A method for measuring urban walking space perception based on embodied perception

Abstract

The application discloses a kind of city walking space perception measure method based on embodiment perception, comprising: the environmental elements of the city walking space to be perceived measure are determined;Through the walking of the subject in the city walking space to be perceived measure, perception experimental data are obtained;Based on physiological data and eye movement data, obtain comprehensive perception stimulation index;Based on subjective evaluation data, determine subjective cognitive comprehensive index, by the fusion calculation of the subjective cognitive comprehensive index and comprehensive perception stimulation index, obtain cognitive response intensity index, based on space-time behavior index, determine static behavior index and dynamic behavior index;The static behavior index, dynamic behavior index and cognitive response intensity index are coupled calculation, and behavior response index is obtained;Based on comprehensive perception stimulation index, cognitive response intensity index and behavior response index, the perception measure result of the city walking space to be perceived measure is obtained.

Description

Technical Field

[0001] This invention belongs to the field of urban pedestrian spatial perception measurement technology, specifically a method for measuring urban pedestrian spatial perception based on embodied perception. Background Technology

[0002] While existing methods for measuring urban pedestrian spatial perception have made some progress in human-centered exploration, they still have several significant shortcomings. These problems restrict the systematicness, universality, and effectiveness of perception measurement, which can be summarized into the following four points:

[0003] (1) The theoretical foundation is not concise, and the integration of embodied perception processes is not systematic. Although existing methods are based on humanistic perspectives, their theoretical foundations are not yet unified and clear. The "embodied" characteristics of human factors research need to be further explored and refined to form a unified research logic. In addition, existing methods mostly focus on the connection between local links such as "environment-perception" and "environment-behavior", and do not sufficiently analyze the psychological processing mechanism of "perception-cognition" and the decision-making transformation logic of "cognition-behavior", and have not yet formed a systematic connection framework that runs through the entire process.

[0004] (2) Strong scenario dependence and lack of a universal standardized evaluation framework. Existing methods are mostly designed for specific urban spatial types, and the evaluation indicators and processes are significantly scenario-specific due to differences in spatial functional attributes and spatial characteristics. The functional positioning and morphological characteristics of different spaces are significantly different, resulting in evaluation indicators that are highly scenario-bound. Moreover, the various studies have not formed a unified data collection standard and analysis process, making it difficult to reuse across spatial types.

[0005] (3) Limited sample size and insufficient group representativeness. The high cost and complex operation of field experimental equipment limit the recruitment of large-scale samples. At the same time, the access restrictions of specific scenarios and the time cost of subjects further narrow the sample scope, which may lead to the results being affected by the characteristics of specific groups, reducing the universality. As a result, the sample size of existing studies is generally small and it is difficult to reflect a wider range of perceptual patterns.

[0006] (4) Lack of a coupled quantitative mechanism that runs through the entire perception process. Existing methods for measuring urban walking experience generally remain at the level of fragmented analysis of environmental characteristics, subjective feelings, or behavioral performance, lacking a unified mechanism that can continuously express and quantify the various stages of the embodied process. Existing indicator systems are often independent of each other, making it difficult to reflect how environmental factors are gradually transformed into cognitive evaluations and ultimately drive behavioral responses. This results in a lack of calculable causal chains between perceptual results, cognitive interpretations, and behavioral performances, making it difficult for evaluations to form a verifiable end-to-end closed loop.

[0007] In practical research, existing deficiencies do not exist in isolation, but are intertwined, jointly restricting the deepening and application of research on urban pedestrian spatial perception measurement. Specifically, these deficiencies manifest as follows:

[0008] First, the lack of conciseness and systematic integration of the theoretical foundation is the root cause of a series of subsequent problems. Due to the absence of a unified theoretical framework that can span the entire process of "environmental stimulus—perceptual processing—cognitive evaluation—behavioral response," researchers often can only conduct fragmented analyses of specific aspects, making it difficult to develop universally applicable research logic. This directly leads to measurement frameworks relying on empirical and intuitive judgments during construction, lacking a generalizable explanatory structure for the entire process.

[0009] Secondly, the strong scenario dependence and lack of a universally standardized evaluation system further exacerbate the fragmentation of research results. Because different pedestrian spaces differ significantly in their functional positioning and morphological characteristics, existing studies often customize indicator systems and measurement processes for different scenarios, making cross-scenario reuse difficult. Consequently, measurement processes are not universally applicable, results cannot be compared, and research struggles to generate cumulative knowledge.

[0010] Furthermore, existing methods lack a coupled quantitative mechanism that runs through the entire process of perception, cognition, and behavior; that is, they lack a calculable chain that can systematically express "how environmental factors are gradually transformed into cognitive judgments through perceptual stimuli and ultimately affect behavioral performance." Existing indicator systems are mostly discrete, making it difficult to reveal the causal transmission relationship between the three stages. This makes it difficult to form an end-to-end closed loop in evaluation, and ultimately cannot support interpretable reasoning about the mechanism of spatial experience and the precise formulation of spatial optimization strategies.

[0011] Furthermore, the limited sample size and insufficient representativeness of the population further constrain theoretical verification and model extrapolation. Due to the high cost of data collection and the complexity of experimental organization, studies often rely on small, specific sample groups, and the results are easily affected by individual differences, resulting in insufficient generalizability. At the same time, limited samples are insufficient to support iterative feedback of theoretical models, leaving the problem of an unrefined theoretical foundation unresolved for a long time.

[0012] In summary, existing technologies have significant shortcomings in terms of theoretical systematicity, scenario universality, process coupling, and sample representativeness. These problems interact, making it difficult to establish a unified logic in research, hindering the reuse of measurement systems across scenarios, and making it difficult for results to accurately reflect perceptual patterns in urban pedestrian spaces. This limits in-depth analysis of the mechanisms forming the walking experience and weakens the application value of research findings in urban spatial optimization design. Therefore, structural innovation in methodology and measurement systems is urgently needed to address these issues. Summary of the Invention

[0013] Purpose of the invention: To address the problems of strong scene dependence and result distortion caused by laboratory environment in existing urban walking space perception measurement methods, this invention proposes an urban walking space perception measurement method based on embodied perception.

