A new energy vehicle electric side step system with road surface pre-perception and anti-skid functions

Abstract

The present application relates to the technical field of new energy vehicles, and provides a new energy vehicle electric side step system with road pre-perception and anti-skid functions. The system synchronously collects the road surface geometry, surface state and user getting-off habits in the side step area through multi-sensor data fusion; based on this, a road model representing the mechanical safety boundary and a grid risk map quantifying the risk of slipping are constructed; through multi-constraint planning, the safety and anti-skid requirements are fused to generate the preliminary extension length and anti-skid angle strategy of the side step; finally, the strategy is adapted and checked in combination with the user's individual behavior characteristics, and the precise execution instruction for driving the side step mechanism is output. The present application realizes the leap from passive deployment to active environment perception and intelligent decision-making of the side step, effectively avoids the risks of missing the step, collision and slipping, and significantly improves the safety and comfort experience of the user getting on and off the vehicle.

Description

Technical Field

[0001] This invention relates to the field of new energy vehicle technology, and in particular to an electric side step system for new energy vehicles with road surface pre-sensing and anti-slip functions. Background Technology

[0002] With the increasing intelligence of new energy vehicles, electric side steps have become an important feature for improving the convenience and experience of getting in and out of the vehicle. However, traditional electric side step systems can usually only execute pre-set fixed trajectory retraction and extension actions based on the door opening and closing signal. This system lacks the ability to perceive the real-time road conditions outside the vehicle, thus exposing inherent safety defects when facing complex real-world getting-out scenarios such as slopes, water accumulation, ice and snow, or obstacles: because it cannot perceive the road slope, the step may tilt after unfolding, creating a risk of stepping into the air; because it cannot detect side obstacles, the mechanism is prone to collisions during unfolding; at the same time, because it cannot recognize low-traction road surfaces, it loses the possibility of early warning or taking anti-slip measures, posing a significant risk of users slipping when getting out of the vehicle. The fundamental limitation of existing electric side step control technology lies in the lack of real-time, multi-dimensional perception of the external environment, as well as perception-based intelligent decision-making and actuator adaptive control capabilities.

[0003] Therefore, there is an urgent need for an intelligent side pedal system that can deeply integrate environmental perception, risk assessment, and personalized adaptation to ensure the safety of users getting on and off the vehicle in real time and improve the comfort experience. Summary of the Invention

[0004] In order to overcome the defects and shortcomings of the existing technology, the present invention provides an electric side step system for new energy vehicles with road surface pre-sensing and anti-slip functions.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] In a first aspect, the present invention provides an electric side step system for new energy vehicles with road surface pre-sensing and anti-slip functions, comprising the following modules:

[0007] Road surface multimodal data acquisition module: Simultaneously collects road surface geometry data, road surface condition data, and human exit behavior data of the side step area of ​​new energy vehicles;

[0008] Geometric structure analysis module: Based on road surface geometry data, through spatial modeling and motion simulation analysis, it constructs the mechanical safety boundary of the side step in the real-time road surface environment;

[0009] Surface condition analysis module: Based on road surface condition data, through surface material identification and risk mapping analysis, a gridded transition risk map is constructed to predict the risk of slipping when getting out of the vehicle;

[0010] Safety and anti-slip constraint fusion module: Through multi-constraint programming, the mechanical safety boundary and the gridded transition risk map are fused to generate a preliminary attitude strategy that includes the initial side step extension length and the initial anti-slip angle compensation value;

[0011] Personalized execution strategy generation module: Based on the collected human getting-out-of-vehicle behavior habit data, through behavior pattern analysis and action inference, it predicts the user's getting-out-of-vehicle action characteristics; and combined with the preliminary posture strategy, through strategy adaptation and constraint verification, it generates side step posture execution instructions.

[0012] According to the above technical solution, the steps of simultaneously collecting road surface geometry data, road surface condition data, and human vehicle exit behavior data in the side step area of ​​new energy vehicles include:

[0013] S110. Based on multi-sensor data acquisition and fusion of the side step area of ​​new energy vehicles, construct road geometric data including a three-dimensional point cloud dataset of the side step area, a sequence of road surface ground clearance measurements, and obstacle contour point clouds, specifically including:

[0014] A millimeter-wave radar installed on the side skirt of the vehicle collects three-dimensional spatial information of the road surface and obstacles, and constructs a three-dimensional point cloud dataset of the side step area; an ultrasonic sensor array installed on the bottom of the vehicle measures the vertical distance between each detection point on the road surface and the fixed reference surface of the vehicle body, and constructs a sequence of road surface ground clearance measurement values; by fusing the data from the millimeter-wave radar and the ultrasonic sensor array, the shape of the obstacle is identified and extracted, and an obstacle contour point cloud is generated.

[0015] S120. Based on the coordinated acquisition of infrared and visible light from the sidewalk area of ​​new energy vehicles, road surface state data including the road surface infrared polarization feature matrix and the road surface visible light texture image are constructed, specifically including:

[0016] Infrared polarization sensors installed on the vehicle's side skirts collect road surface data to obtain feature data that can identify the temperature differences and polarization characteristics of different road surface materials, and construct an infrared polarization feature matrix of the road surface; protective wide-angle cameras installed on the vehicle's side skirts collect color and visual texture information of the road surface to construct a visible light texture image of the road surface.

[0017] S130. Based on personalized data learning from new energy vehicle users, construct human vehicle disembarkation behavior data that includes a user disembarkation speed dataset and historical side step extension length preference records, specifically including:

[0018] Historical user disembarkation data is obtained from the new energy vehicle user information system. Based on the historical user disembarkation data, the average speed of each user disembarkation action is calculated, organized and stored in chronological order as structured data, thereby constructing a user disembarkation speed dataset. Based on the side pedal extension length data recorded in the new energy vehicle user information system, a structured historical side pedal extension length preference record is constructed through time series segmentation and statistical feature extraction.

[0019] According to the above technical solution, the step of constructing the mechanical safety boundary of the side step in the real-time road environment based on road surface geometry data, through spatial modeling and motion simulation analysis, includes:

[0020] S210. By fusing the three-dimensional point cloud dataset of the side tread area with the road surface ground clearance measurement sequence, a three-dimensional road surface mesh model is reconstructed, and the lateral and longitudinal slope angles of the road surface relative to the vehicle body are calculated based on the three-dimensional road surface mesh model. Specifically, this includes:

[0021] The coordinate system is integrated with the three-dimensional point cloud dataset of the side step area and the road surface ground clearance measurement value sequence. Through spatial interpolation, a three-dimensional road surface mesh model covering the side step area is constructed. Based on the least squares method, the three-dimensional mesh model is locally fitted to the plane. By calculating the projection component of the fitting plane normal vector in the vehicle body coordinate system, the lateral slope angle and longitudinal slope angle of the road surface relative to the vehicle body are obtained.

[0022] S220. Segment and calculate the obstacle contour point cloud to obtain the obstacle's position coordinates and size parameters, specifically including:

[0023] Cluster analysis and parameter calculation are performed on the obstacle contour point cloud. A density-based spatial clustering algorithm is used to segment the point cloud to distinguish independent obstacles. Then, for each segmented independent obstacle point cloud cluster, the minimum and maximum coordinate values ​​of these points in the X, Y, and Z axes of the vehicle coordinate system are determined by traversing the three-dimensional coordinates of all points in the independent obstacle point cloud cluster. A boundary cube aligned with the coordinate axes is defined by these six extreme values, which is the three-dimensional axial bounding box of the obstacle. The center coordinates of the bounding box are used as the position coordinates, and the length, width, and height of the bounding box are used as size parameters to obtain the position coordinates and three-dimensional size parameters of the obstacle.

