Master of Science, The Ohio State University, 2010, Civil Engineering

The passenger flows between origins and destinations on bus routes is information that transportation planners and operators can utilize in planning, designing, and managing the operations of transit services. Manual methods of collecting data on passenger origin-destination (OD) flows are difficult to implement regularly because they require much planning, time, and personnel. Automatic passenger counting (APC) technologies, which provide comprehensive data collection on the boarding and alighting volumes at all bus stops on a route, are increasingly being installed on bus systems. Typically, the APC data provide planners with insights on bus loads along transit routes. However, APC data may also be beneficial in estimating the difficult-to-obtain passenger OD flow information. The research presented in this thesis sought to assess the performance of a simple method that can use APC data to estimate passenger OD flows, namely, the Iterative Proportional Fitting (IPF) procedure used with the uninformative null matrix adopted as the base (or seed) matrix. The base matrix is used as an initial starting point in the IPF procedure to produce an output consistent with the APC data. A null base matrix assumes that each matrix cell is equally likely to be observed (i.e., passengers are equally likely to travel along any feasible OD pair). This equally likely specification can be considered a very crude and uninformative starting point that is based on no knowledge of the passenger flow patterns. The null base matrix is considered in this thesis because its simplicity would allow it to be used easily in practice. Given its unrealistic nature in representing OD flows, using the null base with the IPF procedure to produce OD flow estimates may not be expected to produce good estimates. Therefore, it would be important to assess the performance of the IPF procedure with a null base (referred to as the “IPF-null” procedure in this thesis) in estimating OD flows at the bus trip (open full item for complete abstract)

Committee: Mark McCord PhD (Advisor); Rabi Mishalani PhD (Advisor); Prem Goel PhD (Committee Member) Subjects: Civil Engineering

Master of Science, The Ohio State University, 2021, Civil Engineering

The Iterative Proportional Fitting (IPF) method represents the state-of-the-practice for estimating bus route-level passenger OD flow matrices. It requires boarding and alighting volumes along with a seed matrix as inputs. Although the “Null” matrix, where passengers are equally likely to travel between any of the feasible OD pairs, is used as a default seed, empirical onboard survey data provide an informative seed matrix. Due to the small sample sizes of such surveys, some OD pair entries are of low confidence leading to poorer quality OD flow estimates when such an empirical matrix serves as a seed for the IPF method. To improve the quality of OD flow estimates when using an empirical seed matrix, a linearly weighted combination of an empirical OD matrix and some other OD matrix is proposed to form the seed matrix. A Monte Carlo simulation-based approach is developed to evaluate the use of this approach and understand the conditions that influence its impact on the quality of the estimated OD flow matrix. Empirical data from three operational bus routes at two different transit systems are used to determine important input values. The “Null” matrix and the “IPF-Null” matrix, estimated by applying the IPF method with a “Null” seed matrix, are considered as alternatives to be combined with the empirical OD matrix to form the combined seed matrix. Results demonstrate that using “IPF-Null” is preferred to using “Null” to form the seed matrix. Therefore, an investigation of the quality of the estimated OD flow matrix as a function of the quality of the empirical OD flow matrix is conducted using the “IPF-Null” matrix to form the seed matrix. Moreover, the optimum weighting factor used to combine the two matrices to form the seed matrix as a function of the quality of the empirical OD flow matrix is also explored. For onboard survey sample sizes expected in practice, results suggest that if the optimum value of the weighting factor is known, using a combined (open full item for complete abstract)

Committee: Mark McCord (Advisor) Subjects: Civil Engineering; Engineering; Transportation; Transportation Planning

Doctor of Philosophy, The Ohio State University, 2012, Computer Science and Engineering

In real-world listening environments, speech reaching our ear is often accompanied by acoustic interference such as environmental sounds, music or another voice. Noise distorts speech and poses a substantial difficulty to many applications including hearing aid design and automatic speech recognition. Monaural speech segregation refers to the problem of separating speech based on only one recording and is a widely regarded challenge. In the last decades, significant progress has been made on this problem but the challenge remains. This dissertation addresses monaural speech segregation from different interference. First, we research the problem of unvoiced speech segregation which is less studied compared to voiced speech segregation probably due to its difficulty. We propose to utilize segregated voiced speech to assist unvoiced speech segregation. Specifically, we remove all periodic signals including voiced speech from the noisy input and then estimate noise energy in unvoiced intervals using noise-dominant time-frequency units in neighboring voiced intervals. The estimated interference is used by a subtraction stage to extract unvoiced segments, which are then grouped by either simple thresholding or classification. We demonstrate that the proposed system performs substantially better than speech enhancement methods. Interference can be nonspeech signals or other voices. Cochannel speech refers to a mixture of two speech signals. Cochannel speech separation is often addressed by model-based methods, which assume speaker identities and pretrained speaker models. To address this speaker-dependency limitation, we propose an unsupervised approach to cochannel speech separation. We employ a tandem algorithm to perform simultaneous grouping of speech and develop an unsupervised clustering method to group simultaneous streams across time. The proposed objective function for clustering measures the speaker difference of each hypothesized grouping and incorporates pitch (open full item for complete abstract)

Committee: DeLiang Wang (Committee Chair); Eric Fosler-Lussier (Committee Member); Mikhail Belkin (Committee Member) Subjects: Computer Science