A generative design method for urban pedestrian space layout in cold regions based on dynamic thermal comfort prediction

Abstract

This invention proposes a method for generating and designing pedestrian space layouts in cold-region cities based on dynamic thermal comfort prediction. Based on big data and IoT data, this invention constructs a pathfinding agent model for cold-region pedestrian spaces to obtain pedestrian trajectories at different travel times under different layouts. Based on pedestrian physiological index data, it constructs a mapping relationship between pedestrian thermal sensation and the thermal environment of cold-region pedestrian spaces to obtain pedestrian thermal sensation under different thermal environment changes. Based on the predicted pedestrian thermal sensation results along typical pedestrian routes at different times, the layout of cold-region pedestrian spaces is optimized. Finally, based on machine learning models and designer feedback on visualized solutions, a decision-making model for cold-region pedestrian space layout design guided by designer decision preferences is obtained. For predicting pedestrian thermal sensation in cold-region pedestrian spaces, this invention is more realistic and targeted than existing technologies, while also being highly efficient.

Description

Technical Field

[0001] This invention belongs to the field of thermal comfort prediction technology in cold-region cities, and in particular relates to a method for generating and designing pedestrian space layouts in cold-region cities based on dynamic thermal comfort prediction. Background Technology

[0002] As research progresses, the concept of thermal comfort is constantly evolving to better describe users' thermal sensations in real-world thermal environments. Initially, research equated thermal neutrality with thermal comfort, thus describing a state where users felt neither cold nor hot. This concept emphasized the direct stimulation of the sensory system by the thermal environment, neglecting the subjective influence of psychological factors on thermal sensation. Thermal comfort under this concept was steady-state and discrete. However, research by de Dear et al. showed that thermal sensation is not only related to the current thermal state but is also influenced by thermal history, thermal expectations, and thermal adaptation levels. Therefore, thermal sensation is actually dynamic, continuous, and personalized, with the influence of thermal alliesthesia. In this context, thermal neutrality does not necessarily lead to the most positive thermal experience; their research found that brief deviations from thermal neutrality within a certain range can bring pleasurable sensations. Therefore, thermal pleasure has been incorporated into the consideration of thermal comfort. Based on this, a dynamic thermal comfort theory considering the continuous influence of the environment was proposed to describe users' subjective thermal sensations in non-steady-state environments.

[0003] Currently, most methods for predicting thermal comfort are based on steady-state models, and their predictions only consider deviations from thermal neutrality, not thermal comfort. Existing thermal comfort models include single-node, two-node, and multi-node models. Single-node models treat the human body as a single node, representing the body's overall temperature perception. These models typically assess thermal comfort based on average whole-body temperature and perspiration evaporation rate. Key models include Fanger's PMV and PPD models, and widely used thermal comfort standards ASHRAE 55-2016 and ISO 7730 are based on these. They predict average user responses to ambient temperature based on factors such as ambient temperature, humidity, wind speed, radiant temperature, clothing index, and activity level. This model is a steady-state model, not considering thermal history or changes in temperature and comfort over time in a state of thermal equilibrium. Two-node thermal comfort models divide the human body into two nodes: the core and the skin. Compared to single-node models, this model can more accurately simulate the body's heat conduction and thermoregulation processes, such as heat exchange between the body and the skin. These primarily include the SET model by Gagge et al., which uses physiological equivalent temperature for thermal comfort prediction and evaluation. While this model considers human heat exchange to some extent, it is generally used under steady-state assumptions and does not consider the influence of thermal history. Multi-node thermal comfort models further divide the human body into multiple nodes such as the head, trunk, and limbs to simulate the temperature and thermal perception of different parts, enabling a more refined simulation of temperature distribution and thermoregulation processes in different parts of the body. This mainly includes the UTCI model, which uses complex physiological and thermoregulation models to convert the thermal stress level under different environmental conditions into an equivalent temperature index. Although this model can simulate non-steady-state environments, it also measures thermal comfort levels based on the degree of deviation from thermal neutrality, neglecting thermal pleasure and thus failing to adequately consider the influence of users' psychological thermal perception. In summary, current thermal comfort models do not adequately consider the impact of continuous dynamic environments and users' subjective feelings, resulting in deficiencies in the accuracy of predicting users' thermal comfort in real-world environments.

[0004] The thermal environment of pedestrian spaces in cold regions varies significantly in time and space. Therefore, the temporal and spatial changes experienced by pedestrians during their walk lead to variations in the thermal environment, dynamically influencing their thermal perception and causing fluctuations in skin temperature, resulting in either positive or negative thermal sensations. However, current prediction methods do not fully consider this impact when assessing pedestrian thermal comfort. Therefore, it is essential to develop a dynamic thermal comfort prediction method for pedestrians in cold-region pedestrian spaces based on thermal synesthesia, and to use this method for the overall layout design of such spaces. Summary of the Invention

[0005] The purpose of this invention is to address the problems of insufficient consideration of environmental changes and pedestrian subjective feelings, and low prediction accuracy in existing pedestrian thermal comfort prediction methods. It aims to improve the accuracy of dynamic thermal comfort prediction for pedestrians in cold-region cities, overcome the bottleneck of prediction efficiency in existing methods, strengthen their supporting role in the early design of pedestrian spaces in cold-region cities, and construct a layout generation and design method for pedestrian spaces in cold-region cities based on dynamic thermal comfort prediction. This invention also addresses the problem of insufficient consideration of dynamic thermal comfort in existing pedestrian space design decisions in cold-region urban and rural areas, and proposes a layout generation and design method for pedestrian spaces in cold-region cities based on dynamic thermal comfort prediction.

