Vehicle display screen pose optimization method based on reflected light path attribution constraint

Abstract

The present application relates to a kind of based on reflection light path attribution constraint's vehicle display screen pose optimization method, it is related to the technical field of uncomfortable glare engineering evaluation, solve the technical problems that there is obvious deficiency in optimization efficiency, explainability and reusability of the reflection glare engineering process in the prior art for vehicle cockpit, the method includes the following steps: step 1: reflection light path attribution and constraint generation;Step 2: the minimum solution stage of unified glare value in constraint domain.This based on reflection light path attribution constraint's vehicle display screen pose optimization method of the present application, by reflection light path attribution in UGR High-risk pose is identified before evaluation and exclusion zone constraint is generated, reducing the invalid evaluation of a large number of inevitable high-glare poses, improving the effectiveness and engineering efficiency of optimization search.

Description

Technical Field

[0001] This invention relates to the field of engineering evaluation technology for uncomfortable glare, and in particular to a method for optimizing the pose of a vehicle display screen based on attribution constraints of reflected light paths. Background Technology

[0002] The display terminals in a vehicle's cockpit include the central control display screen, instrument panel, and HUD display components. These displays typically require high brightness to ensure readability during the day or in high-light conditions. However, the light emitted can be reflected mirror-like or quasi-mirror-like by the inner surfaces of the windshield, side windows, dashboard, and some highly reflective interior components. The reflected light creates bright areas or virtual bright spots in the driver's field of vision, easily triggering uncomfortable glare. This manifests as: strong local contrast in the field of vision, bright spots interfering with visual perception, driver discomfort, distraction, and reduced efficiency in recognizing traffic targets. During vehicle driving, the driver's gaze frequently switches between road targets and cockpit displays. The interference from reflected glare not only affects the reading of displayed content but may also impair the continuous perception of the driving situation, exhibiting significant human-caused risks.

[0003] In engineering evaluations of uncomfortable glare, the Unified Glare Rating (UGR) is often used as a quantitative indicator. UGR attributes the degree of glare discomfort to the relative relationship between the brightness contribution of a bright source within the observer's field of view and the background brightness, taking into account the positional influence of the glare source within the field of view. Its typical form is as follows: given an observation point and line of sight, identify visible bright sources or bright areas in the field of view, calculate their brightness and solid angle, and introduce a position index, then combine this with the background brightness to form the glare evaluation metric. The advantages of UGR are: 1) Using a single numerical value to characterize the degree of glare discomfort is suitable for comparing and optimizing solutions; 2) The computational input can be interfaced with the optical simulation results (brightness distribution, visible source list, background brightness), and the engineering implementation path is clear; 3) There are mature and standardized definitions for evaluating indoor lighting and visual comfort, which makes it easy to refer to the standards in technical documents and avoid disputes.

[0004] Regarding the issue of reflected glare in vehicle cockpits, the formation of reflected glare can be understood from a geometric optical path perspective: after the light emitted from the display screen is reflected by the inner surface of the windshield, if the reflected light falls into the driver's field of vision, a visible, bright reflected virtual image is formed within the field of vision. When the reflected virtual image is located near the center of the line of sight or occupies a certain solid angle and its brightness is significantly higher than the background, the UGR value increases, manifesting as uncomfortable glare. Since windshields are usually curved, and there are obstructions, frames, and structural components inside the cockpit, even small changes in the display screen's pose can cause a jump in whether the reflection path is valid or not, resulting in significant nonlinear changes in the glare index. For these reasons, relying solely on manual experience or coarse-grained scanning often makes it difficult to obtain a stable low-glare pose scheme within an acceptable computational load.

[0005] To address glare reflected from the vehicle's cockpit, existing engineering approaches typically follow a closed-loop process of "modeling—simulation—evaluation—adjustment," and their implementation can be summarized as follows: 1) Establish an optical model of the cockpit; Establish a cockpit geometric model in the vehicle coordinate system, including at least the geometric positional relationships of the windshield, side windows, instrument panel, and display screen, and set the display screen as a light source to form a single-condition simulation scene. For transparent parts such as glass, set their reflection / transmission properties or equivalent reflection models so that the simulation can generate a reflected light path of "display screen - glass - driver".

[0006] 2) Define the driver's observation point and line of sight, and calculate the brightness / glare index; An observation point is set at the driver's typical eye position, and the direction of the line of sight is defined (usually along the road or in conjunction with the display reading posture). The brightness distribution or visible light source brightness information in the cockpit is calculated. Based on this, high-brightness sources or high-brightness areas in the field of view are identified according to the calculation caliber of UGR (bright spots of reflected virtual images can be regarded as part of high-brightness sources), and the UGR value is calculated by combining the background brightness and position index, which is used as a glare evaluation index.

