Civil & Environmental Engineering
Permanent URI for this communityhttp://hdl.handle.net/1903/2221
Browse
3 results
Search Results
Item SHORT TERM TRAVEL BEHAVIOR PREDICTION THROUGH GPS AND LAND USE DATA(2015) Krause, Cory; Zhang, Lei; Civil Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The short-term destination prediction problem consists of capturing vehicle Global Positioning System (GPS) traces and learning from historic locations and trajectories to predict a vehicle’s destination. Drivers have predictable trip destinations that can be estimated through probabilistic modeling of past trips. This dissertation has three main hypotheses; 1) Employing a tiered Markov model structure will permit a shorter learning period while achieving similar accuracy results, 2) The addition of derived trip purpose information will increase accuracy of the start of trip and in-route models as a whole, and 3) Similar methodologies of travel pattern inference can be used to accurately predict trip purpose and socio-economic factors. To study these concepts, a database of GPS driving traces (120 participants for 70 days) is collected. To model the user’s trip purpose, a new data source was explored: Point of Interest (POI)/land use data. An open source land use/POI dataset is merged with the GPS dataset. The resulting database includes over 20,000 trips with travel characteristics and land use/POI data. From land use/POI data, and travel patterns, trip purpose is calculated with machine learning methods. A new model structure is developed that uses trip purpose when it is available, yet falls back on traditional spatial temporal Markov models when it is not. The start of trip model has an overall increase of accuracy over other start of trip models of 2%. This comes quickly, needing only 30 days to reach this level of accuracy compared to nearly a year in many other models. When adding trip purpose and the start of trip model to in-route prediction methods, the accuracy of the destination prediction increases significantly: 15-30% improvement of accuracy over similar models between 0-50% of trip progression. Certain trips are predicted more accurately than others: work and home based trips average of 90% correct prediction, whereas shopping and social based trips hover around the 50% mark. In all, the greatest contribution of this dissertation is the trip purpose methodology addition and the tiered Markov model structure in gaining fast results in both the start of trip and in-route models.Item A POSITIVE MODEL OF ROUTE CHOICE BEHAVIOR AND VALUE OF TIME CALCULATION USING LONGITUDINAL GPS SURVEY DATA(2012) Krause, Cory; Zhang, Lei; Civil Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This thesis approaches the topic of value of time calculation and route choice behavior with a new and innovative methodology using a survey dataset that was uniquely designed and implemented for this purpose. The survey is a 70 day, 218 participant GPS travel survey used to track individual location constantly at one minute intervals. Using a positive behavior theory framework, an in depth knowledge database for each user is created that iteratively updates the learned behavior and experienced travel conditions for each trip the user takes. A new approach for calculating value of time is presented; using the cost and trip duration of previous trips. The bounds (or caps and floors) are averaged to achieve the individual's value of time based upon their route (and therefore cost) decisions. Also using this updating knowledge base, route decision rules are derived using machine learning algorithms to tell why a user has decided to take the toll road option for certain days, and under what conditions the user will not take the toll road option. The final contribution is a model that fully takes advantage of longitudinal GPS data to create an adaptive system for value of time calculation and positive route decision making.Item Topology Control and Pointing in Free Space Optical Networks(2007-12-05) Shim, Yohan; Gabriel, Steven A; Civil Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Free space optical (FSO) communication provides functionalities that are different from fiber optic networks and omnidirectional RF wireless communications in that FSO is optical wireless (no infrastructure installation cost involving fibers) and is highly directional (no frequency interference). Moreover, its high-speed data transmission capability is an attractive solution to the first or last mile problem to bridge to current fiber optic network and is a preferable alternative to the low data rate directional point-to-point RF communications for inter-building wireless local area networks. FSO networking depends critically on pointing, acquisition and tracking techniques for rapidly and precisely establishing and maintaining optical wireless links between network nodes (physical reconfiguration), and uses topology reconfiguration algorithms for optimizing network performance in terms of network cost and congestion (logical reconfiguration). The physical and logical reconfiguration process is called Topology Control and can allow FSO networks to offer quality of service by quickly responding to various traffic demands of network users and by efficiently managing network connectivity. The overall objective of this thesis research is to develop a methodology for self-organized pointing along with the associated autonomous and precise pointing technique as well as heuristic optimization methods for Topology Control in bi-connected FSO ring networks, in which each network node has two FSO transceivers. This research provides a unique, autonomous, and precise pointing method using GPS and local angular sensors, which is applicable to both mobile and static nodes in FSO networking and directional point-to-point RF communications with precise tracking. Through medium (264 meter) and short (40 meter) range pointing experiments using an outdoor testbed on the University of Maryland campus in College Park, sub-milliradian pointing accuracy is presented. In addition, this research develops fast and accurate heuristic methods for autonomous logical reconfiguration of bi-connected ring network topologies as well as a formal optimality gap measure tested on an extensive set of problems. The heuristics are polynomial time algorithms for a congestion minimization problem at the network layer and for a multiobjective stochastic optimization of network cost and congestion at both the physical and network layers.