Study On Key Technologies Of Transit Passengers' Travel Pattern Mining And Applications Based On Multiple Sources Of Data

With the rapid progress of urbanization,public transit systems have been regarded as the most critical mode of transportation.Nowadays,urban transit networks are facing several challenges: expanding network scales,changing variability of ridership along with variously growing reachability demands of residents.Unfortunately,the transportation agencies lack of efficient methods to understand and utilize passengers' travel pattern and service preferences so as to benefit the design and adjust of bus network.Therefore,it's of great significance to leverage the huge quantity of passengers' transaction records to mine their travel demands along with behavioral patterns so as to adjust bus headways and dwelling stations thereby contributing the performance enhancement and quality of service in urban transportation systems.In this research,we propose a series of methodologies to mine transit riders' travel pattern and behavioral preferences,and then we use these knowledge to adjust and optimize the transit system of collaborative city.The contributions of this research are:Firstly,We propose a series of methodologies for data preprocessing in order to increase the data validity of our collaborator.Our major contribution of this part are: a)we propose a novel approach to rectify the time discrepancy of data between the AFC(Automated Fare Collection)systems and AVL(Automated Vehicle Location)system,our approach transforms data events into discrete signals and then applies time domain correlation the detect and rectify their relative synchronization discrepancies.b)By combining historical data and passengers' synchronized ticketing time stamps,we induct and compensate missing information in AVL datasets.Our approach can greatly enhance data validity of urban transit systems of China.Secondly,In order to leverage the drawbacks of state-of-art algorithm for passengers' alighting point estimation,we first introduce maximum probabilistic inference and passengers' home place estimation to recover their complete transit trajectory from semi-complete boarding records.Then we propose an enhance transfer activity identification algorithm which is capable of specifying passengers' short-term activity from ordinary transfer procedures.Finally,we provide our analysis on the whole city's temporal-spatial distribution of ridership using recovered passenger trajectory.Thirdly,To discover passengers' rigid travel demand from fragmented passengers' transit trajectories,we propose a novel graph based data fusion mechanism.We first cluster each passenger's trajectory data in multiple days and construct a Hybrid Trip Graph(HTG).We then use a depth search algorithm to derive the spatially closed transit trip chains from HTG;Finally,we use closed transit trip chains of passengers' in the whole city to study passengers' travel pattern from various aspects.Finally,we proposed another option to discover transit corridors of the target city by aggregating the passengers' critical transit chains.Fourthly,to obtain deeper understanding on transit passengers' route choice preference,we first derive eight factors that may have influence on passengers' transit route choice,and then construct each passengers' transit route choice model in regard of different service scenarios they have encountered.Next,we verify and assure our model by using ridership re-distribute simulation experiments.Finally,we conduct a comprehensive analysis on the temporal activity patterns of passengers with different route choice preference.Finally,to make use of the rich information of passengers travel pattern thereby providing better transit services.We first provide a novel transit route optimization method by integrating passengers' transit route choice models and particle swarm optimization algorithm.Then,we then use a real-world case to study the result of various optimization target from the following perspective: passenger flow,passengers' time efficiency,vehicles' headways.Finally,we compare the ridership distribution and passengers' travel time efficiency before and after the establishment of new routes.Our research can provide useful insights in leveraging the power of Big Data to enhance the performance of public transit systems.