[0014] Technical Solution: This invention proposes a method for measuring urban pedestrian spatial perception based on embodied perception, comprising the following steps:

[0015] Identify the environmental elements of the urban pedestrian space to be perceived and measured;

[0016] Perceptual experimental data was obtained by having subjects walk in an urban walking space to be perceived; the perceptual experimental data included: physiological data, eye movement data, spatiotemporal behavioral data, environmental element data, and subjective evaluation data.

[0017] Based on physiological data, environmental factor data, and eye movement data, a comprehensive perceptual stimulation index is obtained;

[0018] Based on subjective evaluation data, a subjective cognitive comprehensive index is determined. By fusing the subjective cognitive comprehensive index with the comprehensive sensory stimulus index, a cognitive response intensity index is obtained.

[0019] Based on spatiotemporal behavioral data, static behavioral indices and dynamic behavioral indices are determined; the static behavioral indices, dynamic behavioral indices, and cognitive response intensity indices are coupled and calculated to obtain a behavioral response index.

[0020] Based on the comprehensive perceptual stimulus index, cognitive response intensity index, and behavioral response index, the perceptual measurement results of the urban pedestrian space to be perceived are obtained.

[0021] Based on the obtained perception measurement results of the urban pedestrian space to be perceived, an optimized design scheme for the urban pedestrian space to be perceived is obtained.

[0022] Furthermore, the physiological data includes: pulse and respiratory rate; the eye movement data includes: fixation duration for each environmental element; the spatiotemporal behavioral data includes: walking route, location of stopping points, walking speed, and duration of standing still; the environmental element data is the proportion of each environmental element in the photos of the scene of interest taken by the subject.

[0023] The subjective evaluation data refers to the data from the participants' scoring of the perceived urban pedestrian space in the form of a questionnaire survey.

[0024] Furthermore, the comprehensive perceptual stimulation index, derived from physiological data, environmental element data, and eye-tracking data, includes:

[0025] S300: Based on physiological data, environmental factor data, and eye-tracking data, a visual saliency fusion index is defined, expressed as:

[0026]

[0027] In the formula, Let represent the visual saliency fusion index value of the j-th type of environmental element in the i-th interest scene photo. This represents the semantic area percentage of the j-th type of environmental element in the i-th interest scene photo. This represents the percentage of fixation time for the j-th type of environmental element in the i-th interest scene photo within the eye-tracking data. This represents the value after Z-Score standardization. These are the weighting coefficients for semantic area ratio and gaze duration ratio, respectively.

[0028] For any photo of a scene of interest, the visual saliency fusion index of that scene of interest is obtained by weighting the visual saliency fusion index of each environmental element therein.

[0029] S310: Based on physiological data, define a physiological response index, expressed as:

[0030]

[0031] In the formula, This represents the physiological response index corresponding to the i-th interest scene photo. This represents the average or representative statistical value of the subject's pulse during the observation period of the i-th interest scene photo or within the corresponding time window. This represents the average or representative statistical value of the subject's breathing rate during the observation period of the i-th interest scene photo or within the corresponding time window. and These are the weighting coefficients of pulse and respiratory rate in the physiological response index, respectively.

[0032] S320: The intensity of environmental stimuli to perception is defined as the Integrated Perceptual Stimulus Index (ESI), expressed as:

[0033]

[0034] in, This represents the overall perceived stimulation index of the i-th interest scene photo. This represents the weighting coefficient of the visual saliency fusion index in the comprehensive perceptual stimulus index.

[0035] Furthermore, the subjective cognitive comprehensive index determined based on subjective evaluation data is expressed as follows:

[0036]

[0037] In the formula, Let K represent the subjective perception index of the i-th interest scene photo; K represents the total number of subjective evaluation data. This represents the rating of the i-th interest scene photo on the k-th subjective evaluation data, where k is the index of the subjective evaluation data;

[0038] The cognitive response intensity index is obtained by fusing the subjective cognitive comprehensive index with the comprehensive sensory stimulus index, and is expressed as follows:

[0039]

[0040] In the formula, Represents the cognitive response intensity index for the i-th interest scene photo; This represents the comprehensive perceived stimulation index of the i-th interest scene photo. Normalized values; This represents the normalized value of the subjective perception comprehensive index of the i-th interest scene photo; This represents the weighting coefficient of the comprehensive perceived stimulus index in the cognitive response intensity index.

[0041] Furthermore, the determination of static and dynamic behavior indices based on spatiotemporal behavior data specifically includes:

[0042] Based on spatiotemporal behavioral data, a static behavioral index is defined, expressed as:

[0043]

[0044] In the formula, Let represent the static behavior index of the i-th interest scene photo. This represents the average dwell time for a single dwelling action at the i-th interest scene photo. This represents the total duration of all pauses at the i-th interest scene photo. The coefficient of variation represents the duration of dwell time at the i-th interest scene photo. ,in, The standard deviation of dwell time, b1 represents the average dwell time; b2 and b3 represent... , , Weighting coefficients in the static behavior index;

[0045] Based on spatiotemporal behavioral data, a dynamic behavioral index is defined as follows:

[0046]

[0047] In the formula, This represents the dynamic behavior index corresponding to the i-th interest scene photo. , These represent the weighting coefficients of the distance-velocity behavior index and the stationary distribution index in the dynamic behavior index, respectively. This represents the distance-speed behavior index corresponding to the h-th travel route. Represents the dwell distribution index of the i-th interest scene photo;

[0048] Wherein, the distance-speed behavior index corresponding to the h-th travel route , is represented as:

[0049]

[0050] In the formula, This represents the total distance traveled along the h-th route; This represents the average walking speed of the h-th route within the observation time window; The coefficient of variation represents the walking speed along the h-th route. =SD(v) / Mean(v), where SD(v) is the standard deviation of walking speed, and Mean(v) is the mean of walking speed; a1, a2, and a3 are the coefficients of variation of total walking distance, walking speed, and the weighting coefficients of average walking speed in the distance-speed behavior index.

[0051] Among them, the dwell distribution index of the i-th interest scene photo , is represented as:

[0052]

[0053] In the formula, N represents the average nearest neighbor distance of all stops within the urban pedestrian space to be perceived; N() represents a function that performs min-max normalization on the variable.

[0054] Furthermore, the coupling calculation of the static behavior index, dynamic behavior index, and cognitive response intensity index to obtain the behavioral response index specifically includes:

[0055] Based on the static and dynamic behavior indices, a baseline behavioral characteristic index is defined, expressed as:

[0056] ;

[0057] In the formula, Represents the baseline index of behavioral features for the i-th interest scene photo; , These are the weighting coefficients of the static behavior index and the dynamic behavior index in the behavior baseline index, respectively.

