| With the increasing attraction of shopping,catering,entertainment and other functions to pedestrians in the urban pedestrian-only space,the excessive concentration of pedestrians has become more and more serious,which has caused a series of serious problems such as poor experience of pedestrian leisure activities and affected personal safety.At present,human management and guidance are mainly used to solve the problem of excessive pedestrian aggregation.Although this method can prevent safety accidents caused by instantaneous aggregation,it seriously affects the leisure experience of pedestrians and limits the full use of pedestrian-only space functions.With the development of UAV data acquisition technology and walking behaviour modeling theory,the method of passive pedestrian guidance provides the possibility to solve the problem of crowd aggregation while satisfying the full function of pedestrian-only space.The purpose of this study is to quantify the comfort and safety of pedestrians by predicting the pedestrian behaviour and density distribution in the block scale urban pedestrian dedicated space,and then to propose an optimization design strategy for the reasonable distribution of passive guided pedestrians.By analyzing the priority of various path planning behaviour prediction models constructed in this study,a tactic layer path planning model is established;Using neural network training,the nonlinear model of pedestrian behaviour in the operational layer is obtained to predict pedestrian acceleration;Combining the two-level prediction model,a pedestrian density distribution prediction model is proposed;This model is used to simulate the pedestrian density of urban pedestrian-only space under various typical spatial forms,and the corresponding optimization design strategy is proposed.According to the attributes of urban pedestrian-only space,this paper analyzes the constituent elements and spatial form,and summarizes the spatial form types of urban pedestrian-only space.According to the top-down hierarchical logic of walking behaviour,this study analyzes the ways in which each level interacts with the environment in the form of perception,action and short-term memory,and then obtains the influencing factors of walking behaviour at all levels.Pedestrian-only space evaluation takes the density distribution results of a large number of pedestrian walking behaviours as the standard to measure the pedestrian leisure experience and stampede accident risk in different urban pedestrian-only space environments,that is,the comfort and safety issues in this paper.Firstly,this paper analyzes the various elements that affect the pedes trian’s walking behaviour in the urban pedestrian-only space,and summarizes three factors that affect the path planning behaviour: the change of pedestrian passage,the spatial physical environment and the walking energy consumption;Then the three factors are divided into five detailed factors: straight ahead behaviour,turning behaviour,passing obstacle,visual field obstacle and vertical walking facilities;The pedestrian path planning model under the influence of a single factor and the path planning priority problem when multiple influencing factors occur at the same time are studied respectively.It is concluded that the priority of passing obstacle is the highest,followed by visual field obstacle,followed by straight travel behaviour,turning behaviour and vertical walking facilities,as the general rule.Then,a pedestrian tactic layer path planning model is established.This paper summarizes the effects of five factors,including the current movement situation,obstacle avoidance behaviour,collision avoidance behaviour,internal and external relations of the social group,which affect the pedestrian operation layer movement behaviour;Based on the neural network,the univariate prediction ability and relative importance of each influencing factor are quantified.Based on the univariate prediction ability and relative importance,the initial independent variable set is screened in two steps,and the input variable set of the prediction model with moderate dimension is determined;Then the nonlinear prediction model of pedestrian acceleration is obtained by training the neural network model;In this paper,the sub-aim determined by the tactic layer model are taken as a key input variable of the operational layer model,and a continuous prediction model of individual pedestrian walking behaviour and walking streamline is established.By incorporating the heterogeneity of pedestrian groups,a prediction model of pedestrian density distribution in urban pedestrian-only space is established;Finally,based on the pedestrian density distribution prediction model established in this paper,the comfort and safety of pedestrians in urban pedestria n-only space under various typical spatial forms are simulated,and the optimization design strategy is proposed.Through the cross research of architecture and transportation disciplines,this research applies data collection methods such as UAV video sho oting and pedestrian automatic tracking,and achieves the following innovative results: the influencing factors of path planning behaviour in urban pedestrian-only space are extracted,and the pedestrian tactic layer path planning model is established;A comprehensive independent variable set is established to predict the real-time movement behaviour of pedestrians at the operational level,and the method of neural dimension control and nonlinear prediction model building at the input layer of the operational layer walking behaviour prediction model based on BP neural network is proposed;An agent-based model is established to predict pedestrian density distribution in urban pedestrian-only space.This study applies the optimized model to predict the comfort and safety of pedestrians in the typical space form of urban pedestrian-only space.By exploring the relationship between different design methods of urban pedestrian–only space and pedestrian density distribution,the design strategy of optimizing comfort and safety is proposed for the overall environment and local space of urban pedestrian-only space,which provides a basis for architects to design urban pedestrian-only space. |