[0024] S230. Combining the lateral slope angle, longitudinal slope angle, and the position coordinates and dimensions of the obstacle, the deployment process is simulated using a kinematic model of the side step mechanism. Collision interference detection and static stability analysis are performed to obtain the maximum allowable extension length. A suggested avoidance area is then delineated based on the outer contour of the obstacle, collectively forming a mechanical safety boundary, specifically including:

[0025] Based on the preset kinematic model of the side step mechanism, the lateral slope angle, longitudinal slope angle and obstacle parameters are used to simulate the process of the side step from retracting to fully unfolding with a fixed step length. In each step, the spatial pose of the mechanism components is calculated and compared with the spatial position of the three-dimensional axial bounding box of the obstacle to determine whether a collision occurs, thereby identifying the critical extension length at which a collision occurs.

[0026] Meanwhile, based on the current slope angle and the position of the center of gravity of the mechanism, the static overturning moment of the side step under different extension lengths is calculated, the static stability of the side step mechanism is evaluated, and the critical extension length based on stability requirements is identified.

[0027] Based on the combined collision and stability analysis results, the smaller of the two critical extension lengths is taken as the maximum permissible extension length. At the same time, the outer contours of all obstacles are merged on the horizontal plane after being expanded with a safety margin to form a suggested avoidance area. The maximum permissible extension length and the suggested avoidance area together constitute the mechanical safety boundary.

[0028] According to the above technical solution, the step of constructing a gridded transition risk map for predicting the risk of slipping after getting off a vehicle, based on road surface condition data and through surface material identification and risk mapping analysis, includes:

[0029] S310. Spatial registration and grid division are performed on the road surface infrared polarization feature matrix and the road surface visible light texture image to generate a multispectral feature grid in which each grid cell contains infrared polarization feature data and visible light texture data. Specifically, this includes:

[0030] The road surface infrared polarization feature matrix and the road surface visible light texture image are registered in a spatial coordinate system and uniformly divided into grid cells of the same size and coordinates. Each grid cell contains both infrared polarization feature data from the road surface infrared polarization feature matrix and visible light texture data from the road surface visible light texture image, thereby generating an aligned multispectral feature grid covering the side step area.

[0031] S320. Obtain infrared polarization feature data and visible light texture data of grid cells in the multispectral feature grid to generate a fused feature vector, input it into the material classification model to obtain the road surface material type; extract grayscale statistical features from the visible light texture data as the texture roughness index of the grid cells; combine the road surface material type and texture roughness index to determine the potential adhesion coefficient of the grid cells, specifically including:

[0032] For each grid cell in the aligned multispectral feature grid, the polarization features in the infrared polarization feature data and the texture features in the visible light texture data of the grid cell are extracted to form a fused feature vector. The fused feature vector is input into the material classification model, and then the road surface material type of the grid cell is output. At the same time, based on the texture features in the visible light texture data, the texture roughness index of the grid cell is calculated by the gray-level co-occurrence matrix algorithm. Combining the road surface material type and the texture roughness index, the potential adhesion coefficient of the grid cell is obtained by querying the preset material roughness adhesion coefficient mapping relationship database.

[0033] S330. Determine the transition risk level of each grid cell in the multispectral feature grid based on the road surface material type and potential adhesion coefficient. Calculate the basic anti-skid angle compensation value for each grid cell based on its transition risk level and potential adhesion coefficient. Integrate the material type, transition risk level, and basic anti-skid angle compensation value information of all grid cells to generate a gridded transition risk map, specifically including:

[0034] Based on the road surface material type and potential adhesion coefficient of each grid cell, and according to the preset risk rating rules centered on the potential adhesion coefficient threshold, the transition risk level of each grid cell is determined. Then, based on the transition risk level and potential adhesion coefficient of each grid cell, the corresponding basic anti-skid angle compensation value is calculated through the anti-skid angle compensation function aimed at compensating for insufficient adhesion. Finally, the location, road surface material type, transition risk level and basic anti-skid angle compensation value information of all grid cells are integrated to generate a structured gridded transition risk map.

[0035] According to the above technical solution, the step of fusing the mechanical safety boundary and the gridded transition risk map through multi-constraint programming to generate a preliminary attitude strategy including the preliminary side step extension length and the preliminary anti-slip angle compensation value includes:

[0036] S410. Set the maximum permissible extension length in the mechanical safety boundary as a rigid constraint, and merge the suggested avoidance area in the mechanical safety boundary with the grid area marked as high risk in the transition risk map into a priority avoidance area, specifically including:

[0037] The maximum allowable extension length in the mechanical safety boundary is used as a rigid constraint. The suggested avoidance area in the mechanical safety boundary for avoiding obstacles is merged with the grid cell area marked as high risk in the gridded transition risk map and defined together as the priority avoidance area where the side step body is prohibited from falling.

[0038] S420. Within the space satisfying the rigid constraints, select all grids from the transition risk map that are not covered by the priority avoidance area and have a transition risk level of low or medium risk to form a candidate support grid set, specifically including:

[0039] Within the side-step reachable space that satisfies the rigid constraints, all grids that are not covered by the priority avoidance area and whose transition risk level is low or medium in the transition risk map are selected to form a candidate support grid set.

[0040] S430. Optimize the candidate support grid set based on multi-constraint programming. The planning objectives include minimizing the transition risk level of the selected grid and minimizing the absolute value of the basic anti-slip angle compensation value corresponding to the selected grid. An optimal support grid is determined by calculating the weighted score of each grid. Based on the position coordinates of this optimal support grid in the transition risk map and the recorded basic anti-slip angle compensation value, the corresponding initial side tread extension length and initial anti-slip angle compensation value are determined, specifically including:

[0041] Multi-constraint programming is performed on the candidate support grid set. The programming objectives include minimizing the sum of the quantized transition risk levels of the selected grids and minimizing the absolute value of the basic anti-slip angle compensation value corresponding to the selected grids. By calculating the weighted score of each candidate support grid, an optimal support grid is determined from the candidate support grid set. Based on the position coordinates of the optimal support grid in the transition risk map, the initial side step extension length is calculated using the kinematic model of the side step mechanism. The basic anti-slip angle compensation value recorded by the optimal support grid in the transition risk map is directly used as the initial anti-slip angle compensation value. The initial side step extension length and the initial anti-slip angle compensation value together constitute the initial attitude strategy.

[0042] According to the above technical solution, the steps of predicting user disembarkation action characteristics based on collected human disembarkation behavior data, through behavior pattern analysis and action deduction, and generating side pedal posture execution commands by combining preliminary posture strategies and through strategy adaptation and constraint verification include:

[0043] S510. Based on the user's dismount speed dataset, determine the user's stability requirement intensity level, and perform statistical analysis on historical side pedal extension length preference records to determine the user's comfort length range; integrate the stability requirement intensity level and the comfort length range to generate structured predicted user dismount action features, specifically including:

[0044] Based on users' historical dismounting action data, the user's stability requirement intensity level is determined by directly analyzing the dismounting speed dataset; the statistical mean and standard deviation of the length value sequence in the historical side pedal extension length preference record are calculated, and the user's comfortable length range is determined based on the mean and standard deviation; the stability requirement intensity level and the comfortable length range are encapsulated into a data structure containing the level and range fields to generate structured predicted user dismounting action features.