[0006] This invention is achieved through the following technical solution: This invention proposes a method for generating and designing pedestrian space layouts in cold-region cities based on dynamic thermal comfort prediction. The method includes the following steps:

[0007] S1. Construction of a spatial pathfinding agent model for cold-region walking;

[0008] Step S1 specifically involves:

[0009] S1.1: Data collection of pedestrian trajectory during different travel periods in a typical cold-region walking space based on IoT;

[0010] S1.2: Construction of gray box model for wayfinding in cold-region pedestrian spaces during different travel periods;

[0011] S1.3: Accurate pedestrian trajectory extraction for different travel periods in typical cold-region walking spaces based on UAV imagery data;

[0012] S1.4: Correction of cold-region walking spatial pathfinding agent model based on typical spatial IoT sensing data;

[0013] S2, Construction of thermal environment data of cold-region pedestrian walking space and dynamic thermal comfort mapping of pedestrians;

[0014] Step S2 specifically involves:

[0015] S2.1: Acquisition of pedestrian thermal sensation in typical cold-region walking spaces based on instantaneous ecological assessment;

[0016] S2.2: Construction of a pedestrian thermal perception model under winter and summer thermal environments in cold-region walking spaces;

[0017] S2.3: Reinforcement Learning-Based Correction of Pedestrian Thermal Perception Model in Cold Regions During Winter and Summer;

[0018] S3, Cold Region Walking Space Layout Generation Design Driven by Dynamic Thermal Comfort Data;

[0019] Step S3 specifically involves:

[0020] S3.1: Obtaining cold-region pedestrian spatial layout schemes driven by generation rules;

[0021] S3.2: Optimization design of cold-region walking space layout driven by dynamic thermal comfort data;

[0022] S3.3: Decision support for the layout design of pedestrian spaces in cold regions based on human-computer interaction.

[0023] Further, step S2.1 specifically includes:

[0024] S2.1.1: Obtaining pedestrian locations in cold-region walking spaces;

[0025] S2.1.2: Real-time thermal sensing data acquisition of pedestrians under changes in pedestrian position in cold-region walking spaces;

[0026] S2.1.3: Real-time acquisition of thermal environment data for typical cold-region walking spaces.

[0027] Furthermore, in step S2.1.2, real-time skin temperature, heart rate, and EDA skin electrical signal data of pedestrians under changes in position in cold-weather walking space are collected based on wearable devices, and subjective evaluation data of pedestrians under changes in position in cold-weather walking space are obtained. Based on this, a mapping relationship between pedestrian physiological index data and thermal sensation in cold-weather walking space is constructed, thereby obtaining real-time thermal sensation information of pedestrians under changes in position in cold-weather walking space.

[0028] Further, in step S3.1, the layout model data of the cold region street is obtained and voxelized. The features of its layout and buildings are extracted by applying a three-dimensional convolutional neural network. Based on the scale features of the cold region street, the predefined cold region pedestrian space is divided into three-dimensional units. Based on the design conditions of the cold region pedestrian space and the extracted features, the three-dimensional matrix is ​​constrained to ensure that the generated scheme meets the geometric and topological requirements. Under the constraints, the three-dimensional units are allocated based on the multi-agent system to realize the cold region pedestrian space layout generation, thereby obtaining the cold region pedestrian space layout optimization prototype.

[0029] Further, step S3.2 specifically includes:

[0030] S3.2.1: Construction of an optimization model for pedestrian spatial layout in cold regions;

[0031] S3.2.2: Acquisition of pedestrian dynamic thermal comfort during travel periods under the pedestrian spatial layout in cold regions;

[0032] S3.2.3: Optimization of pedestrian space layout in cold regions under the guidance of dynamic thermal comfort.

[0033] Further, step S3.2.2 specifically includes:

[0034] S3.2.2.1: Construction of a predictive model for typical thermal environments during different travel periods in winter and summer under the spatial layout of pedestrian walkways in cold regions;

[0035] S3.2.2.2: Thermal environment acquisition of typical pedestrian routes during winter and summer travel periods in cold-region pedestrian spaces;

[0036] S3.2.2.3: Acquisition of pedestrian thermal sensation during typical route travel times under the pedestrian spatial layout in cold regions.

[0037] Further, in step S3.2.2.1, thermal environment images of winter and summer travel periods under the cold-region walking spatial layout are obtained, and the thermal environment images of each travel period are clustered respectively. Then, the mapping relationship between the cold-region walking spatial layout and the clustered images of winter and summer travel periods is constructed.

[0038] Furthermore, in step S3.3, after obtaining the optimal layout scheme for cold-region walking spaces under dynamic thermal comfort guidance, a decision model for cold-region walking spaces is constructed using a random forest model based on the dynamic thermal comfort data of the layout scheme and the morphological evaluation data of the designer on the layout scheme, to obtain a preliminary decision scheme; VR devices and environmental control devices are used to provide users with a walking experience simulation of the scheme, to obtain their evaluation of the scheme, and to use this as feedback to the decision model, so as to further adjust the decision model for cold-region walking spaces, thereby providing designers with efficient and comprehensive decision support.