[0058] When the urban pedestrian space to be perceived and measured is a specific scene, the original values ​​of the behavioral response index for each interest scene photo are obtained based on the behavioral feature baseline index and the cognitive response intensity index, according to the following formula:

[0059] ;

[0060] In the formula, This represents the original value of the behavioral response index for the i-th interest scene photo; This represents the cognitive response intensity index for the i-th interest scene photo. For cognitive modulation coefficients; a space includes multiple scenes;

[0061] The behavioral response index is obtained by normalizing the raw values ​​of the behavioral response index of each interest scene photo.

[0062] Furthermore, when the urban pedestrian space to be perceived and measured consists of multiple scenes, the original values ​​of the behavioral response index for each interest scene photo are obtained based on the behavioral feature baseline index and the cognitive response intensity index, according to the following formula:

[0063] ;

[0064] In the formula, This represents the original value of the behavioral response index for the i-th interest scene photo; This is the weighting coefficient of the cognitive response intensity index in the behavioral response index;

[0065] The behavioral response index is obtained by normalizing the raw values ​​of the behavioral response index of each interest scene photo.

[0066] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0067] (1) In view of the problems of the lack of concise theoretical foundation and unsystematic integration of embodied perception process in existing technologies, this invention constructs a unified and clear embodied perception theoretical framework. By deeply exploring the core characteristics of "embodied" in human factors research, it forms a systematic correlation logic that runs through the entire process of "environment-perception-cognition-behavior". This makes up for the limitations of existing methods in focusing on the correlation of local links and provides a unified and systematic theoretical guidance for the measurement of urban walking space perception.

[0068] (2) In view of the problems of strong dependence on existing technology scenarios and lack of universal standardized evaluation framework, this invention aims to establish a universal measurement system applicable to multiple types of urban pedestrian spaces. By formulating unified environmental element classification standards, standardizing data collection processes and analysis methods, the binding of evaluation schemes to specific space types is broken, and the problems of inconsistent measurement indicators and difficulty in cross-scenario reuse caused by differences in space functions and forms are solved, so as to realize the universal adaptability of evaluation schemes in various types of pedestrian spaces.

[0069] (3) In view of the problem that the sample size of existing technologies is limited and the group representativeness is insufficient, this invention aims to optimize the sample sampling strategy and data collection mode. By innovating a low-cost and easy-to-operate field experimental scheme, it breaks through the limitations of equipment cost and operation complexity on the sample size, while expanding the sample coverage to improve the group representativeness, solving the problem of insufficient universality of results caused by the limited sample in existing studies, so that the measurement results can more broadly reflect the perception patterns of urban walking space.

[0070] (4) In view of the lack of a coupling and quantification mechanism that runs through the entire perception process in the existing technology, the present invention constructs an embodied perception coupling model with the three stages of “environment-triggered perception-perception-cognition-cognition-driven behavior” as the core, and establishes an end-to-end index system of ESI (Comprehensive Perceptual Stimulus Index)-CRI (Cognitive Response Intensity Index)-BRI (Behavioral Response Index), so that a quantifiable, deductive and verifiable chain mapping relationship is realized between environmental input, subjective cognition and behavioral feedback, which makes up for the defects of fragmented indicators, stage separation and difficulty in forming a continuous calculation framework in the existing technology.

[0071] In summary, this invention aims to address the core deficiencies of existing technologies in terms of theoretical systematicity, scenario universality, process coupling, and sample representativeness, respond to the core pain points of existing technologies, and provide a set of urban pedestrian space perception measurement schemes that are theoretically clear, widely applicable, yield realistic results, and are cost-controllable. This provides a scientific basis for analyzing pedestrian perception patterns and optimizing pedestrian space design, and helps improve the quality and experience of the urban pedestrian environment. Attached Figure Description

[0072] Figure 1 This is a schematic diagram of existing methods for measuring urban pedestrian spatial perception.

[0073] Figure 2 A schematic diagram of the experimental equipment for multi-source data acquisition;

[0074] Figure 3 For the experimental procedure;

[0075] Figure 4 This invention provides a computational framework for an urban walking spatial perception measurement method based on embodied perception.

[0076] Figure 5 This is a schematic diagram of the embodied perception mechanism of the urban walking space perception measurement method based on embodied perception proposed in this invention.

[0077] Figure 6 This is a flowchart of an urban walking space perception measurement method based on embodied perception proposed in this invention. Detailed Implementation

[0078] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will further explain, in conjunction with the accompanying drawings and embodiments, a method for measuring urban pedestrian spatial perception based on embodied perception proposed in this invention.

[0079] like Figure 6 As shown in the figure, this invention proposes a method for measuring urban walking spatial perception based on embodied perception. Based on a comprehensive embodied perception framework, and through optimized embodied experiments, multidimensional data is acquired using pedestrian behaviors such as free walking, stopping to take photos, eye-tracking observation, and subjective evaluation, under legal and compliant conditions. The behavioral and psychological mechanisms at different embodied stages of the walking process are analyzed, and finally, targeted design optimization suggestions are proposed based on the analysis results. This fully presents the research technical path from perceptual experience to spatial intervention. The specific steps are as follows:

[0080] Step 1: Select the urban pedestrian space to be perceived and measure, and define its boundaries and environmental elements. Environmental elements include natural elements and artificial elements. Natural elements include the sky, water bodies, and landscapes, while artificial elements include roads, buildings, and structures. If necessary, the object of measurement itself can be listed as a separate element. The classification criteria and specific meanings of environmental elements are shown in Table 1.

[0081] Table 1 Classification of Environmental Elements

[0082]

[0083] Step 2: Multi-source data acquisition, specific operations include:

[0084] S200: Recruit 10-20 qualified participants, including those who meet the following criteria:

[0085] 1) Good health condition, with no motor or visual impairment.

[0086] 2) All participants were visiting the pedestrian spaces of the city to be tested for the first time to avoid interference from environmental familiarity with the perception data.

[0087] 3) Possess spatial aesthetic perception and visual expression sensitivity.

[0088] S210: Configure experimental equipment, such as Figure 2 As shown, it includes:

[0089] 1) Physiological sensors: used to synchronously collect the pulse and respiratory rate of the subject during walking, including ear clip pulse sensors and belt-type respiratory sensors.