[0007] 3) Iterative adjustment or scanning search with the display screen pose as the variable; Parameterize the display pose as the independent variable x, and find a solution that meets the requirements using one of the following methods: Manual parameter adjustment: Based on simulation images and experience, engineers adjust the display screen position or pitch / yaw angle, and then recalculate the UGR; Parameter scanning: Perform discrete step size scanning on the pose parameters, calculate the UGR of each candidate pose and select the minimum value; General optimization: Treat UGR(x) as a black-box objective function and try to find a better solution in the pose space through random search or simple iterative strategy.

[0008] 4) Recommendations for low-glare output layout; Finally, a specific display screen pose scheme and corresponding UGR evaluation results are provided for structural design or layout decisions of the vehicle's cockpit.

[0009] While the above approach can achieve the basic goal of "calculating UGR and adjusting pose accordingly", it treats the glare problem mainly as an "objective function optimization problem", which has obvious shortcomings in terms of optimization efficiency, interpretability and reusability. Summary of the Invention

[0010] This invention aims to address the significant shortcomings in the existing engineering processes for vehicle cockpit reflection glare in terms of optimization efficiency, interpretability, and reusability, and provides a vehicle display screen pose optimization method based on reflection light path attribution constraints.

[0011] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: A method for optimizing the pose of a vehicle display screen based on reflection optical path attribution constraints includes the following steps: Step 1: Attribution and constraint generation of reflected light paths; Based on geometric optical path determination, the system identifies the set of effective reflection paths that cause glare within the driver's field of vision and generates safety domain constraints accordingly. Step 2: Solve for minimizing the uniform glare value within the constraint domain; Within the safe domain generated in step 1, a uniform glare value is calculated for the candidate poses, and the optimal pose is output by using the covariance matrix adaptive evolution algorithm for constraint optimization.

[0012] In the above technical solution, before step 1, there is a step: input / output and variable definition; defining input data, output data and display pose vector; in: Input data includes: vehicle cockpit geometry model, key reflective interface set, driver observation parameters, and display assembly feasible domain constraints; The output data includes: optimal display pose parameters, corresponding to Values, and descriptions of pose safety domains and restricted areas; Display Pose Vector Defined as: in, Indicates the translation parameter. These represent displacements along the x, y, and z directions, respectively. Indicates transpose; Indicates attitude parameters, These represent the rotation angles around the x-axis, y-axis, and z-axis, respectively.

[0013] In the above technical solution, step 1 specifically includes the following steps: Step 11: View sampling and ray generation; From the driver's perspective Construct a line-of-sight cone starting from the point of view, and sample the line-of-sight directions within the cone to generate a set of line-of-sight ray directions. The sampling direction must meet the following requirements: in, This represents the unit vector in the direction of the driver's line of sight. Indicates half-angle of the view, Indicates the first The direction unit vector of the sampled field-of-view ray, Indicates the ray number, = 1, 2, …, N; Step 12: Find the intersection with the key reflection interface; Geometric intersection is performed on each ray, and the ray is represented as: in, Indicates the eye point. Represents the corresponding parameters on the ray. Spatial location point, Represents ray parameters, used to indicate the distance scale along the ray direction. Represents the unit vector of the ray direction; Will Find the nearest intersection point with the set of key reflective interfaces S. and the normal direction of that point ; Step 13: Calculation of reflection direction and determination of effective path; At the most recent intersection At that point, construct a line pointing from the intersection to the eye point. unit direction vector Calculate the reflection direction according to the laws of specular reflection: in, This represents the unit vector indicating the direction of reflection. Represents the unit vector of the incident direction. This represents the unit normal vector at the intersection point; Then along Emit a reflected ray and determine whether the reflected ray intersects with the geometry of the display screen; Step 14: Risk path aggregation and risk degree function construction; Define the risk function for: in, This represents the angle between the incident direction and the line-of-sight direction corresponding to the risk path. Indicates the weight decay scale. Indicates a valid reflection path. Represents the pose vector on the display screen The set of effective reflection paths below, Indicates the effective reflection path The weight function value, Represents the display screen pose vector; Step 15: Generation of security domain and restricted area constraints; Generate the display pose safety domain based on the risk factor function. Set risk thresholds ,definition: in, This indicates the feasible domain for display screen assembly.