[0039] The present invention also proposes an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method for generating and designing pedestrian space layout in cold-region cities based on dynamic thermal comfort prediction.

[0040] The present invention also proposes a computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the steps of the method for generating and designing pedestrian space layouts in cold-region cities based on dynamic thermal comfort prediction.

[0041] The beneficial effects of this invention are:

[0042] This invention provides a method for generating and designing cold-region pedestrian space layouts based on dynamic thermal comfort prediction. Based on more realistic pedestrian thermal perception prediction results, it offers designers efficient design decision support for cold-region pedestrian space layouts. This invention constructs a pathfinding agent model for cold-region pedestrian spaces based on big data and IoT data to obtain pedestrian trajectories at different travel times under different layouts. Based on pedestrian physiological index data, it constructs a mapping relationship between pedestrian thermal perception and the thermal environment of cold-region pedestrian spaces to obtain pedestrian thermal perception under different thermal environment changes. Based on the predicted pedestrian thermal perception results along typical pedestrian routes at different times, it optimizes the cold-region pedestrian space layout. Based on machine learning models and designer feedback on visualized solutions, it obtains a cold-region pedestrian space layout design decision model guided by designer decision preferences. For pedestrian thermal perception prediction in cold-region pedestrian spaces, this invention is more realistic and targeted than existing technologies, while also being highly efficient. For cold-region pedestrian space layout decisions, this invention, compared to existing technologies, better considers designer preferences while ensuring decision stability, thus being more conducive to application in practical engineering scenarios. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0044] Figure 1 This is a flowchart of the method for generating and designing pedestrian space layouts in cold-region cities based on dynamic thermal comfort prediction, as described in this invention. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Combination Figure 1 This invention proposes a method for generating and designing pedestrian space layouts in cold-region cities based on dynamic thermal comfort prediction. The method includes the following steps:

[0047] S1. Construction of a spatial pathfinding agent model for cold-region walking;

[0048] Step S1 specifically involves:

[0049] S1.1: Collection of pedestrian trajectory data for different travel periods in typical cold-region walking spaces based on IoT; In step S1.1, select a portion of typical cold-region spaces, deploy multiple Wi-Fi and Bluetooth access points or beacons in the target area, record the MAC address and signal strength of each device at different travel periods, and use triangulation and data fusion technology to accurately estimate the location and trajectory of each device, so as to realize pedestrian positioning and trajectory tracking at different travel periods in typical cold-region walking spaces;

[0050] S1.2: Construction of a gray box model for pathfinding in cold-region pedestrian spaces during different travel periods; In step S1.2, pedestrian trajectory information is transformed into pedestrian path graph structure data. Based on this, a graph convolutional neural network model is applied to construct a pedestrian path selection model, achieving efficient prediction of pedestrian path node selection; On this basis, a social force model, a multi-agent system, and a particle filter algorithm are integrated to construct a pedestrian pathfinding gray box model. The model parameters are optimized and adjusted based on pedestrian trajectory data, integrating mathematical models and data-driven methods to achieve prediction of specific trajectories between pedestrian nodes;

[0051] S1.3: Extracting precise pedestrian trajectories for each travel time in typical cold-region walking spaces based on UAV imagery data; In step S1.3, multi-angle imagery data of typical cold-region walking spaces for each time period are acquired based on UAV low-altitude photography. A high-precision 3D information model of the cold-region street is constructed based on the multi-angle images, and the positions of pedestrians in typical frames are marked in the high-precision 3D information model. At the same time, the YOLO algorithm is applied to perform target detection in continuous UAV image frames, and the CSRT algorithm is applied to track each detected pedestrian separately to obtain the changes in the position of each pedestrian. The pedestrian positions in typical frames and the information on changes in pedestrian positions are integrated to reconstruct the pedestrian trajectory with high precision.

[0052] S1.4: Correction of cold-region walking spatial pathfinding agent model based on typical spatial IoT sensing data;

[0053] S2, Construction of thermal environment data of cold-region pedestrian walking space and dynamic thermal comfort mapping of pedestrians;

[0054] Step S2 specifically involves:

[0055] S2.1: Acquisition of pedestrian thermal sensation in typical cold-region walking spaces based on instantaneous ecological assessment; Step S2.1 specifically includes:

[0056] S2.1.1: Obtaining pedestrian locations in cold-region walking spaces;

[0057] S2.1.2: Real-time thermal sensation acquisition of pedestrians under changes in pedestrian position in cold-weather walking space; In step S2.1.2, real-time skin temperature, heart rate and EDA skin conductance signal data of pedestrians under changes in pedestrian position in cold-weather walking space are collected based on wearable devices, and subjective evaluation data of pedestrians under changes in pedestrian position in cold-weather walking space are obtained. Based on this, a mapping relationship between pedestrian physiological index data and thermal sensation in cold-weather walking space is constructed, thereby obtaining real-time thermal sensation information of pedestrians under changes in pedestrian position in cold-weather walking space;

[0058] S2.1.3: Real-time acquisition of thermal environment data for typical cold-region walking spaces;

[0059] S2.2: Construction of a pedestrian thermal perception model under winter and summer thermal environments in cold-region walking spaces;