[0090] 2) Eye-tracking devices: used to record gaze coordinates, gaze duration and gaze trajectory, such as the Tobii ProGlasses 3 wearable eye tracker.

[0091] 3) Behavior recording platform: used to collect travel routes, walking speed, stop locations and dwell time, such as Tianjin ErgoLAB spatiotemporal behavior collection software; the "Two Steps" APP can be used to mark stop locations.

[0092] 4) Subjective evaluation tools are used to measure the subjective experience of the participants. A 5-point Likert scale (1-5 points, 1 = very dissatisfied, 5 = very satisfied) is used, which includes 6 indicators: recognizability, sense of security, sense of openness, sense of aesthetics, sense of comfort, and overall experience.

[0093] 5) A camera device used to capture photos of scenes of interest, with one photo corresponding to each scene of interest.

[0094] 6) The experimenter possesses a real-time eye-tracking monitoring platform and a real-time physiological monitoring platform.

[0095] S220: Experiments should be conducted in a daytime environment with uniform and soft outdoor lighting to avoid interference from strong light or extreme weather conditions on the sensing data. For example... Figure 3 As shown, the experiment includes:

[0096] (1) Experimental preparation:

[0097] 1) Fit the subjects with physiological sensors, eye trackers, and positioning devices, and calibrate the device signals;

[0098] 2) Explain the experimental requirements to the participants: walk at a natural walking speed, stop and observe the scene of interest for 10-15 seconds and take a photo. When taking the photo, keep the horizontal angle and line of sight level.

[0099] (2) Recovery phase (2 minutes):

[0100] Subjects were guided to the starting point of the urban walking space to be tested, kept still, and their initial physiological fluctuations were eliminated.

[0101] (3) Free walking phase (approximately 30 minutes):

[0102] 1) The participants started walking freely from the starting point. When passing through scenes of interest, they stopped and observed for 10-15 seconds, and the gaze data was recorded simultaneously through an eye tracker.

[0103] 2) After stopping, use the "Two Steps" APP to mark the current coordinates, take a photo of the scene of interest, and ensure that the composition of the photo is consistent with the field of vision.

[0104] (4) Immediate scoring phase (within 15 minutes after the experiment):

[0105] Participants completed a 5-point Likert scale based on photos of scenes of interest, rating six subjective indicators to ensure the timeliness of the evaluation.

[0106] Step 3: Collect multi-source data in the experiment and process it using the following steps to form a basic dataset based on interest scenes, including:

[0107] Physiological data: The subjects' pulse and respiratory rate were recorded in real time using physiological sensors and synchronously linked to spatiotemporal coordinates. The Tianjin ErgoLAB system's Biometrics module was used to filter and reduce noise and perform baseline correction on the pulse and respiratory rate.

[0108] Eye-tracking data: The eye-tracking device records the fixation duration of each environmental element in the scene of interest, and generates a fixation percentage statistics for each environmental element based on the ErgoLABEyetracking module.

[0109] Spatiotemporal behavioral data: Record the walking route, location of stops, walking speed and duration of stops, preprocess the trajectory data using the ErgoLABBehavior module, and generate a visual walking trajectory map and a heat map of stops using MATLAB programming.

[0110] Environmental element data: The OneFormer semantic segmentation framework is used to calculate the proportion of each type of environmental element in the image for the photos of the scene of interest. In this embodiment of the invention, the OneFormer semantic segmentation framework is trained based on the ADE20K dataset.

[0111] Subjective evaluation data: After collecting the questionnaires, the original scores of each indicator were recorded, the subject number and corresponding interest scene were marked, and the Z-Score standardization formula was used to eliminate individual score differences and unify the scale.

[0112] Table 2 Embodied Experiment Data

[0113]

[0114] Step 4: Based on embodied perception theory, construct the environment-triggered perception stage, the perception-establishing cognition stage, and the cognition-driven behavior stage.

[0115] In the environmental trigger perception stage, the visual saliency fusion index and the physiological response index are formed and combined into the comprehensive perceptual stimulus index (ESI), which is used to characterize the degree of immediate stimulation of the pedestrian's perceptual system by environmental elements.

[0116] In the perception-based cognition stage, a subjective cognitive comprehensive index (SCI) is constructed based on the subjective cognitive evaluation results. This SCI is then fused with the comprehensive perceptual stimulus index (ESI) to obtain the cognitive response intensity index (CRI), which reflects the strength and consistency of pedestrians' cognitive judgments based on perception.

[0117] In the cognitive-driven behavior stage, a static behavior index (SBI) and a dynamic behavior index (DBI) are constructed using static standing and dynamic movement data. These are then coupled with a cognitive response intensity index (CRI) to obtain the final behavioral response index (BRI), which is used to quantify the impact and explicit strength of cognitive processing on behavioral decisions. Figure 4 As shown.

[0118] In the environmentally triggered perception phase, the Integrated Perception Stimulus Index (ESI) is calculated according to the following steps:

[0119] S410: The Visual Salience Index (VSI) is defined as the comprehensive visual stimulus intensity that measures both the semantic proportion of each environmental element in an image and the proportion of actual gaze duration in a given scene. It is expressed as:

[0120]

[0121] In the formula, Let represent the visual saliency fusion index value of the j-th type of environmental element in the i-th interest scene photo. This represents the semantic area percentage of the j-th type of environmental element in the i-th interest scene photo. This represents the percentage of fixation time for the j-th type of environmental element in the i-th interest scene photo within the eye-tracking data. This represents the value after Z-Score standardization. The weighting coefficients for semantic area proportion and gaze duration proportion are respectively. They can be initially set to 0.5 and 0.5, or determined by data-driven methods such as principal component analysis (PCA) and entropy weighting.

[0122] For any photograph of a scene of interest, the visual saliency fusion index of that scene is obtained by weighting the visual saliency fusion index of each environmental element within it:

[0123]

[0124] in, The weights of each environmental element's VSI can be assigned based on the element's importance or determined using PCA.

[0125] S411: The intensity of the impact of a scenario on the body's physiological activation / load (heart rate, respiratory rate, etc.) is defined as the Physiological Response Index (PRI), calculated according to the following:

[0126]

[0127] In the formula, This represents the physiological response index corresponding to the i-th interest scene photo. This represents the average or representative statistical value of the subject's pulse during the observation period of the i-th interest scene photo or within the corresponding time window. This represents the average or representative statistical value of the subject's breathing rate during the observation period of the i-th interest scene photo or within the corresponding time window. and These are the weighting coefficients of pulse and respiratory rate in the physiological response index, respectively. They can be initially set to 0.5 and 0.5, or determined by methods such as principal component analysis (PCA) or entropy weighting.