[0014] In the above technical solution, step 2 specifically includes the following steps: Step 21 Indicator definition; Calculated according to the unified glare value definition: in, Indicates background brightness; Indicates the first Brightness of individual glare sources or high-brightness areas; Represents solid angle; Indicates location index; Indicates the number of glare sources. Indicates the glare source number. Indicates a uniform glare value; Step 22: Objective function and constraint penalty function; In the CMA-ES (Covariance Matrix Adaptive Evolution) algorithm search process, the risk constraint is incorporated into the objective function in the form of a penalty function: in, Represents the risk level function. Indicates the risk threshold. Indicates the penalty factor. This indicates the feasible domain for display screen assembly. This represents the display screen pose vector. Represents the display pose vector The uniform glare value, Represent the objective function; Step 23: Solve using CMA-ES; CMA-ES uses the display pose vector x as the optimization variable and the objective function as the objective function. Perform an iterative search for the fitness function.

[0015] In the above technical solution, the CMA-ES solution process in step 23 specifically includes: The initial display pose is determined as the search center, and the initial sampling scale and sampling distribution parameters are set so that the covariance matrix adaptive evolution algorithm CMA-ES can generate candidate poses within the feasible region of display assembly. Entering the iteration process: In each generation, several candidate poses are generated based on the current sampling distribution, forming a candidate pose set; For each candidate pose in the candidate pose set, risk degree calculation is performed first for rapid feasibility determination. The reflection path attribution result from step 1 is then used to quantitatively evaluate the risk of the reflection path corresponding to the candidate pose: when the risk degree exceeds a preset risk threshold, the candidate pose is determined to not meet the safety domain constraint, and a penalty value is assigned to its fitness or it is directly eliminated; when the risk degree meets the risk threshold requirement, a uniform glare value is applied. Calculate and Together with the risk penalty term, they constitute the objective function of the candidate pose. ; After completing the fitness evaluation of the candidate pose set, the candidate poses are sorted according to the objective function value, and several candidate poses with better fitness are selected as winning samples. The sampling distribution is then updated based on the winning samples: the center parameter of the next generation sampling distribution is adjusted to the weighted clustering region of the winning samples to guide the search to converge toward the pose region with a smaller objective function. Based on the statistical distribution characteristics of the winning samples in each pose parameter dimension and their combination direction, the shape and scale of the sampling distribution are adaptively updated. When the stopping criterion is met, the iteration terminates and the optimal pose is output.

[0016] In the above technical solution, the stopping criterion is: The number of iterations reaches a preset upper limit, the improvement of the optimal value of the objective function is less than a preset threshold for several consecutive generations, the sampling scale converges to a preset lower limit, or The risk level is reduced to the preset target threshold while simultaneously meeting the safety domain constraints.

[0017] The present invention has the following beneficial effects: The vehicle display pose optimization method based on reflected light path attribution constraints of the present invention optimizes the vehicle display pose by attributing reflected light paths. Before evaluation, high-risk poses are identified and restricted areas are generated, reducing a large number of invalid evaluations of poses that inevitably have high glare, and improving the effectiveness and engineering efficiency of optimization search.

[0018] The vehicle display pose optimization method based on reflected light path attribution constraints of the present invention explicitly transforms the reflected glare formation mechanism into structured information of "effective reflection path set - risk degree - safety domain", making the layout decision interpretable and allowing restricted area constraints to be precipitated into reusable rules.

[0019] The vehicle display pose optimization method based on reflected optical path attribution constraints of the present invention optimizes the pose of the vehicle display screen within the safety domain. Optimize the pose of the target to make the search process more stable and controllable, reduce the reliance on manual experience for parameter tuning, and be suitable for the iterative static design phase of vehicle display layout. Attached Figure Description

[0020] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0021] Figure 1 This is a schematic diagram of the overall process of the vehicle display pose optimization method based on reflected light path attribution constraints of the present invention.

[0022] Figure 2 This is a schematic diagram of the reflected light path attribution in the vehicle display pose optimization method based on reflected light path attribution constraints of the present invention. Detailed Implementation

[0023] The vehicle display pose optimization method based on reflected optical path attribution constraints of the present invention specifically includes the following inventive concept: 1) In calculation Previously, high-risk pose areas for reflected glare were identified to reduce invalid assessments; By constructing a set of the driver's eye-point field of view and key reflective interfaces, the system performs reflection light path attribution, determines the conditions for the "reflected virtual image entering the field of view," and automatically generates no-placement zones / no-pose zones or safe domain constraints for the display screen's pose, enabling subsequent... The computation and optimization search are focused on effective candidate regions, thereby reducing computational costs and improving iteration efficiency.

[0024] 2) Transform the reflection glare problem from a black-box optimization problem into an interpretable optical path constraint problem, thereby improving convergence stability; By making the causes of glare explicit through path-level attribution, a structured description of "key reflection interface - reflection point - reflection direction - line of sight relationship - pose constraint" is established, which gives pose adjustment a clear direction, reduces blind trial and error caused by curved surface reflection and occlusion boundary, and improves the stability and controllability of the optimization process.