[0060] S2.3: Reinforcement Learning-Based Correction of Pedestrian Thermal Perception Model in Cold Regions During Winter and Summer;

[0061] S3, Cold Region Walking Space Layout Generation Design Driven by Dynamic Thermal Comfort Data;

[0062] Step S3 specifically involves:

[0063] S3.1: Obtaining the cold-region pedestrian space layout scheme driven by generation rules; In step S3.1, the cold-region street layout model data is obtained and voxelized, and the layout and building features are extracted by applying a three-dimensional convolutional neural network; Based on the scale features of the cold-region street, the predefined cold-region pedestrian space is divided into three-dimensional units, and the three-dimensional matrix is ​​constrained based on the cold-region pedestrian space design conditions and the extracted features to make its generation scheme meet geometric and topological requirements. Under the constraints, the three-dimensional units are allocated based on the multi-agent system to realize the cold-region pedestrian space layout generation, thereby obtaining the cold-region pedestrian space layout optimization prototype;

[0064] S3.2: Optimization design of cold-region walking space layout driven by dynamic thermal comfort data; Step S3.2 specifically includes:

[0065] S3.2.1: Construction of an optimization model for pedestrian spatial layout in cold regions;

[0066] S3.2.2: Acquisition of pedestrian dynamic thermal comfort during travel periods under the cold-region walking space layout; Step S3.2.2 specifically includes:

[0067] S3.2.2.1: Construction of typical thermal environment prediction model for winter and summer travel periods under the cold-region walking spatial layout; In step S3.2.2.1, thermal environment images of winter and summer travel periods under the cold-region walking spatial layout are obtained, and the thermal environment images of each travel period are clustered respectively. Then, the mapping relationship between the cold-region walking spatial layout and the clustered images of winter and summer travel periods is constructed.

[0068] S3.2.2.2: Thermal environment acquisition of typical pedestrian routes during winter and summer travel periods in cold-region pedestrian spaces;

[0069] S3.2.2.3: Acquisition of pedestrian thermal sensation during typical travel times on pedestrian routes under the cold-region pedestrian spatial layout;

[0070] S3.2.3: Optimization of pedestrian space layout in cold regions under the guidance of dynamic thermal comfort;

[0071] S3.3: Human-Computer Interaction-Based Decision Support for Cold-Region Pedestrian Space Layout Design. In step S3.3, after obtaining the optimal cold-region pedestrian space layout scheme under dynamic thermal comfort guidance, a decision model for the cold-region pedestrian space layout is constructed using a random forest model based on the dynamic thermal comfort data of the layout scheme and the designer's morphological evaluation data of the layout scheme, to obtain a preliminary decision scheme. VR devices and environmental control devices are used to provide users with a simulated walking experience of the scheme, to obtain their evaluation of the scheme, and to use this as feedback to the decision model for further adjustment, thereby providing designers with efficient and comprehensive decision support.

[0072] Example

[0073] according to Figure 1 As shown, this invention provides a method for generating and designing the spatial layout of pedestrian areas in cold regions based on dynamic thermal comfort prediction, comprising the following steps:

[0074] Step 1: Construction of a spatial pathfinding agent model for cold-region walking;

[0075] Step 1 specifically involves:

[0076] Step 1.1: Data collection of pedestrian trajectory during different travel periods in a typical cold-region walking space based on IoT;

[0077] Step 1.2: Construction of gray box model for wayfinding in cold-region pedestrian spaces during different travel periods;

[0078] Step 1.3: Extracting precise pedestrian trajectories for different travel periods in typical cold-region walking spaces based on UAV imagery data;

[0079] Step 1.4: Correction of the cold-region walking space pathfinding agent model based on accurate pedestrian trajectory information. Using accurate pedestrian trajectory data of each travel time in a typical cold-region walking space as the fine-tuning dataset, the pathfinding gray box model of each travel time in the cold-region walking space is further trained, and the model parameters are gradually adjusted and optimized.

[0080] Step 1.1 specifically includes:

[0081] A representative cold-weather walking area was selected as the study area. Multiple Wi-Fi and Bluetooth access points or beacons were deployed within the target area to ensure signal coverage throughout the region. The spacing between access points should be reasonably arranged based on environmental conditions and positioning accuracy requirements. During each travel period, the MAC addresses and signal strengths of all devices entering the target area were recorded. The Wi-Fi and Bluetooth access points periodically scanned the surrounding environment, capturing device MAC addresses and received signal strength (RSSI). The collected MAC address and signal strength data were stored in a database. Each record should include a timestamp, access point ID, MAC address, and signal strength information for subsequent data processing and analysis.

[0082] Using triangulation, the location of the device is calculated based on the signal strength received from the same device at different access points. First, based on the received signal strength indication, the signal strength is converted into distance using a path loss model. Then, based on the calculated distance and the location of the access points, the device's location is estimated using triangulation. Because signals are affected by unpredictable and varied obstacles (such as walls and furniture) in cold-weather walking spaces, the free-space path loss model and the ITU indoor path loss model are not used here. Instead, a more flexible empirical model is employed, with the calculation formula as follows:

[0083] RSSI = RSSI0 - 10nlog 10 (d)

[0084] Here, RSSI stands for Received Signal Strength Indicator, measured in dBm. RSSI0 is the signal strength at a reference distance d0, typically measured at 1 meter. d is the distance between the receiver and transmitter, measured in meters (m). n is the path loss index, which depends on the environment and is generally between 2 and 4.

