Research On Site Selection Of Charging Stations For Electric Taxi Based On Spatio Temporal Trajectory Data

In the process of the continuous promotion of carbon peaking and carbon neutrality goals,the overall electrification of vehicles in the urban public domain has been widely carried out,and taxis have become the main object of promotion after the electrification of buses.This paper focuses on the practical needs of the urgent need to improve the charging infrastructure service level and guarantee capacity of electric taxis,and takes the vehicle history space-time trajectory data mining as the main way.To explore the location scheme of electric taxi charging station driven by real operation space-time trajectory data,in order to provide reference for alleviating the charging problem of urban electric taxi and improving the profit efficiency of charging station.This paper takes the trajectory data processing flow as a reference,for the purpose of charging station location to clean the irregular data in the original track,and uses the map matching method based on road geometry to correct the position offset track points to improve the data quality.In order to better mine the charging demand contained in the space-time trajectory,this paper uses the time,space and speed attributes contained in the trajectory to capture the corresponding trajectory point of the taxi stopping behavior that meets the charging characteristics;in order to make up for the lack of the stopping point which is difficult to describe the charging demand in the taxi operation pr ocess,the drop-off point data contained in the OmurD order is extracted as the potential charging demand location.Secondly,the dense area in the stay data is obtained as the pre-selected address by the spatio-temporal density clustering algorithm ST-DBSCAN.In the process of landing point clustering,the reachable distance sequence generated by OPTICS is used to obtain the appropriate K value and filter the noise point data,and then the centroid is extracted as the pre-selected location by using Kmurmeanski+algorithm.Considering the large scale and uneven density of the landing point data,the density partition is carried out in the clustering process,and the corresponding parameters are set for different density regions to improve the clustering effect of the clustering algorithm in the case of large difference in data density.reduce charging supply blanks.In order to avoid the overlap of service range caused by the dense pre-location distribution of electric taxi charging stations,this paper establishes a location optimization model based on POI data and sub-regional service radius to maximize coverage demand and traffic influence,and uses immune optimization algorithm to solve the problem,and a relatively reasonable layout of electric taxi charging stations is obtained.In order to test the rationality and feasibility of the location method proposed in this paper,an example analysis is carried out based on the open spatio-temporal trajectory data set of taxis in Shenzhen.The experimental results show that the location scheme in this paper accords with the operation law of electric taxis.It can cover more charging requirements represented by parking points,thus verifying the effectiveness of the model.This research work is helpful to broaden the application field of spatio-temporal trajectory data in urban construction and provide some auxiliary decision support for the location selection of electric taxi charging stations.