[0025] 3) Generate reusable constraint outputs to enable rapid application across vehicles or modification iterations; By solidifying the laws of reflection risk into exportable restricted / safe domain constraint files or constraint configurations, subsequent vehicle platforms or modification schemes can reuse existing constraint logic under the same caliber, reducing repeated trial and error and accumulating design knowledge.

[0026] 4) Output verifiable chain of attribution evidence to support engineering design review and verification; In addition to outputting the optimal pose, it also outputs key reflection interfaces, typical effective reflection paths, or risk contribution statistics, enabling engineers to verify the sources of glare risk and the basis for pose optimization, forming a reviewable and traceable chain of technical evidence.

[0027] The present invention will now be described in detail with reference to the accompanying drawings.

[0028] The vehicle display pose optimization method based on reflected light path attribution constraints of the present invention includes the following steps: Step 1: Attribution and constraint generation of reflected light paths; Based on geometric optical path determination, without relying on a large number of Based on the assessment, the set of effective reflection paths that cause glare within the driver's field of vision is identified, and a safe domain (positional no-go zone) constraint is generated accordingly; specifically, the following steps are included: Step 11: View sampling and ray generation; From the driver's perspective Construct a line-of-sight cone starting from the point of view, and sample the line-of-sight directions within the cone to generate a set of line-of-sight ray directions. The sampling direction must meet the following requirements: in, This represents the unit vector in the direction of the driver's line of sight. Indicates half-angle of the view, Indicates the first The direction unit vector of the sampled field-of-view ray, Indicates the ray number, = 1, 2, …, N; Step 12: Find the intersection with the key reflection interface; Geometric intersection is performed on each ray, and the ray is represented as: in, Indicates the eye point. Represents the corresponding parameters on the ray. Spatial location point, Represents ray parameters, used to indicate the distance scale along the ray direction. Represents the unit vector of the ray direction; Will Find the nearest intersection point with the set of key reflective interfaces S. and the normal direction of that point ; Step 13: Calculation of reflection direction and determination of effective path; At the most recent intersection At that point, construct a line pointing from the intersection to the eye point. unit direction vector Calculate the reflection direction according to the laws of specular reflection: in, This represents the unit vector indicating the direction of reflection. Represents the unit vector of the incident direction. This represents the unit normal vector at the intersection point; Then along Emit a reflected ray and determine whether the reflected ray intersects with the geometry of the display screen; Step 14: Risk path aggregation and risk degree function construction; Define the risk function for: in, This represents the angle between the incident direction and the line-of-sight direction corresponding to the risk path. Indicates the weight decay scale. Indicates a valid reflection path. Represents the pose vector on the display screen The set of effective reflection paths below, Indicates the effective reflection path The weight function value, Represents the display screen pose vector; Step 15: Generation of security domain and restricted area constraints; Generate the display pose safety domain based on the risk factor function. Set risk thresholds ,definition: in, This indicates the feasible domain for display screen assembly.

[0029] Step 2: Uniform glare value within the constraint domain Minimize the solution; Within the safety domain generated in step 1, calculate a uniform glare value for the candidate pose. The optimal pose is then obtained by using the Covariance Matrix Adaptive Evolutionary Algorithm (CMA-ES) for constrained optimization. In CMA-ES, CMA stands for Covariance Matrix Adaptation, while ES stands for Evolution Strategies, a gradient-free stochastic optimization algorithm. CMA-ES is a stochastic or randomized method for optimizing real-parameter (continuous domain) nonlinear, non-convex functions; it specifically includes the following steps: Step 21 Indicator definition; Calculated according to the unified glare value definition: in, Indicates background brightness; Indicates the first Brightness of individual glare sources or high-brightness areas; Represents solid angle; Indicates location index; Indicates the number of glare sources. Indicates the glare source number. Indicates a uniform glare value; Step 22: Objective function and constraint penalty function; In the CMA-ES (Covariance Matrix Adaptive Evolution) algorithm search process, the risk constraint is incorporated into the objective function in the form of a penalty function: in, Represents the risk level function. Indicates the risk threshold. Indicates the penalty factor. This indicates the feasible domain for display screen assembly. This represents the display screen pose vector. Represents the display pose vector The uniform glare value, Represent the objective function; Step 23: Solve using CMA-ES; CMA-ES uses the display pose vector x as the optimization variable and the objective function as the objective function. Perform an iterative search for the fitness function.