[0045] S520. Extract the comfortable length range from the predicted user disembarkation action features, limit the initial side pedal extension length within the comfortable length range, and obtain the comfort adjustment length; based on the stability requirement intensity level, determine the angle adjustment coefficient through a preset level coefficient mapping relationship, and multiply the initial anti-slip angle compensation value by the angle adjustment coefficient to obtain the stability adaptation angle, specifically including:

[0046] Based on the comfortable length range in the predicted user dismounting action characteristics, the initial side step extension length in the initial posture strategy is adjusted: if the initial side step extension length is less than or equal to the lower limit of the comfortable length range, the initial side step extension length is set to the lower limit of the comfortable length range; if the initial side step extension length is greater than or equal to the upper limit of the comfortable length range, the initial side step extension length is set to the upper limit of the comfortable length range; otherwise, the initial side step extension length remains unchanged, thus obtaining the comfort adjustment length. At the same time, based on the stability requirement intensity level in the predicted user dismounting action characteristics, the angle adjustment coefficient is obtained according to the preset level coefficient mapping relationship. The initial anti-slip angle compensation value in the initial posture strategy is multiplied by the angle adjustment coefficient to obtain the stability adaptation angle.

[0047] S530. Perform constraint verification on the comfort adjustment length and stability adaptation angle, constraining the comfort adjustment length to not exceed the maximum allowable extension length within the mechanical safety boundary, and constraining the stability adaptation angle within the angle adjustment range of the side pedal mechanism; combine the final comfort adjustment length and final stability adaptation angle determined after this constraint verification to generate a side pedal posture execution command that drives the side pedal mechanism, specifically including:

[0048] The comfort adjustment length and stability adaptation angle are constrained and verified: the comfort adjustment length is compared with the maximum allowable extension length in the mechanical safety boundary. If the comfort adjustment length is greater than the maximum allowable extension length, the comfort adjustment length is corrected to the maximum allowable extension length; otherwise, the comfort adjustment length remains unchanged.

[0049] The stability adaptation angle is compared with the angle adjustment range of the side pedal mechanism. If the stability adaptation angle is greater than or equal to the upper limit of the side pedal mechanism's angle adjustment range, the stability adaptation angle is corrected to the upper limit of the side pedal mechanism's angle adjustment range; if the stability adaptation angle is less than or equal to the lower limit of the side pedal mechanism's angle adjustment range, the stability adaptation angle is corrected to the lower limit of the side pedal mechanism's angle adjustment range; otherwise, the stability adaptation angle remains unchanged. The final comfort adjustment length determined after this constraint verification and the final stability adaptation angle are combined to generate the side pedal posture execution command that drives the side pedal mechanism.

[0050] Secondly, this application provides an electronic device, including a processor and a memory, wherein the memory stores a computer program that can be called by the processor, and the processor executes an electric side pedal system for a new energy vehicle with road surface pre-sensing and anti-slip functions by calling the computer program stored in the memory.

[0051] Thirdly, this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to execute a new energy vehicle electric side pedal system with road surface pre-sensing and anti-slip functions.

[0052] Compared with the prior art, this application has the following advantages and beneficial effects:

[0053] This application uses multi-sensor fusion to perceive road geometry and surface conditions in real time, and integrates user disembarkation habit data to construct a mechanical safety boundary and transition risk map. Through multi-constraint planning and personalized adaptation, it generates side pedal execution commands, realizing active safety adjustment and anti-slip optimization of side pedal extension posture under complex road conditions. This fundamentally improves user disembarkation safety and comfort, forming a closed-loop intelligent control system from environmental perception and risk decision-making to personalized execution. Attached Figure Description

[0054] Figure 1 This is a schematic diagram of a new energy vehicle electric side step system with road surface pre-sensing and anti-slip function provided in an embodiment of this application;

[0055] Figure 2 This is a schematic diagram of the process of constructing mechanical safety boundaries using the geometric structure analysis module provided in this application embodiment;

[0056] Figure 3 This is a schematic diagram of the process of constructing a transition risk map using the surface state analysis module provided in this application embodiment;

[0057] Figure 4 This is a flowchart illustrating the process by which the safety and anti-slip constraint fusion module provided in this application performs multi-constraint planning and generates a preliminary attitude strategy.

[0058] Figure 5 This is a flowchart illustrating the process of the personalized execution strategy generation module performing strategy adaptation and constraint verification provided in this application embodiment. Detailed Implementation

[0059] The technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solution of this application, rather than limitations thereof. In the absence of conflict, the embodiments and technical features in the embodiments can be combined with each other.

[0060] Please see Figure 1 , Figure 1 This is a schematic diagram of a new energy vehicle electric side step system with road surface pre-sensing and anti-skid functions provided in an embodiment of this application, specifically including the following modules:

[0061] Road surface multimodal data acquisition module: Simultaneously collects road surface geometry data, road surface condition data, and human exit behavior data of the side step area of ​​new energy vehicles;

[0062] In this embodiment, the simultaneous collection of road surface geometry data, road surface condition data, and human exit behavior data in the side step area of ​​new energy vehicles includes the following specific contents:

[0063] S110. Based on the multi-sensor data acquisition and fusion of the side step area of ​​new energy vehicles, construct road geometric data including a three-dimensional point cloud dataset of the side step area, a sequence of road surface ground clearance measurements, and obstacle contour point clouds.

[0064] The raw echo signal collected by the millimeter-wave radar installed on the vehicle side skirt is acquired. This raw echo signal is then demodulated, filtered, and processed to generate a point cloud, resulting in a set of three-dimensional spatial coordinates representing reflection points on the road surface and obstacle surfaces within the side skirt coverage area. Outlier removal and noise filtering are then applied to this three-dimensional spatial coordinate set to form a structured three-dimensional point cloud dataset of the side skirt area.

[0065] The flight time data collected by each detection unit of the ultrasonic sensor array installed on the bottom of the vehicle body is acquired. Based on the physical principle that the propagation speed of ultrasound in air is proportional to the flight time, the flight time data of each detection unit is converted into the vertical distance between the road surface and a preset fixed reference plane of the vehicle body at that detection point. According to the physical layout order of the sensor array, the vertical distance values ​​of all detection points are arranged into an ordered sequence of road surface ground clearance measurement values.

[0066] First, the three-dimensional point cloud dataset of the side step area and the road surface ground clearance measurement sequence are spatiotemporally registered. The point cloud is unified to the vehicle coordinate system through coordinate transformation. The timestamps of the two sets of data are synchronized and aligned based on the system clock. Then, a grid of fixed size is divided on the XY plane of the vehicle coordinate system. The variance of the Z coordinate of the registered point cloud falling into each grid is calculated. Grids with variance greater than the height variance abnormality threshold are marked as height abnormality areas. The height variance abnormality threshold is set based on the statistical upper limit value of the height variance of the point cloud of flat paved road surface.

[0067] At the same time, the difference between the vertical distances measured by each two adjacent detection points in the gap sequence is calculated, and the points whose absolute difference is greater than the gap change judgment threshold are marked as gap change points. The gap change judgment threshold is determined by the expected height difference of typical curbs or pothole edges and the measurement error of ultrasonic sensors, and is mapped to the corresponding position on the XY plane of the vehicle body coordinate system.

[0068] By combining the mapping locations of all height anomaly areas and gap mutation points, candidate obstacle regions are delineated. A clustering algorithm based on Euclidean distance is used to set the neighborhood search radius according to the average spacing of the point cloud. By iteratively merging spatially neighboring points, the point cloud located in these regions is divided into independent obstacle point cloud clusters.

[0069] Finally, to extract the projected contours of obstacles on the horizontal plane and delineate the avoidance zone, each independent obstacle point cloud cluster is projected onto the XY plane of the vehicle coordinate system, forming a two-dimensional point set. Then, the Graham scan algorithm is used to calculate the convex hull of each point set, obtaining an ordered sequence of convex hull vertices. Based on this vertex sequence, corresponding ordered spatial boundary points are extracted from the 3D point cloud dataset of the side tread area to generate an obstacle contour point cloud describing the horizontal shape of the obstacle.

[0070] S120. Based on the coordinated acquisition of infrared and visible light in the side step area of ​​new energy vehicles, road surface state data including road infrared polarization feature matrix and road visible light texture image are constructed.

