CN103198182B - The pedestrian guide sign design method of view-based access control model perception simulation technology - Google Patents

Abstract

The invention relates to the technical field of traffic signs and signs. To provide theoretically specific basis for the setting of traffic signs and signs, the technical solution adopted by the present invention is a design method for pedestrian guide signs based on visual perception simulation simulation technology, including the following steps of data collection, model design, simulation platform construction and There are four parts: parameter calibration and method application; the model design part is mainly to describe the pedestrian walking environment statically and dynamically, model the pedestrian as an intelligent body, and combine and serialize the characteristics, external environment and behavior of pedestrians; the construction of the simulation platform and parameters The calibration part is mainly to use the standard and feasible three-dimensional virtual simulation experiment method of pedestrians in the practical application of computerized model design, select typical scenes for experiments, compare and correct them with the visual recognition data of pedestrian walking scenes in the actual hub, and give recommended values under different conditions. The range of experimental parameters. The invention is mainly applied to the design of traffic signs and signs.

Description

Design method of pedestrian guide signs based on visual perception simulation technology

technical field

The invention relates to the technical field of traffic signs and signs, in particular to a design method for pedestrian guide signs based on visual perception simulation technology.

Background technique

The design of pedestrian guidance signs needs to be designed to continuously solve spatial problems, including perception and cognition of the environment, using environmental information to make wayfinding decisions and action plans, and then putting decisions into action at appropriate locations. The research on the perceived behavior of pedestrian traffic signs needs to involve multidisciplinary fields such as environmental psychology, behavior, visual perception, ergonomics, etc. Currently, the main research includes the research on the interaction between traffic signs and pedestrian wayfinding behavior.

Abroad, Scholar Funahashi Kunio of Osaka University has done a series of research on the sign system of the surrounding environment of Umeda subway station in Osaka. Through the on-site investigation of the use of sign information, we can understand the utilization efficiency of the sign system, and propose that people in the process of finding a way In addition to inquiries, the most important thing is to rely on the identification system, which has also been confirmed by Canadian scholar S.C.Wimsinghe in collaboration with several Hong Kong scholars in the study of "Hong Kong International Airport Hub Wayfinding".

At present, there are still few domestic researches on the design methods of pedestrian guidance signs. Professor Xu Muqing from the School of Architecture and Urban Planning of Tongji University is more in-depth. Disaster prevention and evacuation design strategy research" two aspects, with transportation hubs, underground public spaces and building complexes as the research objects, according to the wayfinding experiment, using CCD synchronous photography, tracking recording, wayfinding path recording, questionnaire filling and other investigations The method is to study the influence of directional signs on travel wayfinding behavior, so as to make a directional evaluation of the sign system in the urban traffic hub commercial space, and determine the shortcomings and improvement methods of the existing urban transport hub commercial space, sign system design. At the same time, in the commercial space of urban transportation hub, the degree of influence of different types and positions of signs on people's wayfinding is sorted, and it provides reference and guidance for the design of sign types and their key positions, aiming to establish an efficient urban public space. The advanced wayfinding system enhances the orientation and recognizability of the space.

In addition, other researchers conducted on-site surveys to investigate the wayfinding behavior of a certain number of subjects under the guidance of clear signs and those without clear signs, trying to understand the characteristics of spatial cognition, the role of differences in spatial features, sense of direction and number of turns In order to study the design method of pedestrian guidance signs in the urban underground public space, and put forward design suggestions that are conducive to space orientation.

Contents of the invention

The present invention aims to overcome the deficiencies of the prior art, provide a design method for pedestrian guide signs, and provide a theoretical basis for the design of traffic signs and signs. In order to achieve the above-mentioned purpose, the technical solution adopted by the present invention is to The pedestrian guide sign design method of simulation technology includes the following four parts: data collection, model design, simulation platform construction and parameter calibration, and method application; the data collection part is mainly through questionnaire survey, 3D virtual reality scene construction and simulation experiments and pedestrian Quantification of perceptual factors; the model design part mainly describes the pedestrian walking environment statically and dynamically; based on the data collection in the first part, the pedestrian is modeled as an intelligent body, and the pedestrian characteristics, external environment and behavior are combined and serialized; simulation The part of platform construction and parameter calibration is mainly to use the standard and feasible 3D virtual simulation experiment method of pedestrians in the practical application of model computerized design, select typical scenes for experiments, compare and correct with the visual recognition data of pedestrian walking scenes in actual hubs, and give different The recommended range of experimental parameters under the conditions.