[0030] Step 1, based on preset geometric and optical constraints, achieves rapid convergence of the feasible design domain by identifying and eliminating deterministic high-risk poses; Step 2 then performs refined numerical optimization within the selected feasible domain to ensure uniform glare values. To achieve the theoretically or engineering-permissible global optimum. The two steps are logically progressive, together forming a systematic design closed loop from risk pre-elimination to performance optimization.

[0031] Specifically, such as Figure 1 and 2 As shown, the vehicle display pose optimization method based on reflected light path attribution constraints of the present invention includes the following steps and contents: 1. Input, output and variable definition.

[0032] 1.1 Input data (corresponding) Figure 1 The input parameters include: 1) Vehicle cockpit geometry model: including the inner surface of the windshield, the inner surface of the side window glass, the instrument panel structure and display screen geometry, etc., all expressed in the vehicle coordinate system.

[0033] 2) Set of critical reflective interfaces: This set includes at least the inner surface of the windshield, denoted as the set of critical reflective interfaces S; optionally, it includes the inner surfaces of the side window glass and the transparent dashboard cover, etc. The interfaces can be curved surfaces or triangular meshes, and local normals can be calculated.

[0034] 3) Driver observation parameters: driver's eye point Driver's line of sight unit vector (i.e., direction of gaze), half-angle of visual field .

[0035] 4) Display assembly feasible domain constraints: the installation space boundary of the display in the vehicle's cockpit, the allowable pose range, and the interference / obstruction constraints with structural components, etc.

[0036] 1.2 Output Data (corresponding) Figure 1 The output parameters include: 1) Optimal display screen pose parameters: position and attitude parameters .

[0037] 2) Corresponding value: .

[0038] 3) Description of pose safety domains and restricted areas: Output in the form of constraint configuration, discrete annotation table or restricted area rules, so as to facilitate subsequent deployment and reuse.

[0039] 1.3 Design variables; Display Pose Vector Defined as: (1) in, This represents the translation parameters of the display screen in the vehicle coordinate system. These represent displacements along the x, y, and z directions, respectively. Indicates transpose; The attitude parameters are represented using Euler angles. These represent the rotation angles around the x-axis, y-axis, and z-axis, respectively, used to describe the spatial attitude of the display screen's local coordinate system relative to the vehicle's coordinate system.

[0040] 2. Stage A: Attribution and Constraint Generation of Reflected Light Path - Corresponding to Step 1.

[0041] The core of Phase A is to transform the "reflection glare risk" from a black-box problem of indicators into a decisionable geometric optical path problem, and to generate safety domain constraints that can be used for optimization.

[0042] 2.1 View sampling and ray generation - corresponding to step 11; A line-of-sight cone is constructed starting from the driver's eye point E. Within this cone, the line-of-sight direction is sampled to generate a set of ray directions for the field of view. The sampling direction must meet the following requirements: (2) in, Indicates the first The direction unit vector of the sampled field-of-view ray, Indicates the ray number, = 1, 2,…, N, This represents the unit vector in the direction of the driver's line of sight. This represents the half-angle of the field of view, used to define the maximum angle between the sampling direction and the line-of-sight direction. This step ensures that attribution only considers potential reflection paths entering the driver's field of view, reducing the number of invalid paths.

[0043] 2.2 Intersection with the key reflection interface - corresponding to step 12; Geometric intersection is performed on each ray, and the ray is represented as: (3) in, Indicates the eye point. Represents the corresponding parameters on the ray. Spatial location point, Represents ray parameters, used to indicate the distance scale along the ray direction. Represents the unit vector of the ray direction; Will Find the nearest intersection point with the set of key reflective interfaces S. and the normal direction of that point If there is no intersection point or the intersection point is not within the valid interface range, the ray will not proceed to the subsequent determination.

[0044] 2.3 Calculation of Reflection Direction and Determination of Effective Path - Corresponding to Step 13; At the most recent intersection At that point, construct a line pointing from the intersection to the eye point. unit direction vector Calculate the reflection direction according to the laws of specular reflection: (4) in, This represents the unit vector indicating the direction of reflection. Represents the unit vector of the incident direction. This represents the unit normal vector at the intersection point; Then along Emit a reflected ray and determine if it intersects with the geometric boundaries of the display screen. The display screen can be described by both its planar equation and its boundary boundaries: if the planar equation of the display screen is... Then the parameters of the intersection point between the reflected ray and the plane are: (5) in, This indicates the parameters of the intersection point between the reflected ray and the display screen plane; This represents the unit normal vector of the display screen plane. Represents the spatial position vector of a point on the display screen plane; when >0, and the intersection point If the light falls within the boundary of the display screen, then an "effective reflected glare path" is considered to exist. This path means that, in the current pose, the light emitted by the display screen may enter the driver's field of vision after being reflected by the key reflective interface and form a reflected virtual image or a bright area.