[0128] The physiological response index was normalized to 0–1.

[0129] S412: The "external stimulus intensity" index, which integrates the Visual Salience Fusion Index (VSI) and the Physiological Response Index (PRI) in a spatial reference frame, is used to identify the strength of environmental stimuli to the perceptual system. This index is defined as the Integrated Perceptual Stimulus Index (ESI), expressed as:

[0130]

[0131] In the formula, Represents the comprehensive perceived stimulation index of the i-th interest scene photo; This represents the visual saliency fusion index of the i-th interest scene photo. This represents the physiological response index of the i-th interest scene photo. This represents the weighting coefficient of visual saliency in the comprehensive perceptual stimulus index. These are the weighting coefficients for the physiological response. It can be determined through regression analysis with subjective overall evaluation as the dependent variable, or an appropriate value can be selected within the range of 0 to 1 based on experience.

[0132] The comprehensive perceived stimulus index is normalized to [0,1] to obtain the final comprehensive perceived stimulus index.

[0133] In the perception-cognition stage, the Subjective Cognitive Composite Index (SCI) is calculated using the following steps:

[0134] The subjects' multidimensional subjective evaluations of the scene of interest (identifiability, aesthetics, safety, openness, comfort, etc.) are combined into a single indicator reflecting the intensity of subjective cognition, which is defined as the Subjective Cognition Index (SCI).

[0135]

[0136] In the formula, The subjective perception comprehensive index value of the i-th interest scene photo is represented; K represents the total number of subjective evaluation data. In this embodiment of the invention, K=6, corresponding to six indicators: recognizability, sense of security, sense of openness, sense of aesthetics, sense of comfort, and sense of overall experience. This represents the rating of the i-th interest scene photo on the k-th subjective evaluation data, where k is the index of the subjective evaluation data.

[0137] The cognitive intensity / clarity formed after a scene of interest is perceived is defined as the Cognitive Response Index (CRI). The CRI is calculated by fusing the Subjective Cognitive Composite Index (SCI) and the Integrated Perceptual Stimulus Index (ESI), and is expressed as follows:

[0138]

[0139] In the formula, Represents the cognitive response intensity index for the i-th interest scene photo; This represents the comprehensive perceived stimulation index of the i-th interest scene. The value after normalization or standardization; This represents the subjective cognitive comprehensive index of the i-th interest scene. The value after normalization or standardization; The weighting coefficients of the perceived stimulus in the cognitive response intensity index are used to determine the overall weighting coefficients. The weighting coefficients of the subjective perception comprehensive index. This can be achieved by using the overall subjective experience rating as the dependent variable, and... and The regression coefficients in a multiple regression model with independent variables are obtained by normalization, and can also be set to 0.5 equal weights based on experience.

[0140] In the cognitive-driven behavior stage, the behavioral response index BRI is calculated so that it includes CRI (and may include ESI) to reflect the numerical link of "cognitive-driven behavior".

[0141] The intensity of dwell behavior, including the duration of a single dwelling, the total duration of dwelling, and the variability of dwelling duration, is defined as the SBI (Static Behavior Index), expressed as:

[0142]

[0143] In the formula, Represents the static behavior index of the i-th interest scene photo; This represents the average dwell time for a single dwelling action at the i-th interest scene photo; This represents the total duration of all pauses at the i-th interest scene photo; represents the coefficient of variation of the dwell time at the i-th interest scene photo. ,in The standard deviation of dwell time, b1 represents the average dwell time; b2 and b3 represent... , and The weighting coefficients in the static behavior index can be determined through methods such as expert weighting, entropy weighting, or regression analysis.

[0144] The Dynamic Behavior Index (DBI) consists of two parts: the Distance-Speed ​​Index (DSI) and the Pause Distribution Index (PDI).

[0145] The exploratory / proactive nature of pedestrian paths, measured by factors such as path length, average speed, and speed variability, is defined as the DSI (Distance-Velocity Behavior Index), expressed as:

[0146]

[0147] In the formula, This represents the distance-speed behavior index corresponding to the h-th travel route; This represents the total distance traveled along the h-th route; This represents the average walking speed of the h-th route within the observation time window; The coefficient of variation represents the walking speed along the h-th route. = SD(v) / Mean(v), where SD(v) is the standard deviation of walking speed, and Mean(v) is the mean of walking speed; a1, a2, and a3 are the coefficients of variation of total walking distance, walking speed, and the weighting coefficients of average walking speed in the distance-speed behavior index.

[0148] The spatial concentration / dispersion of rest points (represented by the average shortest distance to nearest neighbors or kernel density) is defined as the PDI (Pause Distribution Index). Concentration indicates focused behavior, while dispersion indicates broad exploration, expressed as:

[0149]

[0150] In the formula, Indicates the dwelling distribution index; N represents the average nearest neighbor distance of all stops within the urban pedestrian space to be perceived and measured, used to characterize the overall concentration or dispersion of stops in space; N() represents the function that performs min-max normalization on the variables, linearly mapping NND to the interval [0,1].

[0151] Calculate the minimum distance from each stop point to other stop points. .

[0152] Take the sample average NND:

[0153]

[0154] Therefore, the Dynamic Behavior Index (DBI) is expressed as:

[0155]

[0156] In the formula, This represents the dynamic behavior index corresponding to the i-th interest scene photo. , These represent the weighting coefficients of the distance-speed behavior index and the dwelling distribution index in the dynamic behavior index, respectively. They can be determined through expert weighting, entropy weighting, or regression analysis with a certain behavior output variable (such as total dwelling time) as the dependent variable.

[0157] Finally, the DBI is normalized to 0–1.

[0158] The comprehensive reflection of how cognition (CRI) is transformed into observable behavior (SBI, DBI), while retaining the direct or indirect impact on the stimulus (ESI), is defined as the Behavioral Response Index (BRI). The BRI indicates whether a scenario truly elicits a behavioral response (such as dwelling, close observation, or wandering). Specific calculations include:

[0159] Based on static and dynamic behavioral indices, a baseline behavioral feature index (BFI) is defined:

[0160]

[0161] In the formula, Represents the baseline index of behavioral features for the i-th interest scene photo; , These are the weighting coefficients of the static behavior index and the dynamic behavior index in the behavior baseline index, respectively.