[0071] The raw data stream is acquired synchronously from infrared polarization sensors installed on the vehicle side skirts. The raw data stream is demodulated, temperature compensated, and non-uniformity corrected. The infrared radiation intensity value and polarization degree value corresponding to each pixel are extracted. The processed pixel data is directly mapped and arranged into a two-dimensional matrix according to the row and column order of the physical position of each infrared polarization sensor in the sensor focal plane array. The row and column indices of this two-dimensional matrix correspond one-to-one with the spatial grid position of the road surface area in the vehicle coordinate system, thereby generating a road surface infrared polarization feature matrix. Each element in this two-dimensional matrix contains a pair of values: infrared radiation intensity value and polarization degree value.

[0072] Acquire raw image data synchronously from a protective wide-angle camera mounted on the vehicle's side skirt. Perform lens distortion correction and color space standardization on the raw images. Crop the corrected images according to the geometric correspondence between the camera's field of view and the side skirt area, and achieve spatial alignment through projection transformation between the image and the road surface to generate a standardized visible light texture image of the road surface.

[0073] S130. Based on personalized data learning of new energy vehicle users, construct human vehicle disembarkation behavior data including user disembarkation speed dataset and historical side step extension length preference records.

[0074] Historical user disembarkation data is obtained from the new energy vehicle user information system. Based on this data, the average speed of each disembarkation action is calculated, organized and stored chronologically as structured data, thus constructing a user disembarkation speed dataset. Historical side pedal extension length data is preprocessed, first sorted by timestamp to form a time series; the processed time series is then divided into several length segments according to a preset time window; statistical features are extracted from each segment, including segment mean, segment variance, and record frequency; the feature data from all segments are integrated to generate a structured historical side pedal extension length preference record.

[0075] Geometric structure analysis module: Based on road surface geometry data, through spatial modeling and motion simulation analysis, it constructs the mechanical safety boundary of the side step in the real-time road surface environment;

[0076] In this embodiment, as Figure 2 As shown, based on road surface geometry data, and through spatial modeling and motion simulation analysis, the mechanical safety boundary of the side step in the real-time road environment is constructed through the following specific steps:

[0077] S210. By integrating the three-dimensional point cloud dataset of the side step area with the road surface ground clearance measurement sequence, a three-dimensional road surface mesh model is reconstructed and generated. Based on the three-dimensional road surface mesh model, the lateral slope angle and longitudinal slope angle of the road surface relative to the vehicle body are calculated.

[0078] First, all 3D points in the 3D point cloud dataset of the side tread area, as well as the 3D coordinate points generated from the road surface ground clearance measurement sequence, are uniformly transformed into the vehicle coordinate system to form all valid data points for road reconstruction. Each measurement value in the road surface ground clearance measurement sequence corresponds to an ultrasonic sensor detection unit installed on the bottom of the vehicle. Its horizontal coordinates are obtained from the known installation location of this unit, and combined with its measured vertical distance value, a data point with 3D coordinates is constructed. Each data point includes horizontal coordinates. with height coordinates .

[0079] A continuous digital elevation surface of the road surface is reconstructed using an inverse distance weighted interpolation algorithm based on a regular grid. A rectangular grid covering the side tread region is defined on the XY plane of the vehicle coordinate system, with grid node coordinates as follows: .

[0080] For any grid node Using it as the center, set a fixed spatial search radius. All known data points within this circular search area (from the 3D point cloud dataset of the side tread area and the sequence of road surface ground clearance measurements) will be included in the calculation. Search radius. The settings are determined based on the minimum resolvable support area size of the side pedal mechanism and the average spatial density of sensor data.

[0081] For any grid node, the corresponding road surface height value The following formula is used to calculate: ; in, The target point estimate, i.e., at the target point The interpolation result at the location; For the first The measured values ​​of a known sample point are the data basis for the model to perform spatial inference; For the first From known sample points to target point Euclidean distance, The value is obtained by calculating the coordinate difference between the known sample point and the target mesh node in the horizontal coordinate system of the vehicle body, and by applying the Pythagorean theorem. This distance value directly determines the weight of the known sample point in the inverse distance weighted interpolation. The weights are distance-based, and the weights are inversely proportional to the square of the Euclidean distance; the numerator term... Indicates the first The contribution of the measured value of each sample point according to its weight; denominator term Then, the summation and normalization of all weight contributions are performed to ensure that the sum of the weights at each point is 1, thereby making the final estimate... It is a weighted average of the sample values, and it is kept within a reasonable range.

[0082] Triangular patches are formed by connecting all adjacent mesh nodes in the XY plane of the vehicle coordinate system, and the road surface height value corresponding to each mesh node is used as the basis for the process. The three-dimensional spatial coordinates of the mesh nodes are determined to construct a complete three-dimensional mesh model of the road surface represented by triangular meshes. Each triangular facet in the three-dimensional mesh model of the road surface represents a continuous infinitesimal element in the actual road surface, and its vertex set fully characterizes the spatial geometry of the road surface in the side-step area.

[0083] Based on the completed 3D mesh model of the road surface The process then moves to the road surface attitude analysis stage. First, based on the known design dimensions of the side pedal mechanism and the range of joint motion, the maximum theoretical coverage area (i.e., theoretical motion envelope) of the outer edge of the pedal on the horizontal plane is determined, and this is applied to the three-dimensional mesh model of the road surface. On the corresponding XY plane of the vehicle body coordinate system, a rectangular analysis region is projected to match it. The length and width of this rectangle are determined by the horizontal projection range of the theoretical motion envelope.

[0084] Subsequently, the 3D coordinates of all triangular mesh vertices within the rectangular analysis area are directly read and used as input data for plane fitting using the least squares method. Specifically, to calculate the plane that best represents the overall slope trend of the road surface in this area from the discrete vertex coordinates, the following plane equation is used as the fitting target: ; in, Let be the parameters of the plane equation to be fitted. This represents the predicted height value calculated based on the planar model. The goal of fitting is to find a set of parameters that minimizes the sum of the squared vertical distances between the plane and all data points, i.e., solving the following optimization problem: ; in, That is, the objective function to be minimized; The optimization objective is to find the one that makes The parameter with the smallest value ; This refers to the observed road height at that vertex, provided by the three-dimensional road surface mesh model. Then, according to the plane equation The predicted height value is calculated at this point; therefore, the technical process of plane fitting can be summarized as: adjusting parameters through optimization algorithms. So that the above objective function When the value of is minimized, the corresponding plane is the best-fit plane.

[0085] The optimal parameters can be obtained directly by solving the normal equation corresponding to this least squares problem. and At this point, the normal vector of the best-fit plane can be expressed in the vehicle body coordinate system as: The normal vector represents the spatial orientation of the fitted plane.

[0086] The calculation process of the longitudinal slope angle is as follows: First, based on the component A of the normal vector on the X-axis and the component -1 on the Z-axis, the slope of the fitting plane along the front and rear directions of the vehicle body is determined. Then, the corresponding angle value is obtained by the arctangent operation of the slope. This angle is defined as the longitudinal slope angle, and its physical meaning is the pitch angle of the road surface in the direction of vehicle travel.

[0087] The calculation process of the lateral slope angle is as follows: Based on the component B of the normal vector on the Y-axis and the component -1 on the Z-axis, the tilt slope of the fitting plane along the left and right directions of the vehicle body is determined. Then, the corresponding angle value is obtained by the arctangent operation of the slope. This angle is defined as the lateral slope angle, and its physical meaning is the side tilt angle of the road surface in the lateral direction of the vehicle.