[0045] Based on the above determination, the set of valid reflection paths is obtained. The set of elements can record information such as: the reflection interface number, the intersection point position, the reflection direction, the position of the hit display screen, and the angle between the path and the line of sight.

[0046] 2.4 Risk path aggregation and risk degree function construction - corresponding to step 14; To transform the set of effective reflection paths into a quantitative description that can be used for constraint generation, a risk degree function is introduced. The level of risk is reflected in the positioning pose (i.e., the display screen pose vector). The intensity of the risk of downward reflected glare can be defined using the form of "effective path accumulation weighted by line-of-sight proximity": (6) in, This represents the angle between the incident direction and the line-of-sight direction corresponding to the risk path. Indicates the weight decay scale. Indicates a valid reflection path. Represents the pose vector on the display screen The set of effective reflection paths below, Indicates the effective reflection path The weight function value, This represents the display screen pose vector. This definition reflects that reflection paths near the driver's line of sight are more sensitive to uncomfortable glare, thus highlighting "line of sight risk" during the constraint generation stage.

[0047] 2.5 Generation of security domain and restricted area constraints - corresponding to step 15; Generate the display pose safety domain based on the risk factor function. Set risk thresholds. ,definition: (7) in, This represents the feasible region for display assembly (defined by installation space boundaries, orientation range, interference constraints, etc.). The pose space is divided into safe and risk areas: when the pose causes a large number of effective reflection paths to enter the field of view or the path is close to the center of the line of sight, the risk level increases and is constrained and eliminated.

[0048] Engineering representations of security domain constraints may include: Constraint configuration file: contains , Parameters and interface set definitions are used for reuse; Discrete labeling table: Labels the safety / risk of sampled poses for rapid feasibility determination in the optimization process; Restricted Area Rules: These rules express the boundaries of restricted posture areas and restricted arrangement areas using geometric conditions or combinations of inequalities, and are used for rule-based review during the design phase.

[0049] The safety domain constraints output in stage A above are for stage B. Minimization provides the basis for the feasible region.

[0050] 3. Stage B: Within the constraint domain UGR Minimize the solution -- corresponding to step 2.

[0051] Phase B is conducted within the security domain. Evaluating and solving for the optimal pose includes: candidate pose generation, constraint selection, The calculation, iterative update, and convergence output are performed. This invention consistently uses CMA-ES for pose search, and its solution objective and constraint handling are as follows.

[0052] 3.1 UGR Indicator definition - corresponding step 21; Calculated according to the unified glare value definition: (8) in, Indicates background brightness; Indicates the first Brightness of individual glare sources or high-brightness areas; Represents solid angle; Indicates location index; Indicates the number of glare sources. Indicates the glare source number. This indicates a uniform glare value; for vehicle-reflected glare issues, the glare source can be composed of bright spots or high-brightness areas of reflected virtual images formed by the display screen through key reflective interfaces such as the windshield, and the background brightness is determined by the ambient brightness of the cockpit.

[0053] 3.2 Objective function and constraint penalty function - corresponding to step 22; During the CMA-ES search process, to ensure that candidate poses satisfy the safety domain constraints, the risk constraint is incorporated into the objective function in the form of a penalty function: (9) in, This represents the risk function defined in stage A. Indicates the risk threshold. Indicates the penalty factor. This indicates the feasible domain for display screen assembly. This represents the display screen pose vector. Represents the display pose vector The uniform glare value, Let represent the objective function. This objective function ensures that candidate poses are significantly penalized once they enter the risk region, thereby guiding the search to focus on... The effective area within.

[0054] 3.3 CMA-ES Solution - Corresponding Step 23; CMA-ES uses display pose vectors To optimize variables, with the objective function An iterative search is performed for the fitness function. First, the initial display pose is determined as the search center, and the initial sampling scale and sampling distribution parameters are set so that the Covariance Matrix Adaptive Evolution Algorithm (CMA-ES) can generate candidate poses within the assembly feasible region. Then, the iterative process begins: in each generation, several candidate poses are generated based on the current sampling distribution, forming a candidate pose set. For each candidate pose in the set, risk calculation is prioritized for rapid feasibility determination, i.e., the reflection path attribution result from stage A is called to quantify the reflection path risk corresponding to the candidate pose. When the risk exceeds a preset risk threshold, the candidate pose is determined to not meet the safety region constraints, and a penalty value is assigned to its fitness or it is directly removed, thus avoiding high-cost glare calculations for significantly high-risk poses. When the risk meets the risk threshold requirement, a unified glare value is further calculated. Calculate and Together with the risk penalty term, they constitute the objective function value of the candidate pose. .