[0085] Step 1.2 specifically includes:

[0086] First, the collected pedestrian trajectory data is preprocessed, including noise removal, missing data filling, and standardization. Next, the preprocessed trajectory data is transformed into a path graph structure, where nodes are defined as key locations in the pedestrian space, edges represent pedestrian paths, and each edge is assigned a corresponding weight (such as pedestrian flow, movement time, etc.).

[0087] Based on the transformed path graph structure data, a pedestrian path selection model is constructed using a Graph Convolutional Neural Network (GCN) model. First, a GCN model is designed and trained, using the path graph structure data as input to predict the probability of a pedestrian selecting nodes in the path graph. The model training process includes feature extraction, graph convolution operations, and node classification. Through multiple iterative optimizations, the model's prediction accuracy and robustness are improved.

[0088] Building upon this foundation, a gray-box model for pedestrian pathfinding is constructed by integrating a social force model, a multi-agent system, and a particle filter algorithm. The social force model simulates the interaction forces among pedestrians and the influence of the environment on pedestrians; the multi-agent system simulates the collective behavior and path selection strategies of pedestrians; and the particle filter algorithm is used for real-time tracking and prediction of pedestrian trajectories.

[0089] In the construction of the pathfinding gray box model, each agent is first assigned an initial position and velocity to simulate the initial state of the pedestrian crowd. These initial states reflect the pedestrian's position and direction of movement at a specific time point. Next, a set of particles is initialized for each agent, representing possible trajectory states. A particle filtering algorithm uses these particles to represent the pedestrian's possible position and path in future time steps.

[0090] After initialization, the trained GCN model is used to predict the next target node on the pedestrian path, providing each agent with a clear short-term objective and direction for subsequent motion calculations. Then, the motion of each agent at the current time step is calculated based on the social force model. The social force model calculates the pedestrian's acceleration and velocity changes by considering the interaction forces between pedestrians, the attractive forces between pedestrians and targets, and the repulsive forces between pedestrians and obstacles, thereby updating the pedestrian's position.

[0091] Building upon this foundation, a particle filter algorithm is applied to predict and track the trajectory of the intelligent agent. The particle filter uses a social force model to predict the path of each particle while adjusting its weight based on actual observation data. By adjusting the weights, the particle filter improves the accuracy of estimating the pedestrian's true trajectory. The resampling step, by selecting particles with high weights, reduces impossible trajectories, thereby enhancing the reliability of trajectory prediction.

[0092] Based on the results of the particle filtering algorithm, the state of each agent is updated. This process ensures that each agent's trajectory not only accurately reflects its own motion patterns but also aligns with the actual situation of interaction with the environment and other pedestrians. Through this dynamic adjustment, the motion trajectory of each agent becomes more precise and reliable.

[0093] Finally, through iterative steps, the agent's trajectory is gradually optimized and accurately predicted. At each time step, the agent's trajectory is adjusted and optimized based on new observation data and the prediction results of the particle filter algorithm, until predictions for all time steps are completed, resulting in the final gray-box model for pedestrian pathfinding.

[0094] Step 1.3 specifically includes:

[0095] Deploy drones within typical cold-region pedestrian areas, and plan drone flight paths based on the terrain and layout of the target area to ensure coverage of the entire pedestrian space. Drones should conduct multiple low-altitude flights during various time periods (including 08:00-09:00, 09:00-10:00, 10:00-11:00, 11:00-12:00, 12:00-13:00, 13:00-14:00, 14:00-15:00, 15:00-16:00, 16:00-17:00, and 17:00-18:00) to collect multi-angle image data. Use high-resolution cameras for low-altitude photography to ensure image clarity and coverage, avoiding obstruction and ghosting. Flight paths can be planned using Altizure software. After planning the survey area, altitude, forward overlap, lateral overlap, and camera tilt angle, the software will automatically calculate the shooting time.

[0096] Next, the multi-angle image data captured by the UAV is preprocessed, including noise removal and color correction. Then, it is imported into ContextCapture Master software, where the coordinates of corresponding points in the UAV-acquired images are matched against multiple data sources. By setting the coordinates of connection points and control points, a correction equation is established to improve image matching accuracy. The modeling process automatically matches the best images from different angles at each location and uses the SMF algorithm for feature point matching, converting the multi-angle image data into point cloud data. After segmenting the dense point cloud data obtained from aerial triangulation, different levels of TIN triangulation are generated to transform the point cloud data into irregular triangular meshes. The triangular mesh model is simplified, and then automatically textured to generate a high-precision 3D information model of the cold-region street, ensuring the model's accuracy and detail, and clearly displaying information such as terrain, buildings, and roads in the pedestrian space.

[0097] Based on this, several typical frames are selected from the image data captured by the drone. These frames should represent the pedestrian distribution in the pedestrian space over different time periods. Then, the position of each pedestrian is located in the selected typical frames, and this position data is recorded in a high-precision 3D information model.

[0098] Next, the YOLO algorithm is trained using a publicly available pedestrian detection dataset or a custom dataset to accurately detect pedestrian targets in the imagery. The trained YOLO model is then applied to consecutive UAV image frames to automatically detect pedestrian targets and record the position and time information of each pedestrian. At the location of each detected pedestrian target, a CSRT (Discriminative Correlation Filter with Channel and Spatial Reliability) tracker is initialized. The CSRT algorithm is then applied to track each detected pedestrian in consecutive UAV image frames, recording the positional changes of each pedestrian in each time frame.