Model design refers to the construction of a simulation model of pedestrian traffic sign perception behavior in the hub, which specifically includes the following aspects:

(1) Determine the pedestrian activity neighborhood;

(2) Construct pedestrian visual perception and attention model:

2.1 Static and dynamic vision

make Respectively represent the line-of-sight angles of the left, right, up, and down directions of the static field of view, and L _s represents the static effective line-of-sight distance of pedestrians, then the sight range of pedestrians in the four directions of left, right, up, and down is When the pedestrian stops searching for the target, the target point within the above range is regarded as the visual potential sensitive point, which is used as the input of bottom-up perception;

make Respectively represent the line-of-sight angles of the left, right, up and down directions of the dynamic field of view, L _d (v) represents the dynamic effective line-of-sight distance of pedestrians, where v is the walking speed of pedestrians, then the left, right, up , the field of view of the next four directions is ; When pedestrians are walking, the target points within the above range are regarded as potential visual sensitive points, which are used as the input of bottom-up perception;

2.2 Pedestrian Walking Attention Focus Extraction

The p neighborhood is used to construct the scene focus. The p neighborhood of the scene is the p neighborhood of v in the set A that is the most adjacent to a certain visual focus v. The symbol N _p (v, A) express;

2.3 Description of pedestrian visual trajectory preference

Let v _h and v _v denote the moving speed of the line of sight in the horizontal and vertical directions, a _v denotes the angle between the line of sight and the vertical plane formed by the walking direction of the pedestrian, and a _h denotes the angle between the line of sight and the horizontal ground in the direction of the pedestrian's walking, It can form the focus area of sight;

Let v ₀ represent the position of the pedestrian's head, v ₁ , v ₂ , v ₃ , v ₄ represent the position of the visual focus, corresponding to quadrants 1, 2, 3, and 4 of the Cartesian coordinate system, and a _{i, j} represent v _i The angle between the position of v j and v _j , L _{i, j} represents the distance between v _i and v _j position, A _k represents the angle of view of the kth type of pedestrian, L(t) represents the distance between the pedestrian and the visual target plane at time t vertical distance, the established mathematical model is as follows:

1) When L _1,4 ≤L(t).tanA _k , L _1,2 ≤L(t).tanA _k , if L _1,4 ≤L _1,2 η _k , where η _k represents the kth class Pedestrian's line of sight horizontal transfer priority coefficient, calibrated through visual simulation experiments, then the probability of visual focus shifting from v ₁ to v ₂ ; If L _1,4 >L _1,2 η _k , the probability of visual focus shifting from v ₁ to v ₄ $P(v_4>v_2)=\frac{1}{1+\exp (L_{\mathrm{1,4}}-L_{\mathrm{1,2}}{\mathrm{\η}}_k)};$

2) When the line of sight turns from focus v ₁ to focus v _j , if 0 ≤ _{a 1, j} ≤ π/4, or 7 π/4 ≤ _{a 1, j} ≤ 2 π, then the direction of v ₁ turning to focus v _j is Horizontal from left to right; if 3π/4≤a _{1, j} ≤5π/4, the direction of v ₁ turning to focus v _j is horizontal from right to left; if π/4≤a _{1, j} ≤3π/4 , then the direction of v ₁ turning to focus v _j is vertical from bottom to top; if 5π/4≤a _{1, j} ≤7π/4, then the direction of v ₁ turning to focus v _j is vertical top to bottom;

2.4 Pedestrian walking visual search rules

The pedestrian eye movement search rules in the hub scene are obtained by repeated iterations according to the p-neighborhood rules on the premise of satisfying the taboo rules.