[0055] After evaluating the fitness of the candidate pose set, the candidate poses are sorted according to the objective function value. Several candidate poses with better fitness are selected as winning samples, and the sampling distribution is updated based on these winning samples: the center parameter of the next-generation sampling distribution is adjusted to the weighted clustering region of the winning samples to guide the search towards pose regions with smaller objective functions; simultaneously, based on the statistical distribution characteristics of the winning samples in each candidate pose parameter dimension and its combination direction, the shape and scale of the sampling distribution are adaptively updated, so that subsequent candidate pose generation increases the search intensity in high-yield directions and decreases the search intensity in low-yield directions, thereby improving search efficiency and suppressing ineffective exploration. This process of "candidate generation—risk screening—" is implemented. The cyclical process of "calculation - fitness ranking - sampling distribution update" enables continuous optimization of the display pose.

[0056] The iteration terminates and the optimal pose is output when the stopping criterion is met. The stopping criteria include, but are not limited to: the number of iterations reaching a preset upper limit; the improvement in the optimal value of the objective function being less than a preset threshold for several consecutive iterations; the sampling scale converging to a preset lower limit; or The risk level is reduced to a preset target threshold while simultaneously satisfying the safety domain constraint. Optimal pose. It can be defined as the candidate pose with the minimum objective function value during the iteration process, and the corresponding optimal pose can be output synchronously. The values, risk levels, and safety domain constraints are used for subsequent engineering layout review and scheme confirmation.

[0057] In other specific embodiments of the present invention: 1) The key reflective interface can be represented by a surface patch or a triangular mesh, and the normal can be obtained analytically or numerically.

[0058] 2) The risk level function can take the form of path counting, weighted path counting, or incorporating line-of-sight center weighting, etc., with a risk threshold. It can be set to zero or non-zero.

[0059] 3) Security domain (i.e., the feasible domain for display screen assembly) It can output as a constraint configuration file, a discrete annotation table, or an explicit restricted area rule.

[0060] 4) The driver's eye point can be a single eye point, or it can be expanded to the left and right eyes or a small range of head tolerance sets.

[0061] 5) Display screen attitude parameters can be equivalently represented by Euler angles, rotation matrices, or quaternions, and still belong to pose variables.

[0062] In summary, the vehicle display pose optimization method based on reflected light path attribution constraints of the present invention optimizes the vehicle display pose by attributing reflected light paths. Before evaluation, high-risk poses are identified and restricted areas are generated, reducing a large number of invalid evaluations for poses that inevitably produce high glare, thus improving the effectiveness and engineering efficiency of optimization searches. The mechanism of reflected glare formation is explicitly represented as structured information of "effective reflection path set—risk level—safety domain," making deployment decisions interpretable and allowing restricted area constraints to be precipitated as reusable rules. Within the safety domain... Optimize the pose of the target to make the search process more stable and controllable, reduce the reliance on manual experience for parameter tuning, and be suitable for the iterative static design phase of vehicle display layout.

[0063] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for optimizing the pose of a vehicle display screen based on reflection optical path attribution constraints, characterized in that, Includes the following steps: Step 1: Attribution and constraint generation of reflected light paths; Based on geometric optical path determination, the system identifies the set of effective reflection paths that cause glare within the driver's field of vision and generates safety domain constraints accordingly. Step 2: Solve for minimizing the uniform glare value within the constraint domain; Within the safe domain generated in step 1, a uniform glare value is calculated for the candidate poses, and the optimal pose is output by using the covariance matrix adaptive evolution algorithm for constraint optimization.

2. The vehicle display pose optimization method based on reflected optical path attribution constraints according to claim 1, characterized in that, Before step 1, there are steps for: defining input, output and variables; defining input data, output data and display pose vector; in: Input data includes: vehicle cockpit geometry model, key reflective interface set, driver observation parameters, and display assembly feasible domain constraints; The output data includes: optimal display pose parameters, corresponding to Values, and descriptions of pose safety domains and restricted areas; Display Pose Vector Defined as: in, Indicates the translation parameter. These represent displacements along the x, y, and z directions, respectively. Indicates transpose; Indicates attitude parameters, These represent the rotation angles around the x-axis, y-axis, and z-axis, respectively.