[0099] Finally, the manually labeled pedestrian positions in typical frames are integrated with the pedestrian position change information obtained through the CSRT algorithm, and these data are fused into a high-precision 3D information model. Based on the integrated pedestrian position data, a trajectory reconstruction algorithm is used to reconstruct the pedestrian's movement trajectory with high precision, ensuring that the reconstructed trajectory accurately reflects the pedestrian's movement path and behavior pattern in the cold-region walking space. The reconstructed pedestrian trajectory is then visualized in the high-precision 3D information model using a GIS platform or 3D visualization tools, providing intuitive pedestrian behavior analysis results.

[0100] Step 2: Constructing a mapping between thermal environment data of cold-region walking spaces and pedestrian thermal sensation;

[0101] Step 2 specifically involves:

[0102] Step 2.1: Acquisition of pedestrian thermal sensation in typical cold-region walking spaces based on instantaneous ecological assessment;

[0103] Step 2.2: Construction of pedestrian thermal perception model under winter and summer thermal environment conditions in cold-region walking spaces;

[0104] Step 2.3: Correction of winter and summer pedestrian thermal perception model in cold-region walking space based on reinforcement learning.

[0105] Step 2.1 specifically involves:

[0106] Step 2.1.1: Based on the mobile devices (such as smartphones) carried by pedestrians, record the geographical location of pedestrians in real time to obtain the location of pedestrians in cold-region walking spaces;

[0107] Step 2.1.2: Real-time thermal sensing data acquisition of pedestrians under changes in pedestrian position in cold-region walking spaces;

[0108] Step 2.1.3: Deploy temperature, humidity, wind speed and radiative heat sensors in a grid pattern in a typical cold-region walking space to collect thermal environment data of the typical cold-region walking space in real time.

[0109] Step 2.1.2 specifically includes:

[0110] Based on wearable devices with sensors, including smart wearable ear clip sensors, smart wearable wrist sensors, and smart wearable finger sensors, real-time skin temperature, heart rate, and EDA skin conductance data are collected as pedestrians change position in cold-weather walking spaces, and this data is transmitted to a central data processing server via wireless communication (such as Bluetooth or Wi-Fi).

[0111] Meanwhile, during pedestrian walks, subjective evaluation data on current thermal sensations will be collected periodically through questionnaires, including subjective indicators such as thermal acceptability, thermal comfort, temperature perception, and thermal pleasure. This subjective evaluation data will be matched with physiological indicator data in time and space to form a complete dataset.

[0112] Subsequently, big data analytics and machine learning techniques were used to construct a mapping relationship between pedestrian physiological indicators and thermal sensation in cold-region walking spaces. First, the collected data was cleaned and normalized to eliminate noise and outliers. Then, key features, such as skin temperature change rate, heart rate fluctuation patterns, and EDA signal changes, were extracted as input variables for the model. Next, a neural network model was used to train the mapping model to predict pedestrian thermal sensation. After model validation and optimization, real-time thermal sensation prediction under pedestrian position changes in cold-region walking spaces was achieved.

[0113] Step 3: Generation and design of cold-region walking space layout driven by dynamic thermal comfort data;

[0114] Step 3 specifically involves:

[0115] Step 3.1: Obtaining a rule-driven spatial layout scheme for cold-region pedestrian areas;

[0116] Step 3.2: Optimization design of cold-region walking space layout driven by dynamic thermal comfort data;

[0117] Step 3.3: Decision support for the layout design of pedestrian spaces in cold regions based on human-computer interaction.

[0118] Step 3.1 specifically involves:

[0119] First, a large amount of cold-region street layout model data was acquired. This data was then voxelized, meaning the 3D model was divided into a series of regular small cubic units. Next, a 3D convolutional neural network was applied to process the voxelized data to extract the layout features and building features of the cold-region streets.

[0120] Based on the extracted scale features of cold-region blocks, the predefined cold-region pedestrian space is divided into three-dimensional units of appropriate size. The size of these three-dimensional units should be set according to the actual design conditions and spatial requirements to ensure that each unit can accurately reflect the details of the pedestrian space. Then, constraints are set on the three-dimensional matrix according to the design conditions of the cold-region pedestrian space and the extracted layout features. These constraints include geometric and topological requirements, such as building scale restrictions, road width requirements, and the adjacency of blocks and roads.

[0121] Under constraints, a multi-agent system is used to allocate three-dimensional units. The system simulates the combination patterns of these units, with each agent representing a unit. Specific rules and decision-making mechanisms are used for combination and layout optimization. The agents interact with each other based on design and constraints, and through an iterative process, gradually adjust the position and distribution of each three-dimensional unit to generate a pedestrian space layout scheme that meets the design objectives.

[0122] Step 3.2 specifically involves:

[0123] Step 3.2.1: Construct a mathematical model and a parametric model for optimizing the spatial layout of pedestrian areas in cold regions based on design objectives and constraints;

[0124] Step 3.2.2: Acquisition of pedestrian dynamic thermal comfort during travel periods under the pedestrian spatial layout in cold regions;

[0125] Step 3.2.3: Based on the pedestrian dynamic thermal comfort objective, apply a multi-objective optimization algorithm to optimize the spatial layout of pedestrian walkways in cold regions and obtain the Palento solution.

