ABSTRACT The focus of this paper is to investigate methods to cluster activity patterns of indivi... more ABSTRACT The focus of this paper is to investigate methods to cluster activity patterns of individuals such that the results of the clustering can be used for inference purposes in activity-based behavioral travel demand models. Such questions as how to classify activity patterns in the most informative manner, how well the clusters and their representative patterns can represent activity patterns at the population level, or how the patterns of out-of-sample individuals can be classified to the clusters for further analysis are addressed. We use a two-stage clustering technique of affinity propagation and K-means to classify activity patterns that are created by segmenting daily activities into ten-minute intervals, carrying information about activity types, duration, schedule, and travel distance. The measure used to estimate the distance between pairs of patterns is a weighted combination of agenda dissimilarity and edit-distance. Decision trees are employed to model dependencies between individuals’ socio-demographics and cluster allocation that then enable the methodology to be used as a model for pattern inference and practical applications.
Chains of activities performed during the course of the day are interconnected such that particip... more Chains of activities performed during the course of the day are interconnected such that participation in one activity and the time allocated to that specific activity correspondingly influences the time use behavior of a traveler along the course of the day. This points to the importance of analyzing trajectories of patterns as a set of activities with such specific characteristics as start time, duration, and sequence, rather than simply analyzing participation in each activity singularly. In this paper we present a methodology to answer a main question in trajectory analysis— ‘how to generate activity patterns trajectories, and how to conduct useful analysis that eventually makes inferences drawn from the time use behavior of individuals applicable to the population-at-large possible?’ The methodology presented here can be applied to synthetize chains of activities and their space time distribution. It starts with clustering activity patterns into a small set of representative patterns by using message passing algorithms, and then capturing the correlation among demographic profiles of travelers to the bundles of activities performed and their corresponding time sequence using multivariate probit models. We apply the methodology to two sets of data: (1) California household travel survey data for year 2000-2001, and (2) California household travel survey data for year 2010-2011. The longitudinal analysis performed in this work: (1) proves the robustness of proposed methodology in replicating time-use behavior and synthetizing activity chains; (2) reveals dynamics of changes in the trajectories of activity patterns during a 10-year time span; and (3) quantifies the influence of different sociodemographic variables on the trajectories of activities performed by travelers by implementing a statistical analysis on the distribution of estimates.
Recently, path flow estimator (PFE) has been used for the estimation of origin-destination (O-D) ... more Recently, path flow estimator (PFE) has been used for the estimation of origin-destination (O-D) matrices. This paper develops a formulation that incorporates the decoupled path flow estimator in a generalized least squares (GLS) framework. The approach seeks to solve a GLS problem that minimizes the sum of errors in traffic counts and O-D matrices based on equilibrium assignment mapping, which is derived from a K-shortest path ranking procedure exogenously. Solving the GLS-PFE model inevitably involves non-invertible linear systems and nonnegative constraints. A solution algorithm is thus designed to iteratively identify active constraints and solve linear systems by computing the pseudo-inverse. A simplified version of this algorithm is further developed to improve its computational efficiency. The solution properties and computational efficiency of the two methods are tested and compared with small to mid-size networks. It is concluded that the simplified algorithm is efficient in solving the large-scale decoupled GLS-PFE model for O-D estimation.
Center For Traffic Simulation Studies, Aug 1, 2002
ABSTRACT This paper presents a micro-simulation method to evaluate potential ATMIS applications. ... more ABSTRACT This paper presents a micro-simulation method to evaluate potential ATMIS applications. Based on the commercial PARAMICS model, a capability-enhanced PARAMICS simulation environment, PARAMICS-E, has been developed through integrating a ...
The California Advanced Transportation Management Systems (ATMS) Testbed program focuses on the d... more The California Advanced Transportation Management Systems (ATMS) Testbed program focuses on the development of an integrated transportation operations system (TOS) based on real-time computer-assisted traffic management and communication. As distinct from previous “smart corridor” developments, this TOS is structured to integrate network-wide traffic information (both surface street and freeway) in a real-time environment and to provide “intelligent” computer-assisted decision support to traffic management personnel. With funding from the California Department of Transportation (Caltrans) Division of New Technology and Research, a research program to accomplish this ambitious objective has been designed as a symbiosis between traffic management research and technology development, with provision for implementation of promising research developments in a real-world testbed. The effort is organized along three principal dimensions: (1) TOS decision support; (2) ATMS strategic response; and (3) database management, integration and communication
ABSTRACT The focus of this paper is to investigate methods to cluster activity patterns of indivi... more ABSTRACT The focus of this paper is to investigate methods to cluster activity patterns of individuals such that the results of the clustering can be used for inference purposes in activity-based behavioral travel demand models. Such questions as how to classify activity patterns in the most informative manner, how well the clusters and their representative patterns can represent activity patterns at the population level, or how the patterns of out-of-sample individuals can be classified to the clusters for further analysis are addressed. We use a two-stage clustering technique of affinity propagation and K-means to classify activity patterns that are created by segmenting daily activities into ten-minute intervals, carrying information about activity types, duration, schedule, and travel distance. The measure used to estimate the distance between pairs of patterns is a weighted combination of agenda dissimilarity and edit-distance. Decision trees are employed to model dependencies between individuals’ socio-demographics and cluster allocation that then enable the methodology to be used as a model for pattern inference and practical applications.