[0088] S220. Segment and calculate the point cloud of the obstacle outline to obtain the position coordinates and size parameters of the obstacle;

[0089] First, the obstacle contour point cloud is preprocessed. By statistically analyzing the spatial distances between points in the point cloud, sparse outliers are filtered out to eliminate sensor noise. Then, a density-based spatial clustering algorithm is used to segment the preprocessed point cloud. This algorithm automatically aggregates adjacent points with sufficient density to meet connectivity conditions into independent point cloud clusters based on the three-dimensional spatial distribution of the obstacle contour point cloud in the vehicle coordinate system, while filtering out sparse and discrete noise points. Through this process, the overall point cloud scene covering the side step area is segmented into multiple spatially separated point cloud clusters, each corresponding to a potential obstacle that exists independently in physical space.

[0090] For each segmented point cloud cluster, the minimum and maximum coordinates of all its points on the X, Y, and Z axes of the vehicle coordinate system are directly read to determine its spatial distribution range. Based on this coordinate range, a three-dimensional axial bounding box with sides parallel to the coordinate axes is constructed for the point cloud cluster. The coordinates of the geometric center point of this three-dimensional axial bounding box are taken as the position coordinates of the obstacle, and the side lengths in the X, Y, and Z directions are taken as the length, width, and height dimensions of the obstacle, respectively.

[0091] S230. Combining the lateral slope angle, longitudinal slope angle, and the position coordinates and size parameters of the obstacle, the deployment process is simulated through the kinematic model of the side step mechanism. Collision interference detection and static stability analysis are performed to obtain the maximum allowable extension length. Combined with the outer contour of the obstacle, a suggested avoidance area is delineated, which together constitutes the mechanical safety boundary.

[0092] First, a pre-established kinematic model of the side pedal mechanism is invoked. This model precisely describes the geometric connections and relative motion constraints between all moving components of the side pedal. The input to the model is the side pedal extension length. For any given side pedal extension length value, the kinematic model can, through simulation calculation, output the accurate three-dimensional position and spatial attitude of each moving component in the side pedal mechanism in the vehicle body coordinate system.

[0093] Using a kinematic model of the side pedal mechanism, the entire deployment process from the fully retracted state to the designed maximum length is simulated step-by-step with a fixed step size. In each discrete step of the simulation, the specific spatial pose of each moving component of the side pedal mechanism at the current extension length is obtained. Based on this, corresponding 3D collision bounding boxes are established for components prone to interference. Subsequently, the collision risk is assessed by detecting the interference between the 3D collision bounding box of the side pedal component and the 3D axial bounding box of the obstacle. Once interference is detected, it indicates a critical state for collision, and the corresponding extension length value is recorded as the critical extension length value for interference.

[0094] Simultaneously performing static stability analysis, the kinematic model of the side pedal mechanism can provide the overall center of gravity position of the side pedal mechanism under different extension lengths. This is combined with the longitudinal slope angle of the current road surface. With lateral slope angle Calculate the static overturning moment generated by the gravity of the side step mechanism about its inner support hinge axis. The formula for calculating this static overturning moment is: ; in: This is the static overturning moment; The total weight of the side pedal moving parts; The extension length of the side step refers to the horizontal displacement of the outer edge of the electric side step relative to its fixed mounting point on the vehicle body. This is the vector distance from the overall center of gravity of the mechanism to the axis of the inner support hinge, given the current extension length. For the comprehensive slope angle and Then, the effective overturning angle formed by the direction of gravity relative to the support axis is calculated. The calculated overturning moment is compared with a preset stability safety threshold. This stability safety threshold is set based on the minimum destructive torque that the key load-bearing components of the side pedal mechanism can withstand in terms of materials and structure, and is comprehensively determined with reference to the safety redundancy requirements for dynamic loads in relevant vehicle industry safety standards. During the simulation, when the overturning moment first exceeds this safety threshold, the corresponding extension length is recorded as the critical extension length based on stability requirements.

[0095] After simulating the entire deployment process, the smaller of the two critical extension lengths obtained from the above collision detection and stability analysis is selected as the maximum allowable extension length.

[0096] Finally, the projected contours of all identified obstacles on the horizontal plane are extended outward by a preset safety margin. The value of the safety margin takes into account the minimum movement clearance of the side pedal mechanism, the measurement error of the sensor system, and the typical range of motion of the user, so as to ensure that the side pedal maintains a sufficient safe distance from the obstacle during actual deployment and user use. These extended areas are then merged to form a complete two-dimensional polygonal area, which is defined as the recommended avoidance area.

[0097] Surface condition analysis module: Based on road surface condition data, through surface material identification and risk mapping analysis, a gridded transition risk map is constructed to predict the risk of slipping when getting out of a vehicle.

[0098] In this embodiment, as Figure 3 As shown, based on road surface condition data, and through surface material identification and risk mapping analysis, a gridded transition risk map for predicting the risk of slipping after getting out of a vehicle is constructed, including the following specific steps:

[0099] S310. Spatial registration and grid division are performed on the infrared polarization feature matrix of the road surface and the visible light texture image of the road surface to generate a multispectral feature grid in which each grid cell contains infrared polarization feature data and visible light texture data.

[0100] Based on the known relative installation position and angle relationship between the infrared polarization light sensor and the protective wide-angle camera, the visible light texture image of the road surface is geometrically transformed and resampled to achieve spatial alignment with the infrared polarization feature matrix of the road surface on the XY plane of the vehicle body coordinate system, thereby completing the spatial registration of the infrared polarization feature matrix of the road surface and the visible light texture image of the road surface.

[0101] Subsequently, a rectangular grid covering the side tread unfolded area is defined on the registered coordinate plane. The side length of the grid cells is determined with reference to the theoretical ground sampling distance of a single pixel on the road surface from the infrared polarized light sensor.

[0102] Next, using the rectangular grid as a reference, the registered road surface infrared polarization feature matrix and the road surface visible light texture image are simultaneously divided and data extracted. For each grid cell, all pixels falling into that grid cell are extracted from the infrared polarization feature matrix, and the arithmetic mean of the infrared radiation intensity value and polarization degree value of that grid cell is calculated to generate the infrared polarization feature data of that grid cell; at the same time, the corresponding pixel blocks (i.e., grayscale image blocks) are extracted from the visible light texture image and directly used as the visible light texture data of that grid cell.

[0103] Finally, a structured multispectral feature grid is generated, in which each grid cell contains infrared polarization feature data and visible light texture data extracted and encapsulated from its corresponding physical region.

[0104] S320. Obtain infrared polarization feature data and visible light texture data of grid cells in the multispectral feature grid to generate a fused feature vector, input it into the material classification model to obtain the road surface material type; extract grayscale statistical features from the visible light texture data as the texture roughness index of the grid cells; combine the road surface material type and texture roughness index to determine the potential adhesion coefficient of the grid cells.

[0105] The infrared polarization feature data (including the mean infrared intensity and the mean polarization degree) of the grid cell are extracted, as well as the local binary pattern (LBP) texture spectrum histogram features extracted from the grayscale image patch corresponding to its visible light texture data. The above features are concatenated to form the fusion feature vector of the grid cell, and then input into a pre-trained material classification model.

[0106] The material classification model is a multi-classifier based on support vector machines. Its training data comes from a sufficient number of road surface material samples collected under various typical environmental and lighting conditions and precisely labeled by humans. The model learns the distribution differences of different materials in the fused feature space to establish a non-linear mapping relationship from feature vectors to material categories. During online recognition, the model directly outputs the road surface material type to which the current grid cell belongs.

[0107] Texture analysis is performed on visible light texture grayscale image patches within the same grid cell, and the texture roughness index of that grid cell is calculated using the gray-level co-occurrence matrix (GLCM) algorithm. The GLCM algorithm constructs GLCMs for the visible light texture grayscale image patch in multiple directions and extracts contrast feature values ​​from these GLCMs as output. These contrast feature values ​​directly reflect the degree of difference in grayscale values ​​between pixels in the image, thus quantitatively characterizing the intensity of local changes in the road surface microstructure, i.e., the texture roughness index.

