectory conflict points under different collaborative strategies. Based on the conflict resolution results of trajectory conflict points, a vehicle group passage sequence is dynamically generated. After verifying the global feasibility through a cloud platform, it is issued as an executable passage instruction. Step S3: Based on the generated executable passage instructions, quantify the throughput efficiency and latency threshold of each path node to construct a road network pressure situation field; overlay the traffic demand forecast from the cloud with the dynamic risk heat map to identify the spatiotemporal elastic capacity of key bottleneck areas in the road network pressure situation field; and redefine lane functional boundaries based on the elastic capacity.

10. An electronic device, comprising: Processor and memory; The memory is used to store at least one executable instruction, which causes the processor to perform the functions of each structure in the intelligent road transportation test platform based on human-vehicle-road-cloud collaboration as described in any one of claims 1-8, or causes the processor to perform the intelligent transportation test method based on human-vehicle-road-cloud collaboration as described in claim 9.

Images