[0126] Step 3.2.2 specifically involves:

[0127] Step 3.2.2.1: Construct a predictive model of typical thermal environment during different travel periods in winter and summer under the spatial layout of cold-region walking;

[0128] Step 3.2.2.2: Using the typical thermal environment prediction model for each travel period in winter and summer constructed in Step 3.2.1, obtain the typical thermal environment prediction images for each travel period in winter and summer. Using the cold-region walking space pathfinding agent model constructed in Step 1, obtain the pedestrian routes for each travel period under the current cold-region walking space layout, so as to realize the prediction of the thermal environment under the typical pedestrian routes for each travel period in winter and summer in the cold-region walking space.

[0129] Step 3.2.2.3: Based on the thermal environment of pedestrians during winter and summer travel periods in the cold-region pedestrian space as described in Step 3.2.2, obtain the pedestrian thermal sensation during typical travel periods under the cold-region pedestrian space layout by using the mapping model between the thermal environment data of the cold-region pedestrian space and pedestrian thermal sensation constructed in Step 2.

[0130] Step 3.2.2.1 specifically includes:

[0131] Microclimate simulations were performed on 1000 typical cold-region walking spaces using ENVI-MET software to obtain thermal environment images for different travel periods in winter and summer. The images were preprocessed, converting from RGB color space to HSV (hue, saturation, brightness) color space to better separate color and brightness information. Then, color histograms were calculated, statistically analyzing the color distribution of each of the three HSV channels, obtaining the color histogram for each channel, and normalizing it to ensure values ​​were between 0 and 1. To capture color distribution patterns, the images were divided into grids or blocks, and the color histogram for each block was calculated and combined into a feature vector. This method reflects the spatial distribution of colors in the image. By extracting and combining the color histograms of each grid block, a complete color distribution feature vector was generated. After feature extraction, K-means clustering was used to cluster the extracted color distribution feature vectors, obtaining thermal environment clustering images for each of the 1000 typical cold-region walking spaces during different travel periods in winter and summer. Finally, a mapping relationship between the spatial layout of cold-region walking and clustered images of different travel periods in winter and summer is constructed using a convolutional neural network.

[0132] Step 3.3 specifically involves:

[0133] First, 200 designers were invited to evaluate and score the morphology of 1000 cold-region walking space designs, and dynamic thermal comfort data for each design was obtained. Next, the collected data was cleaned to remove noise and outliers. The dynamic thermal comfort data and morphological evaluation data were then integrated into a comprehensive dataset, ready to be input into a random forest model.

[0134] Next, key features were extracted from the comprehensive dataset. Statistical and correlation analysis methods were used to select features that had a significant impact on layout decisions, thereby improving the model's training efficiency and prediction accuracy. A random forest algorithm was then used to train the processed data, constructing a preliminary layout decision model. Cross-validation was employed to evaluate the model's performance, ensuring it possesses good generalization ability and prediction accuracy.

[0135] Next, VR devices and environmental control equipment were used to provide users with a simulated walking experience of the proposed solution. High-performance VR headsets and controllers were selected to ensure an immersive walking experience. Environmental control equipment, such as air conditioners and humidifiers, was configured to control ambient temperature, humidity, and wind speed, simulating real-world climate conditions and environmental changes. A virtual scene of the cold-region walking space layout scheme was constructed using 3D modeling software, including elements such as buildings, roads, and green spaces. Users were then invited to wear VR devices and experience a virtual walk. During this process, environmental parameters in the virtual scene and controllable environment were set based on pedestrian dynamic thermal comfort data to simulate a realistic thermal comfort experience. Feedback on the scheme, including comfort, aesthetics, and functionality, was recorded during and after the experience and compiled into a structured evaluation dataset.

[0136] Finally, using the collected user evaluation data, the machine learning model was optimized and trained, and the parameters and optimization strategies of the layout decision model were adjusted to gradually improve the model's decision accuracy and matching degree. Based on the optimized cold-region pedestrian space decision model, the optimal cold-region pedestrian space layout design scheme was obtained, providing designers with efficient and comprehensive decision support.

[0137] The present invention also proposes an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method for generating and designing pedestrian space layout in cold-region cities based on dynamic thermal comfort prediction.

[0138] The present invention also proposes a computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the steps of the method for generating and designing pedestrian space layouts in cold-region cities based on dynamic thermal comfort prediction.

[0139] The memory in this application embodiment can be volatile memory or non-volatile memory, or it can include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct rambus RAM (DRRAM). It should be noted that the memory used in the methods described in this invention is intended to include, but is not limited to, these and any other suitable types of memory.

[0140] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., high-density digital video discs (DVDs)), or semiconductor media (e.g., solid-state disks (SSDs)).

[0141] During implementation, each step of the above method can be completed by an integrated logic circuit of the hardware in the processor or by instructions in the form of software. The steps of the method disclosed in conjunction with the embodiments of the present application can be directly embodied as being executed by a hardware processor, or can be executed by a combination of hardware and software modules in the processor. The software module can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in conjunction with its hardware. To avoid repetition, it will not be described in detail here.