[0108] Subsequently, the road surface material type and texture roughness index value obtained in the above steps are used as joint inputs. A preset material roughness adhesion coefficient mapping relationship database is queried. Through table lookup and interpolation calculation, the potential adhesion coefficient of the grid cell is directly output.

[0109] The material roughness adhesion coefficient mapping database is set up as follows: Under a standardized testing environment, using standard friction testing equipment, the sliding friction coefficients of various typical material samples under different surface texture states are actually measured, thereby establishing and storing a precise correspondence table between material type, texture roughness index value and adhesion coefficient, and storing the standard adhesion coefficient values ​​corresponding to various materials under different roughness.

[0110] Through the above process, the corresponding road surface material type, texture roughness index and potential adhesion coefficient are output for each grid cell in the multispectral feature grid.

[0111] S330. Determine the transition risk level of each grid cell based on the road surface material type and potential adhesion coefficient of each grid cell in the multispectral feature grid. Calculate the basic anti-skid angle compensation value of the grid cell based on its transition risk level and potential adhesion coefficient. Integrate the material type, transition risk level, and basic anti-skid angle compensation value information of all grid cells to generate a gridded transition risk map.

[0112] For each grid cell in the multispectral feature grid, based on the road material type and its potential adhesion coefficient value, a pre-defined risk rating rule base is consulted to determine the transition risk level of the grid cell. The transitional risk level is divided into three levels: low risk, medium risk, and high risk. The low risk level is mapped to the value 1, the medium risk level to the value 2, and the high risk level to the value 3.

[0113] The method for setting the risk rating rule base is as follows: based on the various standard adhesion coefficient values ​​stored in the aforementioned material roughness adhesion coefficient mapping relationship database, and combined with biomechanical research and vehicle industry safety standards, two key thresholds are calibrated: one is the safety adhesion coefficient threshold that ensures the safety of routine vehicle disembarkation. Secondly, the critical adhesion coefficient threshold represents a significant increase in the risk of slipping. The critical adhesion coefficient threshold value is set to 70%. Based on this, a three-level risk assessment rule is defined: when the potential adhesion coefficient μ of the road surface grid unit is ≥ When, it is judged as low risk; when ≤μ< When μ < 0, it is classified as medium risk; when μ < 0, it is classified as medium risk. At that time, it was determined to be high risk.

[0114] After obtaining the transition risk level and potential adhesion coefficient of the mesh element, the basic anti-slip angle compensation value of the element is calculated using an anti-slip angle compensation function. The specific form of the anti-slip angle compensation function is as follows: ; in, Based on the basic anti-slip angle compensation value, For input variables, represents the proportional coefficient related to the transition risk level, and its value is defined according to the following rules: When When it is low risk, The value is 0.8; when When it is medium risk, The value is 1.0; when When it is high risk, The value is 1.5, which allows for greater angle compensation in high-risk areas, providing stronger anti-slip protection. This is a preset reference standard for the safe adhesion coefficient, which is set based on the minimum safe adhesion coefficient required between a person's foot and the pedal surface when getting off the vehicle. The potential adhesion coefficient for the current mesh cell is determined by querying a pre-defined database of material roughness adhesion coefficient mappings. This anti-slip angle compensation function aims to calculate the required angle adjustment to improve anti-slip performance based on the transition risk level and degree of insufficient adhesion.

[0115] Traverse all cells in the multispectral feature grid and record the pavement material type and transition risk level for each cell. and basic anti-slip angle compensation value Finally, this information is integrated to generate a structured dataset with the exact same spatial structure and coordinate correspondence as the multispectral feature grid, namely, a gridded transition risk map. The spatial location of each data entry in this transition risk map is uniquely determined by the position of its corresponding grid cell in the vehicle body coordinate system.

[0116] Safety and anti-slip constraint fusion module: Through multi-constraint programming, the mechanical safety boundary and the gridded transition risk map are fused to generate a preliminary attitude strategy that includes the initial side step extension length and the initial anti-slip angle compensation value.

[0117] In this embodiment, as Figure 4 As shown, the preliminary attitude strategy, which integrates the mechanical safety boundary and the gridded transition risk map through multi-constraint programming, generates a preliminary attitude strategy that includes the initial side step extension length and the initial anti-slip angle compensation value. The specific steps are as follows:

[0118] S410. Set the maximum allowable extension length in the mechanical safety boundary as a rigid constraint, and merge the suggested avoidance area in the mechanical safety boundary with the grid area marked as high risk in the transition risk map into a priority avoidance area.

[0119] The maximum allowable extension length in the mechanical safety boundary is set as a rigid constraint on the extension length of the side step in all subsequent plans, meaning that the actual extension length of the side step must not exceed the maximum allowable extension length.

[0120] The recommended avoidance zone for obstacles is extracted from the mechanical safety boundary. This zone consists of polygonal projections of all obstacles extended with a safety margin. Simultaneously, the gridded transition risk map is traversed, and all grid cells marked as high-risk are identified. The planar areas covered by these cells are aggregated into high-risk road surface areas. Finally, the recommended avoidance zone and the high-risk road surface areas are spatially merged to form a complete, continuous, integrated restricted area, i.e., the priority avoidance zone. This priority avoidance zone defines the spatial range that the sidestep should avoid covering in any extended posture.

[0121] S420. Within the space that satisfies rigid constraints, select all grids that are not covered by priority avoidance areas and whose transition risk level is low or medium from the transition risk map to form a set of candidate support grids.

[0122] Within the space that meets the rigid constraints, the reachable space of the side step is calculated based on the maximum allowable extension length. Then, within this range, the gridded transition risk map is traversed. All grid units that are not covered by the priority avoidance area and have a transition risk level of low or medium risk are selected by spatial location judgment and transition risk level judgment. Finally, these qualified units are aggregated to form a candidate support grid set.

[0123] S430. Optimize the candidate support grid set based on multi-constraint programming. The planning objectives include minimizing the transition risk level of the selected grid and minimizing the absolute value of the basic anti-slip angle compensation value corresponding to the selected grid. By calculating the weighted score of each grid, an optimal support grid is determined. Based on the position coordinates of the optimal support grid in the transition risk map and the recorded basic anti-slip angle compensation value, the corresponding initial side step extension length and initial anti-slip angle compensation value are determined.

[0124] Multi-constraint programming is performed on a set of candidate support grids, where the decision variable is selecting the optimal grid as the final support point. This programming must satisfy the core constraint: the tread extension length calculated based on the selected grid location must not exceed the maximum allowable extension length. The goal of the programming is to minimize the objective function value. This value is determined by the quantized transition risk level of the selected grid. The absolute value of its basic anti-slip angle compensation value The objective function, derived by weighted summation, is: ; in, Let the value of the objective function be minimized. To quantify the transition risk level value, it is the quantified result of the transition risk level recorded by this grid in the transition risk map. It is the absolute value of the basic anti-slip angle compensation value, reflecting the absolute range of anti-slip adjustment. It is a preset weighting coefficient used to balance the relative importance between the risk of slipping and the pedal adjustment range in the objective function. The specific values ​​were obtained through simulation and real-vehicle tests covering typical road surfaces and user scenarios, using this objective function. The empirical value is obtained through multiple rounds of iterative optimization and calibration, with the criterion of achieving the best overall decision-making effect.

[0125] By calculating the objective function value for each candidate grid , choose to The grid with the smallest value is selected as the optimal support grid. Then, based on the position coordinates of this optimal support grid recorded in the transition risk map, the initial side-step extension length required to cover the support point is calculated using the inverse kinematics model of the side-step mechanism. Simultaneously, the basic anti-slip angle compensation value recorded by this optimal support grid is directly used as the initial anti-slip angle compensation value. Finally, the initial side-step extension length and the initial anti-slip angle compensation value together constitute the initial attitude strategy.