Chains of activities performed during the course of the day are interconnected such that particip... more Chains of activities performed during the course of the day are interconnected such that participation in one activity and the time allocated to that specific activity correspondingly influences the time use behavior of a traveler along the course of the day. This points to the importance of analyzing trajectories of patterns as a set of activities with such specific characteristics as start time, duration, and sequence, rather than simply analyzing participation in each activity singularly. In this paper we present a methodology to answer a main question in trajectory analysis— ‘how to generate activity patterns trajectories, and how to conduct useful analysis that eventually makes inferences drawn from the time use behavior of individuals applicable to the population-at-large possible?’ The methodology presented here can be applied to synthetize chains of activities and their space time distribution. It starts with clustering activity patterns into a small set of representative patterns by using message passing algorithms, and then capturing the correlation among demographic profiles of travelers to the bundles of activities performed and their corresponding time sequence using multivariate probit models. We apply the methodology to two sets of data: (1) California household travel survey data for year 2000-2001, and (2) California household travel survey data for year 2010-2011. The longitudinal analysis performed in this work: (1) proves the robustness of proposed methodology in replicating time-use behavior and synthetizing activity chains; (2) reveals dynamics of changes in the trajectories of activity patterns during a 10-year time span; and (3) quantifies the influence of different sociodemographic variables on the trajectories of activities performed by travelers by implementing a statistical analysis on the distribution of estimates.
Recently, path flow estimator (PFE) has been used for the estimation of origin-destination (O-D) ... more Recently, path flow estimator (PFE) has been used for the estimation of origin-destination (O-D) matrices. This paper develops a formulation that incorporates the decoupled path flow estimator in a generalized least squares (GLS) framework. The approach seeks to solve a GLS problem that minimizes the sum of errors in traffic counts and O-D matrices based on equilibrium assignment mapping, which is derived from a K-shortest path ranking procedure exogenously. Solving the GLS-PFE model inevitably involves non-invertible linear systems and nonnegative constraints. A solution algorithm is thus designed to iteratively identify active constraints and solve linear systems by computing the pseudo-inverse. A simplified version of this algorithm is further developed to improve its computational efficiency. The solution properties and computational efficiency of the two methods are tested and compared with small to mid-size networks. It is concluded that the simplified algorithm is efficient in solving the large-scale decoupled GLS-PFE model for O-D estimation.
Center For Traffic Simulation Studies, Aug 1, 2002
ABSTRACT This paper presents a micro-simulation method to evaluate potential ATMIS applications. ... more ABSTRACT This paper presents a micro-simulation method to evaluate potential ATMIS applications. Based on the commercial PARAMICS model, a capability-enhanced PARAMICS simulation environment, PARAMICS-E, has been developed through integrating a ...
The California Advanced Transportation Management Systems (ATMS) Testbed program focuses on the d... more The California Advanced Transportation Management Systems (ATMS) Testbed program focuses on the development of an integrated transportation operations system (TOS) based on real-time computer-assisted traffic management and communication. As distinct from previous “smart corridor” developments, this TOS is structured to integrate network-wide traffic information (both surface street and freeway) in a real-time environment and to provide “intelligent” computer-assisted decision support to traffic management personnel. With funding from the California Department of Transportation (Caltrans) Division of New Technology and Research, a research program to accomplish this ambitious objective has been designed as a symbiosis between traffic management research and technology development, with provision for implementation of promising research developments in a real-world testbed. The effort is organized along three principal dimensions: (1) TOS decision support; (2) ATMS strategic response; and (3) database management, integration and communication
VISLABE is a web application that has been built to visualize and understand lane-changing maneuv... more VISLABE is a web application that has been built to visualize and understand lane-changing maneuvers (LCM) in freeway-to-freeway connectors. It is implemented using a single programming language, R, an open-source GNU project, which significantly simplifies web application development. The web application is dynamic, mobile device ready and cloud based-software installations are not required for the end user. We show how the use of R and a set of recently developed R packages can aid researchers to publish and dynamically visualize their work on the web. VISLABE is initially tailored towards the visualization of lane changing maneuvers from a rich vehicle trajectory dataset collected at four different Southern California freeways. It provides various tools to analyze the relevance between LCM and such explanatory variables as speed, geometry, level of congestion, and vehicle types. To do so, a combination of time series, histograms, heat maps, contour and density plots is used. Concerning LCM analysis, VISLABE is effective in providing insights to traffic flow dynamics and the relationship between LCM and traffic breakdown. The data collected captured the initiation and propagation of traffic congestion. It was found that the causality of traffic breakdown was the increase of LCM since the freeway mainline demands remained relatively constant before and after the occurrence of congestion.
Uploads
Papers by Will Recker