[0162] When the urban pedestrian space to be perceived and measured is a specific scene, the raw values ​​of the behavioral response index for each interest scene photo are obtained based on the behavioral feature baseline index and the cognitive response intensity index, and are expressed as follows:

[0163]

[0164] In the formula, This represents the original value of the behavioral response index for the i-th interest scene photo; This represents the cognitive response intensity index for the i-th interest scene photo. Cognitive modulation coefficient;

[0165] Specifically, when a scenario exhibits high uncertainty or risk perception, and when differences in behaviors such as walking, stopping, and detouring under similar physical environmental conditions are primarily driven by cognitive factors such as sense of security, orientation, and environmental readability, a cognitive modulation approach should be adopted. In this case, the behavioral baseline index represents an individual's potential behavioral capabilities in an objective environment, while the cognitive response intensity index, as a higher-order cognitive state, amplifies or inhibits this behavioral potential. This index is used to characterize the modulation mechanism of "whether behavior is triggered and its intensity changes," making it suitable for mechanism analysis and key scenario identification.

[0166] Alternatively, when the urban pedestrian space to be perceived and measured consists of multiple scenes, the behavioral characteristic baseline index and the cognitive response intensity index are linearly mixed to obtain the raw values ​​of the behavioral response index for each interest scene photo, expressed as:

[0167]

[0168] In the formula, This represents the original value of the behavioral response index for the i-th interest scene photo; This is the weighting coefficient of the cognitive response intensity index in the behavioral response index;

[0169] Specifically, when conducting overall quality assessments and horizontal comparisons of different pedestrian spaces or planning schemes, and when behavioral performance and cognitive perception show a stable statistical correlation, a linear hybrid approach can be adopted. In this case, the behavioral characteristic baseline index and the cognitive response intensity index are considered as indicators that contribute equally to the perceived quality of pedestrian spaces. By weighted fusion, a comprehensive evaluation result is formed to improve the stability and comparability of the indicators, making it suitable for batch analysis, scheme comparison, and engineering application scenarios.

[0170] The output is BRI after normalization.

[0171] The names and meanings of the indicators in this step are shown in Table 3.

[0172] Table 3 Meaning of Indicators

[0173]

[0174]

[0175] Step 5: Based on the causal chain characteristics in the process of pedestrian experience formation, statistical modeling analysis, spatial visualization, and scene classification and identification are performed on the indicator results, and targeted pedestrian space design optimization strategies are proposed accordingly.

[0176] (1) Regression analysis: The significance and weight difference of the path relationship of "environment--perception--cognition--behavior" are tested by using the regression model to verify the validity and dominant direction of the embodied perception mechanism in the research scenario.

[0177] Model example:

[0178]

[0179] The degree of influence at each stage is determined by the values ​​of α, β, and γ.

[0180] (2) Spatial heat map analysis: ESI, CRI and BRI are interpolated in space to generate heat maps, realizing the spatial distribution identification of stimulus intensity, subjective comfort and behavioral response (Table 4).

[0181] Input ESI, CRI, and BRI as point values ​​into ArcGIS and use Kriging heatmaps. The ESI heatmap shows the spatial distribution of environmental stimulus intensity (such as light, noise, visual information density, etc.), the CRI heatmap shows the spatial distribution of subjective experience quality, and the BRI heatmap shows the spatial distribution of behavior (such as staying, exploring, etc.).

[0182] Table 4. Significance and Representation of Indicators

[0183]

[0184] (3) Classification and design optimization: Based on the distribution pattern of the three types of indicator feature values, interest scenes are clustered and classified to identify spatial types with differentiated experience characteristics. Based on the causes of interest scenes, corresponding and implementable environmental design and spatial reshaping strategies are proposed (Table 5).

[0185] Table 5. Interest Scene Type Classification and Design Optimization Strategies

[0186]

[0187] In the field of urban pedestrian space perception measurement, existing technologies suffer from insufficient full-process analysis, strong scene dependence, inefficient sampling strategies, and limited computational coupling, making it difficult to meet the needs of accurate assessment and cross-scenario application. This invention addresses the shortcomings of existing technologies in theoretical framework, application scope, result accuracy, and data reliability by clarifying the core mechanism of embodied perception, overcoming scene limitations, optimizing sampling strategies, and proposing a coupled measurement system. Ultimately, it provides accurate and effective methodological support for the scientific assessment and optimized design of urban pedestrian spaces. Specifically:

[0188] (1) Provide a complete theoretical model of embodied perception mechanism

[0189] To address the issue of insufficiently refined theoretical foundations in existing methods, this invention, centered on embodied perception theory, constructs a clear theoretical model and analytical path. Specifically, this invention leverages the Stimulus-Organism-Response (SOR) model from environmental psychology, extending it to the spatial perception process. Embodied perception is defined as a dynamic cyclical process of "environment-perception-cognition-behavior," encompassing three stages: environment-triggered perception, perception-based cognition, and cognition-driven behavior. Compared to the local correlation analysis of existing methods, the "environment-perception-cognition-behavior" embodied perception mechanism model constructed in this invention significantly enhances the explanatory power of perception measurement.

[0190] Based on this embodied perception mechanism, this invention constructs an analytical method covering the entire embodied process: In the environment-triggered perception stage, by statistically analyzing the image semantics and eye-tracking fixation ratio of various environmental elements, the visual perception of pedestrians is quantitatively analyzed, and the spatiotemporal characteristics of physiological responses are analyzed; In the perception-establishment cognition stage, the process of pedestrians establishing subjective cognition of space is analyzed based on questionnaire survey results, and pedestrians' scene cognition preferences are analyzed by combining Pearson correlation analysis, SHAP interpretability analysis, and other methods; In the cognition-driven behavior stage, pedestrian action patterns and decision-making logic are analyzed from the perspectives of static standing and dynamic trajectories, relying on spatiotemporal behavioral data.

[0191] (2) Provide a perception measurement method that can be widely applied to urban pedestrian spaces.