The specific steps of repeated iterations according to the p neighborhood rule are as follows: set the taboo rule R _s : if a point is selected as the current focus in the N _J generation, it cannot be selected in the N _O generation, and the following search rules for the sight line of pedestrians in the hub scene are obtained:

1) Extract the saliency of the hub scene _P according to the features, and obtain the fused comprehensive hub scene saliency map PS;

2) Let η _d denote the interference noise threshold of the focal salient area in PS, _{S d} denote the focal saliency threshold in PS, and extract the area a _i satisfying the condition in PS, so that | _{S 2} (v _i ) _{-S 2} ( (v _j )≤η _d , S ₂ (v _i )≥S _d or S ₂ (v _j )≥S _d , v _i , v _j ∈a _i , S ₂ (v _i ), S ₂ (v _j ) Indicates the significance of points v _i and v _j ;

3) calculate in Indicates the average saliency of area a _i , |a _i | indicates the number of points contained in area a _i ;

4) Let ρ _e denote the initial line-of-sight focus range threshold in the scene _saliency map PS, and the _salient area a _i in PS is calculated according to the average saliency Sort from high to low, select the most significant in _PS a prominent area, Indicates that ρ _e |a _i | is rounded up; let this The set of points is E _s ;

5) Let x _s (v _i , v _j ) represent the neighborhood selection criteria between points v _i and v _j , randomly select point v _k ∈ E _s , and v _k is the current focus of the scene;

6) Perform p-neighborhood operation on v _k . After obtaining N ₃ (v _k , E _s ), randomly select v _h ∈ N ₃ (v _k , E _s ), and follow the p-neighborhood rule on the premise of satisfying the taboo rule After repeated iterations, the simulated hub scene pedestrian line-of-sight search rules are obtained.

Technical characteristics and effects of the present invention:

The present invention adopts the method of data collection and analysis, and constructs the simulation model of pedestrian traffic sign perception behavior in the hub to simulate the pedestrian traffic sign perception behavior. After the test and verification, it can accurately predict the pedestrian traffic sign perception behavior, and provides strong support for the design of traffic signs and signs. Strong theoretical basis.

Description of drawings

Fig. 1 Schematic diagram of pedestrian simulation step size neighborhood when walking to the right. In the figure, a is fast, b is medium, c is slow, and d is still.

Figure 2 The original image of the interior of the hub.

Fig. 3 Saliency fusion images within hubs.

Figure 4 Eye movement area.

Figure 5. Focus area division.

Fig. 6 Simulation framework of hub pedestrian visual perception and attention model.

Detailed ways

1 Pedestrian traffic sign perception behavior and influencing factors

The environmental and facility factors that affect the perceived behavior of pedestrian traffic signs are the orientation and shape of natural and artificial obstacles between the departure point and the destination, the topological structure of the traffic network and the facilities such as traffic signs in the traffic network can be used; The self-factors of sign perception behavior are pedestrians' cognition degree of geographical environment and traffic environment, that is, the fineness of cognitive map, individual prior knowledge and memory, etc. The factors that affect the perceived behavior of pedestrian traffic signs are shown in the table below.

Table 3-1 List of Factors Affecting Perceived Behavior of Pedestrian Traffic Signs

2 Simulation model of pedestrian traffic sign perception behavior in hub

2.1 Pedestrian activity neighborhood

Pedestrian activity neighborhood refers to the selection area facing the next stage for the current pedestrian's location. This paper divides pedestrian activity neighborhood into three types: step neighborhood, short-term neighborhood and long-term neighborhood. The step size neighborhood refers to the possible position of pedestrians in the next step in the discrete simulation process. In the two-dimensional simulation cell, the grid is generally set according to the average walking speed of the hub. In this paper, the cell length and width are set to 0.4 meters. Figure 1 is a schematic diagram of the step size neighborhood in the direction shown.

As shown in Figure 1, the neighborhoods of pedestrian simulation step lengths are quite different at different walking speeds. When pedestrians are walking at fast and medium speeds, sharp turns are not allowed at the current moment, and the front neighborhood is determined according to the walking speed range. When pedestrians walk slowly, sharp turns are allowed in their step neighborhood, that is, neighborhoods can be set in the left and right areas of the current location. When the pedestrian is stationary, there is currently no walking direction, and the neighborhoods in the four directions of front, back, left, and right are all allowed.