[0142] It should be noted that the processor in the embodiments of the present application can be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method embodiment can be completed by an integrated logic circuit of the hardware in the processor or by instructions in the form of software. The above processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. The various methods, steps, and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiments of the present application can be directly embodied as being executed by a hardware decoding processor, or can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium mature in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, registers, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0143] The foregoing has provided a detailed description of the method for generating and designing pedestrian space layouts in cold-region cities based on dynamic thermal comfort prediction proposed in this invention. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A method for generating and designing pedestrian spatial layouts in cold-region cities based on dynamic thermal comfort prediction, characterized in that, The method includes the following steps: S1. Construction of a spatial pathfinding agent model for cold-region walking; Step S1 specifically involves: S1.1: Data collection of pedestrian trajectory during different travel periods in a typical cold-region walking space based on IoT; S1.2: Construction of gray box model for wayfinding in cold-region pedestrian spaces during different travel periods; S1.3: Accurate pedestrian trajectory extraction for different travel periods in typical cold-region walking spaces based on UAV imagery data; S1.4: Correction of cold-region walking spatial pathfinding agent model based on typical spatial IoT sensing data; S2, Construction of thermal environment data of cold-region pedestrian walking space and dynamic thermal comfort mapping of pedestrians; Step S2 specifically involves: S2.1: Acquisition of pedestrian thermal sensation in typical cold-region walking spaces based on instantaneous ecological assessment; S2.2: Construction of a pedestrian thermal perception model under winter and summer thermal environments in cold-region walking spaces; S2.3: Reinforcement Learning-Based Correction of Pedestrian Thermal Perception Model in Cold Regions During Winter and Summer; S3, Cold Region Walking Space Layout Generation Design Driven by Dynamic Thermal Comfort Data; Step S3 specifically involves: S3.1: Obtaining cold-region pedestrian spatial layout schemes driven by generation rules; In step S3.1, the layout model data of the cold region street is obtained and voxelized. The layout and building features are extracted by applying a three-dimensional convolutional neural network. Based on the scale features of the cold region street, the predefined cold region pedestrian space is divided into three-dimensional units. Based on the design conditions of the cold region pedestrian space and the extracted features, the three-dimensional matrix is ​​constrained to ensure that the generated scheme meets the geometric and topological requirements. Under the constraints, the three-dimensional units are allocated based on the multi-agent system to realize the cold region pedestrian space layout generation, thereby obtaining the cold region pedestrian space layout optimization prototype. S3.2: Optimization design of cold-region walking space layout driven by dynamic thermal comfort data; Step S3.2 specifically includes: S3.2.1: Construction of an optimization model for pedestrian spatial layout in cold regions; S3.2.2: Acquisition of pedestrian dynamic thermal comfort during travel periods under the pedestrian spatial layout in cold regions; S3.2.3: Optimization of pedestrian space layout in cold regions under the guidance of dynamic thermal comfort; S3.3: Decision support for the layout design of pedestrian spaces in cold regions based on human-computer interaction; In step S3.3, after obtaining the optimal layout scheme for cold-region walking spaces under dynamic thermal comfort guidance, a decision model for cold-region walking spaces is constructed using a random forest model based on the dynamic thermal comfort data of the layout scheme and the morphological evaluation data of the designer on the layout scheme, and a preliminary decision scheme is obtained. VR devices and environmental control devices are used to provide users with a walking experience simulation of the scheme, obtain their evaluation of the scheme, and use it as feedback to the decision model to further adjust the decision model for cold-region walking spaces, thereby providing designers with efficient and comprehensive decision support.

2. The method according to claim 1, characterized in that, Step S2.1 specifically includes: S2.1.1: Obtaining pedestrian locations in cold-region walking spaces; S2.1.2: Real-time thermal sensing data acquisition of pedestrians under changes in pedestrian position in cold-region walking spaces; S2.1.3: Real-time acquisition of thermal environment data for typical cold-region walking spaces.

3. The method according to claim 2, characterized in that, In step S2.1.2, real-time skin temperature, heart rate and EDA skin electrical signal data of pedestrians under changes in position in cold-weather walking space are collected based on wearable devices, and subjective evaluation data of pedestrians under changes in position in cold-weather walking space are obtained. Based on this, a mapping relationship between pedestrian physiological index data and thermal sensation in cold-weather walking space is constructed, thereby obtaining real-time thermal sensation information of pedestrians under changes in position in cold-weather walking space.

4. The method according to claim 1, characterized in that, Step S3.2.2 specifically includes: S3.2.2.1: Construction of a predictive model for typical thermal environments during different travel periods in winter and summer under the spatial layout of pedestrian walkways in cold regions; S3.2.2.2: Thermal environment acquisition of typical pedestrian routes during winter and summer travel periods in cold-region pedestrian spaces; S3.2.2.3: Acquisition of pedestrian thermal sensation during typical route travel times under the pedestrian spatial layout in cold regions.

5. The method according to claim 4, characterized in that, In step S3.2.2.1, thermal environment images of winter and summer travel periods under the cold-region walking spatial layout are obtained, and the thermal environment images of each travel period are clustered respectively. Then, the mapping relationship between the cold-region walking spatial layout and the clustered images of winter and summer travel periods is constructed.

6. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1-5.

7. A computer-readable storage medium for storing computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the steps of the method according to any one of claims 1-5.

Images