[0126] Personalized execution strategy generation module: Based on the collected human getting-out-of-vehicle behavior habit data, through behavior pattern analysis and action inference, it predicts the user's getting-out-of-vehicle action characteristics; and combined with the preliminary posture strategy, through strategy adaptation and constraint verification, it generates side step posture execution instructions.

[0127] In this embodiment, as Figure 5 As shown, based on the collected data on human vehicle disembarkation behavior habits, the system predicts user disembarkation action characteristics through behavior pattern analysis and action deduction. Combined with a preliminary posture strategy, and after strategy adaptation and constraint verification, the system generates side-step posture execution commands, including the following specific steps:

[0128] S510. Determine the user's stability requirement intensity level based on the user's dismount speed dataset, and perform statistical analysis on historical side pedal extension length preference records to determine the user's comfortable length range; integrate the stability requirement intensity level and the comfortable length range to generate structured predicted user dismount action features.

[0129] Based on user disembarkation speed datasets from human vehicle disembarkation behavior data, the stability demand intensity level of users is determined. Specifically, this includes: statistically analyzing the user disembarkation speed dataset and calculating the 70th percentile (P70) and 30th percentile (P30) as thresholds; determining the stability demand intensity level based on the average speed value of the most recent N (N≥3, or the most recent 10 if there are more than 10 valid historical records) valid disembarkation records before the current disembarkation action: if the average speed value is greater than or equal to P70, the stability demand intensity level is determined to be high; if the average speed value is greater than P30 but less than P70, the stability demand intensity level is determined to be medium; if the average speed value is less than or equal to P30, the stability demand intensity level is determined to be low.

[0130] Next, a comprehensive analysis was conducted on multiple length record segments contained in the historical side pedal extension length preference records: First, the reliability of each segment was evaluated, and its reliability weight was determined by the recording frequency of the segment and the reciprocal of the segment variance. The higher the recording frequency and the smaller the variance, the more reliable the segment was considered. Subsequently, based on the reliability weight of each segment and its segment mean, a weighted average was calculated as the overall central estimate of the user's habit length, and the weighted standard deviation of each segment mean relative to this central value was calculated to measure the overall dispersion of the habit.

[0131] Finally, combining the central estimate and weighted standard deviation, and introducing a pre-set scaling factor based on statistical coverage principles and personalized adaptability requirements, the user's comfortable length range is calculated. The specific calculation method for the comfortable length range is as follows: subtract the product of the scaling factor and the weighted standard deviation from the central estimate; the result is used as the lower limit of the range. Add the product of the scaling factor and the weighted standard deviation to the central estimate; the result is used as the upper limit of the range. This scaling factor is set based on statistical principles to ensure that the final range covers most of the user's habitual lengths. Finally, the determined stability requirement intensity level and the calculated comfortable length range are integrated and encapsulated into structured data containing level and upper / lower limit fields to generate predicted user disembarkation action characteristics.

[0132] S520. Extract the comfortable length range from the predicted user's dismounting action features, limit the initial side pedal extension length within the comfortable length range, and obtain the comfort adjustment length; based on the stability requirement intensity level, determine the angle adjustment coefficient through the preset level coefficient mapping relationship, and multiply the initial anti-slip angle compensation value with the angle adjustment coefficient to obtain the stability adaptation angle.

[0133] The comfort length range is extracted from the predicted user dismounting action characteristics. The initial side pedal extension length in the initial posture strategy is compared and adjusted with the comfort length range: if the initial side pedal extension length is less than or equal to the lower limit of the comfort length range, it is adjusted to the lower limit; if it is greater than or equal to the upper limit, it is adjusted to the upper limit; if it is within the range, the initial side pedal extension length remains unchanged. The initial side pedal extension length obtained after this adjustment is the comfort adjustment length.

[0134] Simultaneously, personalized adaptation of the anti-slip angle compensation value is performed. The stability requirement intensity level is extracted from the predicted user disembarkation action characteristics, and the corresponding angle adjustment coefficient is determined according to a preset level coefficient mapping relationship. This mapping relationship is set as follows: for users with a high stability requirement level, the angle adjustment coefficient is 1.3; for users with a medium stability requirement level, the angle adjustment coefficient is 1.0; and for users with a low stability requirement level, the angle adjustment coefficient is 0.7. The core principle is to assign a larger adjustment coefficient to users with higher stability requirements to provide more adequate anti-slip protection. Subsequently, the determined angle adjustment coefficient is multiplied by the initial anti-slip angle compensation value in the initial attitude strategy to obtain the final stability adaptation angle.

[0135] S530. Perform constraint verification on the comfort adjustment length and stability adaptation angle, constrain the comfort adjustment length to not exceed the maximum allowable extension length in the mechanical safety boundary, and constrain the stability adaptation angle within the angle adjustment range of the side pedal mechanism; combine the final comfort adjustment length and the final stability adaptation angle determined after this constraint verification to generate the side pedal posture execution command that drives the side pedal mechanism.

[0136] A final constraint verification is performed on the comfort adjustment length and stability adaptation angle to ensure that they meet the mechanical safety boundaries and physical adjustment range of the side pedal mechanism. First, in the length dimension, the comfort adjustment length is compared with the maximum permissible extension length in the mechanical safety boundaries: if the comfort adjustment length is greater than or equal to the maximum permissible extension length, the comfort adjustment length is adjusted to the maximum permissible extension length; otherwise, the comfort adjustment length value remains unchanged. The comfort adjustment length after this verification is determined as the final side pedal extension length.

[0137] In terms of angle, the stability adaptation angle is compared with the inherent angle adjustment range of the side pedal mechanism: if the stability adaptation angle is greater than or equal to the upper limit of the side pedal mechanism's angle adjustment range, the stability adaptation angle is corrected to the upper limit value; if the stability adaptation angle is less than or equal to the lower limit of the side pedal mechanism's angle adjustment range, the stability adaptation angle is corrected to the lower limit value; otherwise, the stability adaptation angle remains unchanged. This verified stability adaptation angle is determined as the final anti-slip angle compensation value.

[0138] Finally, the final side pedal extension length that has passed the above verification is combined with the final anti-slip angle compensation value to generate a set of structured side pedal posture execution commands, which are then sent to the drive controller of the side pedal mechanism to drive the machine to complete precise posture adjustment.

[0139] Embodiments of the present invention also provide an electronic device, including a memory, a processor, and a communication bus; the memory and the processor are connected via the communication bus. The memory stores an electric side pedal system for new energy vehicles with road surface pre-sensing and anti-slip functions, which can be loaded by the processor and executed as provided in the above embodiments.

[0140] The memory can be used to store instructions, programs, code, code sets, or instruction sets. The memory may include a program storage area and a data storage area. The program storage area may store instructions for implementing an operating system, instructions for at least one function, and instructions for implementing a data analysis-based visual communication design evaluation provided in the above embodiments. The data storage area may store data related to a new energy vehicle electric side pedal system with road surface pre-sensing and anti-slip functions provided in the above embodiments.

[0141] A processor may include one or more processing cores. The processor executes instructions, programs, code sets, or instruction sets stored in memory, and calls data stored in memory to perform various functions and process data as described in this application. The processor may be at least one of a specific application-specific integrated circuit, a digital signal processor, a digital signal processing device, a programmable logic device, a field-programmable gate array, a central processing unit, a controller, a microcontroller, and a microprocessor. It is understood that, for different devices, the electronic devices used to implement the above-described processor functions may also be other types, and the embodiments of this application do not specifically limit the specific implementation.

[0142] A communication bus may include a pathway for transmitting information between the aforementioned components. The communication bus can be a PCI bus or an EISA bus, etc. Communication buses can be categorized into address buses, data buses, control buses, etc.