The short-term neighborhood refers to the location that attracts pedestrians due to the existence of traffic facilities such as traffic information boards or other objects that attract pedestrians' attention. The short-term neighborhood is a directional guide for pedestrians to walk. The long-term neighborhood is the possible final destination of pedestrians, such as hub exits, waiting areas, etc. For pedestrians who are not familiar with the layout of the hub, the long-term neighborhood may also change, depending on the update of the short-term neighborhood.

2.2 Pedestrian visual perception and attention model construction

Studies have shown that once human vision has noticed an object, it will have an inhibitory effect on the object, and the object that has been noticed during visual search will no longer attract visual attention. This mechanism is called the forbidden return mechanism of attention. , which corresponds to the short-term memory function of human vision. The visual system needs enough dwell time to complete a visual selection task. It is caused by attention pointing to an area and then moving away from the area. It will delay the response time of attention. The specific delay depends on the difficulty of distinguishing objects. ease. The ban on return typically lasts for a few seconds. Since the object to be noticed is always the most prominent among all the objects participating in the competition, it will always win in the competition, so if there is no specific control mechanism, the focus of attention will always point to the same object, and other objects will not win. Attention opportunity. To avoid this, Kouch and Ullman designed inhibitory negative feedback from the competition network to the interest graph. Once the winning unit is generated, the competition network instantly sends out a pulse, and the interest map then receives an input whose spatial distribution is similar to a differential Gaussian function, and the suppression center is at the target of the winning unit. At this time, the winning target will be shielded, and this part corresponds to disarming the attention on a target. Thereafter, attention shifts to other, more salient objects.

2.2.1 Static and dynamic field of view

The vision of pedestrians in the hub can be divided into static vision and dynamic vision. Static vision refers to the search range of sight when pedestrians stop, and dynamic vision refers to the attention range of pedestrians' sight during walking. In general, static field of view is larger than dynamic field of view.

make Respectively represent the line-of-sight angles of the left, right, up, and down directions of the static field of view, and L _s represents the static effective line-of-sight distance of pedestrians, then the sight range of pedestrians in the four directions of left, right, up, and down is When the pedestrian stops searching for the target, this model takes the target points within the above range as the visual potential sensitive points as the input for bottom-up perception.

make Respectively represent the line-of-sight angles of the left, right, up and down directions of the dynamic field of view, L _d (v) represents the dynamic effective line-of-sight distance of pedestrians, where v is the walking speed of pedestrians. Then the visual range of pedestrians in the four directions of left, right, up and down during walking is When pedestrians are walking, this model regards the target points within the above range as potential visually sensitive points as the input for bottom-up perception. Different from the pedestrian's static field of vision, the size of the pedestrian's dynamic field of vision is related to the walking speed. Generally speaking, the faster the speed, the faster the field of view The narrower it is, the smaller the effective viewing distance L _d (v), and its specific value is calibrated through eye movement experiments.

2.2.2 Pedestrian Walking Attention Focus Extraction

When pedestrians are stationary and walking, the focus of visual attention is constantly changing. How to simulate the focus of attention and visual trajectory of pedestrians is the key to the construction of pedestrian visual perception and attention models.

Use the p neighborhood to construct the scene focus. The p-neighborhood of the scene is the p-neighborhood of v in the set A which is the most adjacent to a certain visual focus v, expressed by the symbol N _p (v, A). As shown in Figure 2 and 3.

Figure 2 is the original image inside the hub, and Figure 3 is the saliency fusion image inside the hub, the most prominent ones in Figure 3 are the 8 focal points as shown in the figure, then A={v ₁ , v ₂ , v ₃ , v ₄ , v ₅ , v ₆ , v ₇ , v ₈ }, if p=3, and the distance between focal points is used as the neighborhood selection criterion, then N ₃ (v ₃ , A)={v ₁ , v ₄ , v ₅ }.

2.2.3 Description of pedestrian visual trajectory preference

During the walking process of pedestrians, the line of sight changes among the focus of attention, forming certain trajectories, and these trajectories obey certain rules. This section describes the description of pedestrian walking visual trajectories.