[0192] Existing technologies rely excessively on the spatial characteristics of specific scenarios, and their overly scenario-specific nature makes it difficult to transfer their evaluation indicators, analytical logic, and processes to other types of pedestrian spaces, resulting in low reusability of research findings. This invention overcomes this limitation by constructing a standardized method: Firstly, it establishes a standardized environmental element classification system, uniformly summarizing the environmental composition of various pedestrian spaces into natural elements (including sky, water, and landscape) and artificial elements (including buildings, structures, roads, and the measurement objects themselves). It also establishes quantitative standards for these elements based on technologies such as semantic segmentation. This classification method removes scenario-specificity, allowing for comparative analysis of environmental characteristics across different spaces on the same dimension. Secondly, it provides a unified data collection and analysis process, standardizing the collection methods for multimodal data such as eye-tracking observation, physiological monitoring, and behavioral recording, and ensuring that the analytical logic does not depend on specific spatial functions. Therefore, this invention can be directly applied to different types of pedestrian spaces such as streets, park trails, and pedestrian bridges. This versatility not only lowers the technical threshold for cross-scenario research but also makes the perceptual patterns of different spaces comparable, providing a unified quantitative analysis tool for the overall optimization of urban pedestrian systems and effectively solving the scenario-dependent problem of existing technologies.

[0193] (3) Provide a method for creating a database by sampling small samples at high frequency.

[0194] Existing technologies suffer from limited sample size and insufficient population representativeness, resulting in limited generalizability of results. This invention addresses this issue by proposing a small-sample, high-frequency sampling method: through rigorous selection of subjects with spatial perception sensitivity, multiple data types, including physiological, eye-movement, behavioral, and subjective evaluation data, are recorded frequently in a single experiment, providing high-density data coverage of perceptual details. In terms of data density, this method can cover multiple environmental factors, subjective indicators, and behavioral patterns, with single-sample data richness superior to traditional methods. Regarding result stability, cross-validation between standardized subjective evaluation data and physiological and behavioral data confirms its reliability. In terms of economic efficiency, compared to large-scale sample studies, it significantly reduces equipment costs, shortens the experimental cycle, and meets design guidance requirements in terms of result accuracy. By compensating for sample size limitations through data depth, this invention can construct a high-quality database reflecting embodied perception patterns, effectively improving the reliability, representativeness, and generalizability of the results.

[0195] (4) Provide a coupled quantitative measurement system that runs through the entire process of "environment-perception-cognition-behavior".

[0196] While existing technologies can describe specific aspects of environmental characteristics, subjective feelings, or behavioral performance, they lack a unified measurement model that connects the four elements of "environmental stimulus—perceptual processing—cognitive evaluation—behavioral response." This results in a lack of inherent logical connections between different stages, with indicators appearing fragmented and failing to reveal the complete mechanism of pedestrian experience formation. To address this issue, this invention proposes a progressively coupled quantitative system based on embodied processes, enabling a computable representation of the entire process.

[0197] Specifically, this invention calculates the ESI (Economic Stimulus Index) in the environmentally triggered perception stage to characterize the intensity of environmental stimuli triggering the perception system; it calculates the CRI (Cognitive Response Intensity Index) in the perception-establishment cognition stage to characterize the depth and consistency of cognitive evaluation formed by pedestrians based on perceptual input; and it calculates the BRI (Behavioral Response Index) in the cognitive-driven behavior stage to quantify the degree of transformation of cognitive evaluation into actual behavioral decisions and action patterns. These three indices correspond to the three stages of the embodied perception process and are coupled step-by-step to form a complete logical chain: “ESI → CRI → BRI”.

[0198] Compared to existing studies where results at each stage are independent and difficult to derive, this invention uses the index of the previous stage as an important input variable for the calculation of the next stage. This allows the process of perception influencing cognition and cognition driving behavior to be quantified and structurally explained. Simultaneously, ESI, CRI, and BRI from different spaces can be mapped to the same indicator space, achieving cross-space comparability, interpretability, and verifiability. Therefore, this invention not only solves the problem of broken chains in experience measurement but also enables the formation process of urban pedestrian space experience to be expressed with clear mathematical logic and a visual structure, effectively improving the analytical depth and model completeness of perception research.

[0199] In summary, this invention comprehensively enhances the technical level of urban pedestrian space perception measurement from four dimensions: systematicity, universality, economy, and scientific rigor, by perfecting the whole-process mechanism analysis, overcoming scenario limitations, optimizing sampling strategies, and proposing a coupled quantification mechanism. These breakthroughs not only effectively solve the core pain points of existing technologies but also form a set of scalable and highly adaptable measurement schemes, providing technical support for the scientific evaluation, optimized design, and quality improvement of urban pedestrian spaces, and possessing significant theoretical and practical value.

Claims

1. A method for measuring urban pedestrian spatial perception based on embodied perception, characterized in that: Includes the following steps: Identify the environmental elements of the urban pedestrian space to be perceived and measured; Perceptual experimental data was obtained by having subjects walk in an urban walking space to be perceived; the perceptual experimental data included: physiological data, eye movement data, spatiotemporal behavioral data, environmental element data, and subjective evaluation data. Based on physiological data, environmental factor data, and eye movement data, a comprehensive perceptual stimulation index is obtained; Based on subjective evaluation data, a subjective cognitive comprehensive index is determined. By fusing the subjective cognitive comprehensive index with the comprehensive sensory stimulus index, a cognitive response intensity index is obtained. Based on spatiotemporal behavioral data, static behavioral indices and dynamic behavioral indices are determined; the static behavioral indices, dynamic behavioral indices, and cognitive response intensity indices are coupled and calculated to obtain a behavioral response index. Based on the comprehensive perceptual stimulus index, cognitive response intensity index, and behavioral response index, the perceptual measurement results of the urban pedestrian space to be perceived are obtained. Based on the obtained perception measurement results of the urban pedestrian space to be perceived, an optimized design scheme for the urban pedestrian space to be perceived is obtained. The physiological data includes pulse and respiratory rate; the eye movement data includes the duration of fixation on various environmental elements; the spatiotemporal behavioral data includes the walking route, location of stopping points, walking speed, and duration of standing still; the environmental element data is the proportion of each environmental element in the photos of the scene of interest taken by the subject. The subjective evaluation data refers to the data from the subjects' scoring of the urban pedestrian space in the form of a questionnaire survey. The comprehensive perceptual stimulation index, derived from physiological data, environmental element data, and eye-tracking data, includes: S300: Based on physiological data, environmental factor data, and eye-tracking data, a visual saliency fusion index is defined, expressed as: ； In the formula, Let represent the visual saliency fusion index value of the j-th environmental element in the i-th interest scene photo, where i is the index of the interest scene photo and j is the index of the environmental element. This represents the semantic area percentage of the j-th type of environmental element in the i-th interest scene photo. This represents the percentage of fixation time for the j-th type of environmental element in the i-th interest scene photo within the eye-tracking data. This represents the value after Z-Score standardization. These are the weighting coefficients for semantic area ratio and gaze duration ratio, respectively. For any photo of a scene of interest, the visual saliency fusion index of that scene of interest is obtained by weighting the visual saliency fusion index of each environmental element therein. S310: Based on physiological data, define a physiological response index, expressed as: ； In the formula, This represents the physiological response index corresponding to the i-th interest scene photo. This represents the average or representative statistical value of the subject's pulse during the observation period of the i-th interest scene photo or within the corresponding time window. This represents the average or representative statistical value of the subject's breathing rate during the observation period of the i-th interest scene photo or within the corresponding time window. and These are the weighting coefficients of pulse and respiratory rate in the physiological response index, respectively. S320: The intensity of environmental stimuli to perception is defined as the Integrated Perceptual Stimulus Index (ESI), expressed as: ； in, This represents the overall perceived stimulation index of the i-th interest scene photo. This represents the weighting coefficient of the visual saliency fusion index in the comprehensive perceptual stimulus index.