[0143] This application provides a computer-readable storage medium storing a computer program that can be loaded by a processor and executed as described in the above embodiments, which is a new energy vehicle electric side pedal system with road surface pre-sensing and anti-slip function.

[0144] In this embodiment, a computer-readable storage medium can be a tangible device that holds and stores instructions used by an instruction execution device. A computer-readable storage medium can be, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof. Specifically, a computer-readable storage medium can be a portable computer disk, a hard disk, a USB flash drive, a random access memory, a read-only memory, an erasable programmable read-only memory, a podium random access memory, a portable compressed disk read-only memory, a digital multifunction disk, a memory stick, a floppy disk, an optical disk, a magnetic disk, a mechanical encoding device, or any combination thereof.

[0145] The terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0146] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing application concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions claimed in this application.

Claims

1. A new energy vehicle electric side step system with road surface pre-sensing and anti-slip function, characterized in that, Includes the following modules: Road surface multimodal data acquisition module: Simultaneously collects road surface geometry data, road surface condition data, and human exit behavior data of the side step area of ​​new energy vehicles; Geometric structure analysis module: Based on road surface geometry data, through spatial modeling and motion simulation analysis, it constructs the mechanical safety boundary of the side step in the real-time road surface environment; Surface condition analysis module: Based on road surface condition data, through surface material identification and risk mapping analysis, a gridded transition risk map is constructed to predict the risk of slipping when getting out of the vehicle; Safety and anti-slip constraint fusion module: Through multi-constraint programming, the mechanical safety boundary and the gridded transition risk map are fused to generate a preliminary attitude strategy that includes the initial side step extension length and the initial anti-slip angle compensation value; The initial attitude strategy for generating the initial side step extension length and initial anti-slip angle compensation value includes: optimizing the candidate support grid set based on multi-constraint programming, with the planning objectives including minimizing the transition risk level value of the selected grid and minimizing the absolute value of the basic anti-slip angle compensation value corresponding to the selected grid; determining an optimal support grid by calculating the weighted score of each grid; and determining the corresponding initial side step extension length and initial anti-slip angle compensation value based on the position coordinates of the optimal support grid in the transition risk map and the recorded basic anti-slip angle compensation value. Personalized execution strategy generation module: Based on the collected data on human behavior habits when getting off the vehicle, the module analyzes behavior patterns and infers actions to predict the characteristics of the user's actions when getting off the vehicle. Combined with the initial attitude strategy, and after strategy adaptation and constraint verification, the side tread attitude execution command is generated.

2. The electric side step system for new energy vehicles with road surface pre-sensing and anti-slip function according to claim 1, characterized in that, The process involves simultaneously collecting road surface geometry data, road surface condition data, and human exit behavior data from the side steps of new energy vehicles, including the following steps: S110. Based on the multi-sensor data acquisition and fusion of the side step area of ​​new energy vehicles, construct road geometric data including a three-dimensional point cloud dataset of the side step area, a sequence of road surface ground clearance measurements, and obstacle contour point clouds; S120. Based on the coordinated acquisition of infrared and visible light in the side step area of ​​new energy vehicles, construct road surface state data including road infrared polarization feature matrix and road visible light texture image; S130. Based on personalized data learning of new energy vehicle users, construct human vehicle disembarkation behavior data that includes user disembarkation speed dataset and historical side step extension length preference records.

3. A new energy vehicle electric side step system with road surface pre-sensing and anti-slip function according to claim 2, characterized in that, Based on road surface geometry data, and through spatial modeling and motion simulation analysis, the mechanical safety boundary of the side step in the real-time road surface environment is constructed, including the following steps: S210. The three-dimensional point cloud dataset of the side step area is fused with the road surface ground clearance measurement sequence to reconstruct and generate a three-dimensional road surface mesh model. Based on the three-dimensional road surface mesh model, the lateral slope angle and longitudinal slope angle of the road surface relative to the vehicle body are calculated. S220. Segment and calculate the obstacle contour point cloud to obtain the obstacle's position coordinates and size parameters; S230. Combining the lateral slope angle, longitudinal slope angle, and the position coordinates and size parameters of the obstacle, the deployment process is simulated through the kinematic model of the side step mechanism. Collision interference detection and static stability analysis are performed to obtain the maximum allowable extension length. Combined with the outer contour of the obstacle, a suggested avoidance area is delineated, which together constitutes the mechanical safety boundary.

4. A new energy vehicle electric side step system with road surface pre-sensing and anti-slip function according to claim 3, characterized in that, Based on road surface condition data, and through surface material identification and risk mapping analysis, a gridded transition risk map is constructed to predict the risk of slipping after getting out of a vehicle. This includes the following steps: S310. Spatial registration and grid division are performed on the road surface infrared polarization feature matrix and the road surface visible light texture image to generate a multispectral feature grid in which each grid cell contains infrared polarization feature data and visible light texture data. S320. Obtain infrared polarization feature data and visible light texture data of grid cells in the multispectral feature grid to generate a fused feature vector, input it into the material classification model to obtain the road surface material type; extract grayscale statistical features from the visible light texture data as the texture roughness index of the grid cells; combine the road surface material type and texture roughness index to determine the potential adhesion coefficient of the grid cells. S330. Determine the transition risk level of each grid cell based on the road surface material type and potential adhesion coefficient of each grid cell in the multispectral feature grid. Calculate the basic anti-skid angle compensation value of the grid cell based on its transition risk level and potential adhesion coefficient. Integrate the material type, transition risk level, and basic anti-skid angle compensation value information of all grid cells to generate a gridded transition risk map.

5. A new energy vehicle electric side step system with road surface pre-sensing and anti-slip function according to claim 4, characterized in that, The mechanical safety boundary and the gridded transition risk map are fused through multi-constraint programming, including the following steps: S410. Set the maximum allowable extension length in the mechanical safety boundary as a rigid constraint, and merge the suggested avoidance area in the mechanical safety boundary with the grid area marked as high risk in the transition risk map into a priority avoidance area. S420. Within the space that satisfies the rigid constraints, select all grids that are not covered by the priority avoidance area and whose transition risk level is low or medium from the transition risk map to form a candidate support grid set.

6. A new energy vehicle electric side step system with road surface pre-sensing and anti-slip function according to claim 5, characterized in that, Based on collected data on human behavior habits when getting out of a vehicle, the system analyzes behavioral patterns and extrapolates actions to predict user behavior characteristics when getting out of the vehicle. Includes the following: Based on the user disembarkation speed dataset, the user's stability requirement intensity level is determined, and statistical analysis is performed on historical side pedal extension length preference records to determine the user's comfortable length range; By integrating stability requirement intensity levels with comfort length ranges, structured predictive user disembarkation action characteristics are generated.

7. A new energy vehicle electric side step system with road surface pre-sensing and anti-slip function according to claim 6, characterized in that, Based on the initial attitude strategy, and after strategy adaptation and constraint verification, the side tread attitude execution instructions are generated, including the following: Extract the comfortable length range from the predicted user disembarkation action features, limit the initial side step extension length within the comfortable length range, and obtain the comfort adjustment length; based on the stability requirement intensity level, determine the angle adjustment coefficient through a preset level coefficient mapping relationship, and multiply the initial anti-slip angle compensation value with the angle adjustment coefficient to obtain the stability adaptation angle; The comfort adjustment length and stability adaptation angle are constrained and verified. The comfort adjustment length is constrained to not exceed the maximum allowable extension length in the mechanical safety boundary, and the stability adaptation angle is constrained to the angle adjustment range of the side pedal mechanism. The final comfort adjustment length and the final stability adaptation angle determined after this constraint verification are combined to generate the side pedal posture execution command that drives the side pedal mechanism.

Images