3. The vehicle display pose optimization method based on reflected optical path attribution constraints according to claim 1, characterized in that, Step 1 specifically includes the following steps: Step 11: View sampling and ray generation; From the driver's perspective Construct a line-of-sight cone starting from the point of view, and sample the line-of-sight directions within the cone to generate a set of line-of-sight ray directions. The sampling direction must meet the following requirements: in, This represents the unit vector in the direction of the driver's line of sight. Indicates half-angle of the view, Indicates the first The direction unit vector of the sampled field-of-view ray, Indicates the ray number, = 1, 2, …, N; Step 12: Find the intersection with the key reflective interface; Geometric intersection is performed on each ray, and the ray is represented as: in, Indicates the eye point. Represents the corresponding parameters on the ray. Spatial location point, Represents ray parameters, used to indicate the distance scale along the ray direction. Represents the unit vector of the ray direction; Will Find the nearest intersection point by finding the intersection with the set of key reflective interfaces S. and the normal at that point ; Step 13: Calculation of reflection direction and determination of effective path; At the most recent intersection At that point, construct a path from the intersection point to the eye point. unit direction vector Calculate the reflection direction according to the laws of specular reflection: in, This represents the unit vector indicating the direction of reflection. Represents the unit vector of the incident direction. This represents the unit normal vector at the intersection point; Then along Emit a reflected ray and determine whether the reflected ray intersects with the geometry of the display screen; Step 14: Risk path aggregation and risk degree function construction; Define the risk function for: in, This represents the angle between the incident direction and the line-of-sight direction corresponding to the risk path. Indicates the weight decay scale. Indicates a valid reflection path. Represents the pose vector on the display screen The set of effective reflection paths below, Indicates the effective reflection path The weight function value, Represents the display screen pose vector; Step 15: Generation of security domain and restricted area constraints; Generate the display pose safety domain based on the risk factor function. Set risk thresholds ,definition: in, This indicates the feasible domain for display screen assembly.

4. The vehicle display pose optimization method based on reflected optical path attribution constraints according to claim 1, characterized in that, Step 2 specifically includes the following steps: Step 21 Indicator definition; Calculated according to the unified glare value definition: in, Indicates background brightness; Indicates the first Brightness of individual glare sources or high-brightness areas; Represents solid angle; Indicates location index; Indicates the number of glare sources. Indicates the glare source number. Indicates a uniform glare value; Step 22: Objective function and constraint penalty function; In the CMA-ES (Covariance Matrix Adaptive Evolution) algorithm search process, the risk constraint is incorporated into the objective function in the form of a penalty function: in, Represents the risk level function. Indicates the risk threshold. Indicates the penalty factor. This indicates the feasible domain for display screen assembly. This represents the display screen pose vector. Represents the display pose vector The uniform glare value, Represent the objective function; Step 23: Solve using CMA-ES; CMA-ES uses the display pose vector x as the optimization variable and the objective function as the objective function. Perform an iterative search for the fitness function.

5. The vehicle display pose optimization method based on reflected optical path attribution constraints according to claim 4, characterized in that, Step 23 of the CMA-ES solution process specifically includes: The initial display pose is determined as the search center, and the initial sampling scale and sampling distribution parameters are set so that the covariance matrix adaptive evolution algorithm CMA-ES can generate candidate poses within the feasible region of display assembly. Entering the iteration process: In each generation, several candidate poses are generated based on the current sampling distribution, forming a candidate pose set; For each candidate pose in the candidate pose set, risk degree calculation is performed first for rapid feasibility determination. The reflection path attribution result from step 1 is then used to quantitatively evaluate the risk of the reflection path corresponding to the candidate pose: when the risk degree exceeds a preset risk threshold, the candidate pose is determined to not meet the safety domain constraint, and a penalty value is assigned to its fitness or it is directly eliminated; when the risk degree meets the risk threshold requirement, a uniform glare value is applied. Calculate and Together with the risk penalty term, they constitute the objective function of the candidate pose. ; After completing the fitness evaluation of the candidate pose set, the candidate poses are sorted according to the objective function value, and several candidate poses with better fitness are selected as winning samples. The sampling distribution is then updated based on the winning samples: the center parameter of the next generation sampling distribution is adjusted to the weighted clustering region of the winning samples to guide the search to converge toward the pose region with a smaller objective function. Based on the statistical distribution characteristics of the winning samples in each pose parameter dimension and their combination direction, the shape and scale of the sampling distribution are adaptively updated. When the stopping criterion is met, the iteration terminates and the optimal pose is output.

6. The vehicle display pose optimization method based on reflected optical path attribution constraints according to claim 5, characterized in that, The stopping criteria are as follows: The number of iterations reaches a preset upper limit, the improvement of the optimal value of the objective function is less than a preset threshold for several consecutive generations, the sampling scale converges to a preset lower limit, or The risk level is reduced to the preset target threshold while simultaneously meeting the safety domain constraints.

Images