2. The method for measuring urban pedestrian spatial perception based on embodied perception according to claim 1, characterized in that: The subjective perception comprehensive index, determined based on subjective evaluation data, is expressed as follows: ； In the formula, Let K represent the subjective perception index of the i-th interest scene photo; K represents the total number of subjective evaluation data. This represents the rating of the i-th interest scene photo on the k-th subjective evaluation data, where k is the index of the subjective evaluation data; The cognitive response intensity index is obtained by fusing the subjective cognitive comprehensive index with the comprehensive sensory stimulus index, and is expressed as follows: ； In the formula, Represents the cognitive response intensity index for the i-th interest scene photo; This represents the comprehensive perceived stimulation index of the i-th interest scene photo. Normalized values; This represents the normalized value of the subjective perception comprehensive index of the i-th interest scene photo; This represents the weighting coefficient of the comprehensive perceived stimulus index in the cognitive response intensity index.

3. The method for measuring urban pedestrian spatial perception based on embodied perception according to claim 2, characterized in that: The determination of static and dynamic behavior indices based on spatiotemporal behavioral data specifically includes: Based on spatiotemporal behavioral data, a static behavioral index is defined, expressed as: ； In the formula, Let represent the static behavior index of the i-th interest scene photo. This represents the average dwell time for a single dwelling action at the i-th interest scene photo. This represents the total duration of all pauses at the i-th interest scene photo. The coefficient of variation represents the duration of dwell time at the i-th interest scene photo. =SD(Stay) / Mean(Stay), where SD(Stay) is the standard deviation of the dwell time, and Mean(Stay) is the mean of the dwell time; b1, b2, and b3 are the weighting coefficients of the above three behavioral quantities in the static behavioral index; based on spatiotemporal behavioral data, a dynamic behavioral index is defined as: ； In the formula, This represents the dynamic behavior index corresponding to the i-th interest scene photo. , These represent the weighting coefficients of the distance-velocity behavior index and the stationary distribution index in the dynamic behavior index, respectively. This represents the distance-speed behavior index corresponding to the h-th travel route. Represents the dwell distribution index of the i-th interest scene photo; Wherein, the distance-speed behavior index corresponding to the h-th travel route , is represented as: ； In the formula, This represents the total distance traveled along the h-th route; This represents the average walking speed of the h-th route within the observation time window; The coefficient of variation represents the walking speed along the h-th route. =SD(v) / Mean(v), where SD(v) is the standard deviation of walking speed, and Mean(v) is the mean of walking speed; a1, a2, and a3 are the coefficients of variation of total walking distance, walking speed, and the weighting coefficients of average walking speed in the distance-speed behavior index. Among them, the dwell distribution index of the i-th interest scene photo , is represented as: ； In the formula, N represents the average nearest neighbor distance of all stops within the urban pedestrian space to be perceived; N() represents a function that performs min-max normalization on the variable.

4. The method for measuring urban pedestrian spatial perception based on embodied perception according to claim 3, characterized in that: The method of coupling and calculating the static behavior index, dynamic behavior index, and cognitive response intensity index to obtain the behavior response index specifically includes: Based on the static and dynamic behavior indices, a baseline behavioral characteristic index is defined, expressed as: ； In the formula, Represents the baseline index of behavioral features for the i-th interest scene photo; , These are the weighting coefficients of the static behavior index and the dynamic behavior index in the behavior baseline index, respectively. Based on the behavioral feature baseline index and the cognitive response intensity index, the original values ​​of the behavioral response index obtained by each interest scene photo under the modulation of the cognitive response intensity index are expressed as follows: ； In the formula, This represents the original value of the behavioral response index obtained by modulating the cognitive response intensity index on the i-th interest scene photo; This represents the cognitive response intensity index for the i-th interest scene photo. Cognitive modulation coefficient; The behavioral response index is obtained by normalizing the original values ​​of the behavioral response index obtained from the photos of each interest scene under the modulation of the cognitive response intensity index.

5. The method for measuring urban pedestrian spatial perception based on embodied perception according to claim 3, characterized in that: The method of coupling and calculating the static behavior index, dynamic behavior index, and cognitive response intensity index to obtain the behavior response index specifically includes: Based on the static and dynamic behavior indices, a baseline behavioral characteristic index is defined, expressed as: ； In the formula, Represents the baseline index of behavioral features for the i-th interest scene photo; , These are the weighting coefficients of the static behavior index and the dynamic behavior index in the behavior baseline index, respectively. By linearly mixing the baseline behavioral characteristic index and the cognitive response intensity index, the original values ​​of the behavioral response index for each interest scene photo under the linear mixing method are obtained, expressed as: ； In the formula, This represents the original value of the behavioral response index obtained by the i-th interest scene photo under linear mixing mode; This is the weighting coefficient of the cognitive response intensity index in the behavioral response index; The behavioral response index is obtained by normalizing the raw values ​​of the behavioral response index obtained from the photos of each interest scene under the linear mixing method.

Images