The eye movement experiment shows that the sight movement process of pedestrians during walking conforms to the following rules:

1) The line of sight moves faster in the horizontal direction than in the vertical direction, and is less prone to fatigue;

2) The change of sight is used to moving from left to right, from top to bottom and clockwise;

3) When the line of sight moves along an oblique direction that is neither horizontal nor vertical, within the range of about 45 degrees up and down in the horizontal direction, the line of sight moves from left to right, and within the range of about 45 degrees left and right in the vertical direction, the line of sight moves from above Move down.

Let v _h and v _v denote the moving speed of the line of sight in the horizontal and vertical directions, a _v denotes the angle between the line of sight and the pedestrian's walking direction and the vertical plane formed by the ground, and a _h denotes the angle of the line of sight in the pedestrian's walking direction and the horizontal ground. Then the focus area of the line of sight as shown in FIG. 4 can be formed.

Let v0 represent the position of the pedestrian's head, v ₁ , v ₂ , v ₃ , and v ₄ represent the position of the visual focus, corresponding to quadrants 1, 2, 3, and 4 of the Cartesian coordinate system. a _{i, j} represents the angle between v _i and v _j positions, L _{i, j} represents the distance between v _i and v _j positions. A _k represents the angle of view of the kth type of pedestrian, and L(t) represents the vertical distance between the pedestrian and the visual target plane at time t. The mathematical model established in this paper for the above laws is as follows:

1) When L _1,4 ≤L(t).tanA _k , L _1,2 ≤L(t).tanA _k , if L _1,4 ≤L _1,2 η _k , where η _k represents the kth class The pedestrian's line of sight horizontal transfer priority coefficient can be calibrated through visual simulation experiments. Then the probability of visual focus shifting from v ₁ to v ₂ If L _1,4 >L _1,2 η _k , the probability of visual focus shifting from v ₁ to v ₄ $P(v_4>v_2)=\frac{1}{1+\exp (L_{\mathrm{1,4}}-L_{\mathrm{1,2}}{\mathrm{\η}}_k)};$

2) When the line of sight turns from focus v ₁ to focus v _j , as shown in the figure below, if 0 ≤ _{a 1, j} ≤ π/4, or 7 π/4 ≤ _{a 1, j} ≤ 2 π, then v ₁ turns to focus The direction of v _j is horizontal from left to right; if 3π/4≤a _{1, j} ≤5π/4, then v ₁ turns to the focal point and the direction of v _j is horizontal from right to left; if π/4≤a _1, If _j ≤ 3π/4, the direction of v ₁ turning to focus v _j is vertical from bottom to top; if 5π/4 ≤ _{a 1, j} ≤ 7π/4, then the direction of v ₁ turning to focus v _j is vertical from top to bottom.

2.2.4 Pedestrian walking visual search rules

In this paper, the pedestrian eye movement search rules in hub scenes are obtained by repeated iterations according to the p-neighborhood rules on the premise of satisfying the taboo rules. Set the taboo rule _{R s} : if a point is selected as the current focus in the _N generation, then it cannot be selected in the N generation, and the following search rules for the sight line of pedestrians in the hub scene are obtained:

1) Extract the saliency of the hub scene _P according to the features, and obtain the fused comprehensive hub scene saliency map PS;

2) Let η _d denote the interference noise threshold of the focal salient area in PS, _{S d} denote the focal saliency threshold in PS, and extract the area a _i satisfying the condition in PS, so that | _{S 2} (v _i ) _{-S 2} ( v _j )|≤η _d , S ₂ (v _i )≥S _d or S ₂ (v _j )≥S _d , v _i , v _j ∈ a _i , S ₂ (v _i ), S ₂ (v _j ) Indicates the significance of points v _i and v _j ;

3) calculate in Indicates the average saliency of area a _i , |a _i | indicates the number of points contained in area a _i ;

4) Let ρ _e denote the initial line-of-sight focus range threshold in the scene _saliency map PS, and the _salient area a _i in PS is calculated according to the average saliency Sort from high to low, select the most significant in _PS a prominent area, Indicates that ρ _e |a _i | is rounded up; let this The set of points is E _s ;

5) Let x _s (v _i , v _j ) represent the neighborhood selection criteria between points v _i and v _j , randomly select point v _k ∈ E _s , and v _k is the current focus of the scene;

6) Perform p-neighborhood operation on v _k . After obtaining N ₃ (v _k , E _s ), randomly select v _h ∈ N ₃ (v _k , E _s ), and follow the p-neighborhood rule on the premise of satisfying the taboo rule After repeated iterations, the simulated hub scene pedestrian line-of-sight search rules are obtained.

