Research On Identification And Characteristics Of Retrograde Cycling Of Shared Bicycle Based On Trajectory Data

Under the background of current green transportation development,cycling has become an important part of the comprehensive transportation system.With the rise of shared bicycle,citizens’ enthusiasm towards green and healthy travel has been further enhanced,and cycling has once again become the main way of travel for citizens.However,at the same time,the problem of inappropriate behaviors have become increasingly prominent,and the retrograde cycling is common,which seriously affects the travel experience and safety of cyclists.Therefore,it is necessary to carry out research into the retrograde cycling of shared bicycles,and analyze the characteristics and causes of the behaviors,also propose solutions in a targeted manner,in order to reduce retrograde cycling behaviors and assure the safety of cyclists.Shared bicycle trajectory data creates a new way for retrograde cycling behavior search.Compared with conventional manual counts methods,shared bicycle trajectory data has wide coverage,strong timeliness,and high analysis efficiency,which can achieve comprehensive identification and evaluation of retrograde behavior.Based on the shared bicycle trajectory data,the paper carries out a study of data processing and cleaning methods,road-network matching methods,and retrograde identification methods,also makes use of the actual trajectory data to conduct an in-depth analysis of the characteristics and causes of retrograde cycling in typical zones,and puts forward solutions and suggestions.The main research contents are as follows:(1)Basic situations of shared bicycles,including the development history,user characteristics,and existing problems,are sorted out;the characteristics of shared bicycle riding and the main influencing factors are analyzed;combined with data structure characteristics and application situation of shared bicycle trajectory data,the data availability is confirmed,and the research strategy of data processing is developed.(2)From the perspective of analysis demands,and targeting at uneven quality of shared bicycle trajectory data,the data processing requirements and processing procedures are built;based on boxplot and spatial relationship,abnormal and mutational points are processed,based on the sliding window technique,the parking points are identified and cleaned.(3)Shared bicycle trajectory and road network matching model based on Bayesian network are established.Under the premise of considering the correlation between multiple adjacent trajectory points and sections of road,the probability of a certain trajectory belonging to each road section is calculated,then taking the highest possible section as the estimated result to achieve the matching between the trajectory and the road network;after eliminating the abnormal trajectory,and according to the relationship between the road vector and the trajectory vector,the retrograde cycling behaviors of the shared bicycle is identified.(4)Taking the core area of Beijing as an example,and applying the research method in this paper,the typical shared bicycle trajectory data is processed and cleaned,also matched with road network,and the proportion and characteristics of retrograde cycling in each road are measured and calculated.In comparison with field investigation result,the effectiveness of the method is verified,and the causes of retrograde cycling are analyzed,thus targeted reconstruction suggestions are put forward in order to support the management of the transportation industry.There are 62 figures,17 tables and 103 references in the paper.