Forecasting the short-term ridership among origin-destination pairs (OD matrix) of a metro system is crucial in real-time operation. However, this problem notoriously difficult due to high-dimensional, sparse, noisy, and skewed nature OD matrices. This paper proposes High-order Weighted Dynamic Mode Decomposition (HW-DMD) model for matrices forecasting. DMD uses Singular Value (SVD) extract low-rank approximation from data, high-order vector autoregression established To address practical...

Accurate forecasting of bus travel time and its uncertainty is critical to service quality operation transit systems: it can help passengers make informed decisions on departure time, route choice, even transport mode also support operators tasks such as crew/vehicle scheduling timetabling. However, most existing approaches in are based deterministic models that provide only point estimation. To this end, we develop paper a Bayesian probabilistic model for estimated arrival (ETA)....

Accurately monitoring road traffic state is crucial for various applications, including travel time prediction, control, and safety. However, the lack of sensors often results in incomplete data, making it challenging to obtain reliable information decision-making. This paper proposes a novel method imputing data using Gaussian processes (GP) address this issue. We propose kernel rotation re-parametrization scheme that transforms standard isotropic GP into an anisotropic kernel, which can...

Metro systems in megacities such as Beijing, Shenzhen and Guangzhou are under great passenger demand pressure. During peak hours, it is common to see oversaturated conditions (i.e., exceeds network capacity), which bring significant operational risks safety issues. A popular control intervention restrict the entering rate during hours by setting up out-of-station queueing with crowd barriers. The \textit{out-of-station waiting} can make a substantial proportion of total travel time but not...

Estimation of link travel time correlation a bus route is essential to many operation applications. Link on could exhibit complex structures, such as long-range correlations, negative and time-varying correlations. This paper develops Bayesian Gaussian model estimate the matrix using smart-card-like data. Our method overcomes small-sample-size problem in estimation by borrowing/integrating those incomplete observations from other routes. We first conduct synthetic experiment results show...

Modeling the relationship between vehicle speed and density on road is a fundamental problem in traffic flow theory. Recent research found that using least-squares (LS) method to calibrate single-regime speed-density models biased because of uneven distribution samples. This paper explains issue LS from statistical perspective: calibration caused by correlations/dependencies regression residuals. Based this explanation, we propose new for modeling covariance residuals via zero-mean Gaussian...

Car-following models are essential for microscopic traffic simulation. While conventional rely on parsimonious formulas with simplified assumptions, recent studies have focused developing data-driven the help of high-resolution trajectory data. This paper presents a model based Bayesian Gaussian mixture (GMM) probabilistic forecasting human car-following behaviors. By incorporating past and future information, our captures temporal dynamics behaviors, providing accurate predictions following...

This paper analyses the effects of congestion and stochastic perceived error in traffic assignment paradox, by measure both actual travel cost. Two different circumstances are studied: improving an existing link adding a new link. It is found that cost functions levels will significantly affect road condition demand level under which paradox happens. Moreover, how interaction between demands O-D pairs affects occurrence illustrated two pairs' network. Besides, counter-intuitive phenomenon...

Predicting metro incident duration is crucial for passengers and transit operators to choose appropriate response strategies. Most existing research focuses on structured data, the rich information embedded within unstructured logs often neglected. This paper incorporates a probabilistic topic model tailored short texts, biterm model, into generic prediction models. By capturing text co-occurrence patterns through Bayesian inference, extracts hidden topics from narratives, each serves as...

Abstract Improving bus travel time reliability can attract more commuters to use transit, and therefore reduces the share of cars alleviates traffic congestion. This paper formulates a new metric that jointly considers two stochastic processes: in-stop waiting process in-vehicle process, function is calculated by convolution independent events’ probabilities. The defined as probability when less than certain threshold be used in both conditions with without transfer. Next, Automatic Vehicle...

Spatiotemporal traffic data imputation is of great significance in intelligent transportation systems and data-driven decision-making processes. To make an accurate reconstruction from partially observed data, we assert the importance characterizing both global local trends time series. In literature, substantial prior works have demonstrated effectiveness utilizing low-rankness property by matrix/tensor completion models. this study, first introduce a Laplacian kernel to temporal...

Accurately monitoring road traffic state is crucial for various applications, including travel time prediction, control, and safety. However, the lack of sensors often results in incomplete data, making it challenging to obtain reliable information decision-making. This paper proposes a novel method imputing data using Gaussian processes (GP) address this issue. We propose kernel rotation re-parametrization scheme that transforms standard isotropic GP into an anisotropic kernel, which can...

Pushbutton control is ideal for midblock crossings with low pedestrian and vehicle demand, but it causes significant interruptions to traffic flow frequent crossing requests. Therefore, we propose an adaptive (AMCC) that minimizes the impact of pushbutton on while maintaining a reasonably short wait time (PWT). We regard two adjacent intersections as integrated system types AMCCs—AMCC-band AMCC-vehicle—based different real-time information. AMCC-band seeks best PWT at minimize green band...

Accurately forecasting bus travel time and passenger occupancy with uncertainty is essential for both travelers transit agencies/operators. However, existing approaches to mainly rely on deterministic models, providing only point estimates. In this paper, we develop a Bayesian Markov regime-switching vector autoregressive model jointly forecast uncertainty. The proposed approach naturally captures the intricate interactions among adjacent buses adapts multimodality skewness of real-world...

Origin-destination (OD) demand matrices are crucial for transit agencies to design and operate systems. This paper presents a novel temporal Bayesian model designed estimate OD at the individual bus-journey level from boarding/alighting counts bus stops. Our approach begins by modeling number of alighting passengers subsequent stops, given boarding stop, through multinomial distribution parameterized probabilities. Given large scale problem, we generate probabilities with latent variable...

In recent years, a multiplicative hybrid (MH) route choice model was proposed to overcome the drawbacks of multinomial logit (MNL) and weibit (MNW) model. This paper compares conditions for stochastic traffic assignment paradox three models. We analyze condition when improving link in an uncongested network counterintuitively increases total travel costs. Using typical flow-independent networks (two links, n independent routes with m overlapping links), we reveal strong relationships further...

Individual mobility prediction is an essential task for transportation demand management and traffic system operation. There exist a large body of works on modeling location sequence predicting the next users; however, little attention paid to trip, which governed by strong spatiotemporal dependencies between diverse attributes, including trip start time $t$, origin $o$, destination $d$. To fill this gap, in paper we propose novel point process-based model -- Attentive Marked temporal...