2.3 Pedestrian Visual Perception and Attention Model Simulation Framework

The simulation framework of the hub pedestrian visual perception and attention model is as follows:

As shown in Figure 6, the simulation framework of the hub pedestrian visual perception and attention model is divided into four parts: data collection, model design, simulation platform construction, parameter calibration, and method application. Simulation experiments and quantification of pedestrian perception factors, analysis of the scope of influence of different types of pedestrians in different scenarios, the relationship between indicators and internal laws, on the one hand provide basic data support for the construction and calibration of pedestrian visual recognition models, and on the other hand On the one hand, the preconditions for model research are determined by analyzing the index data, and the influencing factors that are not within the scope of the conditions are treated as external interference. The model design part mainly describes the pedestrian walking environment statically and dynamically. Based on the data collection in the first part, the pedestrian is modeled as an intelligent body, and the pedestrian characteristics, external environment and behavior are combined and serialized. The construction of the simulation platform and the parameter calibration part are mainly based on the standard and feasible 3D virtual simulation experiment method of pedestrians in the practical application of the model computerized design, select typical scenes for experiments, compare and correct with the visual recognition data of pedestrian walking scenes in the actual hub, and give Recommended experimental parameter ranges under different conditions. The application part of the method is based on the simulation framework, combined with the specific hub scene, to provide decision support in the design of pedestrian guidance signs, the optimization of the information update frequency of the variable information board, etc.

Claims (2)

1. a pedestrian guide sign design method for view-based access control model perception simulation technology, is characterized in that, comprises data acquisition, modelling, emulation platform builds and parameter calibration, method apply four parts; Part of data acquisition obtains data mainly through the mode of survey, three-dimension virtual reality scenario building, simulated experiment and pedestrian's perception factor; Modelling part mainly carries out Static and dynamic description to pedestrian's environment of walking, build hinge pedestrian traffic mark perception behavior simulation model, on the basis of Part I data acquisition, agent model is carried out to pedestrian, pedestrian's feature, external environment and behavior are carried out combining and serializing; Emulation platform structure and parameter calibration part are mainly by normatron, pedestrian's three-dimensional analogue experiment method of specification in design practical application, typical scene is selected to test, and recognize data and compare and correction with actual hinge pedestrian scene visual of walking, the experiment parameter span of recommending under providing different condition; Method applying portion is on simulation frame basis, in conjunction with concrete hinge scene, in the design of pedestrian's boot flag, variable message board information updating frequency optimization, provides decision support; Modelling specifically comprises: determine pedestrian activity's neighborhood, builds pedestrian's visually-perceptible and attention model; Wherein, structure pedestrian's visually-perceptible and attention model comprise:

(1) static state, kinetic perimetry is determined

Order represent the sight angle in left and right, upper and lower four orientation of static vision field respectively, L _srepresent the effective sighting distance of static state of pedestrian, then pedestrian in the static vision field scope in left and right, upper and lower four orientation is when pedestrian stops search target, using the impact point within the scope of above-mentioned static vision field as the potential sensitive spot of vision, and as the input of perception from bottom to top;

Order represent the sight angle in left and right, upper and lower four orientation of kinetic perimetry respectively, L _dv () represents the dynamically effectively sighting distance of pedestrian, wherein v is the speed of travel of pedestrian, then the kinetic perimetry scope in pedestrian left and right, upper and lower four orientation is in the process of walking when pedestrian walks, using the impact point within the scope of above-mentioned kinetic perimetry as the potential sensitive spot of vision, and as the input of perception from bottom to top;

(2) extract pedestrian to walk focus-of-attention

Adopt p neighborhood to build scene focus, scene p neighborhood is the p neighborhood in set A p the visual focus the most adjacent with certain visual focus v being called v, uses symbol N _p(v, A) represents;

(3) lines of description people walks vision track preference

Make v _h, v _vrepresent the movement velocity of sight line in horizontal and vertical direction, a _vrepresent that sight line and pedestrian's direction of travel formed with angle that the is vertical plane on ground, a _hrepresent sight line be expert on people's direction of travel with the angle of level ground, then can form sight line focus area;

Make v ₀represent pedestrian head position, v ₁, v ₂, v ₃, v ₄represent the position at visual focus place, v ₁, v ₂, v ₃, v ₄correspond respectively to 1 of rectangular coordinate system, 2,3,4 quadrants, a _i,jrepresent v _iwith v _jangle between position, L _i,jrepresent v _iwith v _jdistance between position, A _krepresent the visual angle size of kth class pedestrian, L (t) represents that the span of subscript i, j is all 1,2,3,4 t pedestrian from sensation target plane orthogonal distance; The mathematical model then set up is as follows:

1) L is worked as _{isosorbide-5-Nitrae}≤ L (t) tan A _k, L _1,2≤ L (t) tan A _ktime, if L _{isosorbide-5-Nitrae}≤ L _1,2η _k, wherein η _krepresent the sight line horizontal transfer preferred number of kth class pedestrian, demarcated by visual simulation experiment, L _{isosorbide-5-Nitrae}represent that visual focus is from v ₁transfer to v ₂probability if L _{isosorbide-5-Nitrae}>L _1,2η _k, then visual focus is from v ₁transfer to v ₄probability $P(v_4>v_2)=\frac{1}{1+\exp (L_{1,4}-L_{1,2}{\mathrm{\η}}_k)};$

2) when sight line is from focus v ₁forward focus v to _jtime, if 0≤a _{1, j}≤ π/4, or 7 π/4≤a _{1, j}≤ 2 π, then v ₁forward focus v to _jdirection be level from left to right; If 3 π/4≤a _{1, j}≤ 5 π/4, then v ₁forward focus v to _jdirection be level from right to left; If π/4≤a _{1, j}≤ 3 π/4, then v ₁forward focus v to _jdirection be vertical from bottom to top; If 5 π/4≤a _{1, j}≤ 7 π/4, then v ₁forward focus v to _jdirection be vertical from top to bottom;

(4) obtain pedestrian to walk visual search rule

Hinge scene pedestrian eye moves search rule, is meeting under taboo rule prerequisite, and iterating according to p neighborhood rule obtains.

2. the pedestrian guide sign design method of view-based access control model perception simulation technology as claimed in claim 1, is characterized in that, is arrange taboo rule R according to the p neighborhood rule concrete steps that iterate _s, if fruit dot is at N _jin generation, is selected as current focus, then at N _ocan not select for interior, obtain following hinge scene pedestrian sight line search rule:

1) hinge scene P is carried out conspicuousness extraction according to feature, obtain the comprehensive hinge scene after merging and significantly scheme P _s;

2) η is made _drepresent P _smiddle focus marking area interference noise threshold value, S _drepresent P _smiddle focus significance threshold value, extracts P _sin the region a that satisfies condition _i, make | S ₂(v _i)-S ₂(v _j) |≤η _d, S ₂(v _i)>=S _dor S ₂(v _j)>=S _d, v _i, v _j∈ a _i, S ₂(v _i), S ₂(v _j) represent some v _i, v _jsignificance;

3) calculate wherein represent region a _iaverage significance, | a _i| represent region a _iin the quantity of point that comprises;

4) ρ is made _erepresent that scene significantly schemes P _sin initial sight line focus range threshold, by P _smiddle marking area a _iaccording to significantly average sort from high to low, select P _smiddle significance is maximum individual marking area, represent ρ _e| a _i| round up; Make this the set of individual composition is E _s;

5) x is made _s(v _i, v _j) represent some v _i, v _jbetween neighborhood choice standard, Stochastic choice point v _k∈ E _s, now v _kfor scene current focus;

6) to v _kcarry out p neighborhood operation, the N obtained _j(v _k, E _s), Stochastic choice v _h∈ N _j(v _k, E _s), iterate according to p neighborhood rule meeting under taboo rule prerequisite, namely obtain the hinge scene pedestrian sight line search rule of simulation.