Transfer Learning-Based Deep Reinforcement Learning Approach for Robust Route Guidance in Mixed Traffic Environment

Human-driven Vehicles (HVs) and Autonomous Vehicles (AVs) will inevitably coexist in the future, resulting in mixed traffic conditions. Modeling car-following behavior in such mixed traffic is crucial for better understanding the differences in driving behavior between AVs and HVs, thereby facilitating the development of AV driving strategies to enhance traffic safety. In such mixed traffic, the car-following modes typically include AVs following HVs, HVs following HVs and HVs following AVs. Previous studies have compared the driving behavior of HVs following HVs and HVs following AVs, but few have considered the driving behavior of AVs following HVs in the comparison. To bridge this gap, this paper first decomposes car-following events in mixed traffic extracted from a Lyft level-5 dataset into driving primitives using the Hidden Markov Model with Gaussian Mixture Model (GMM-HMM). Then, the Latent Dirichlet Allocation with GMM (GMM-LDA) is introduced to cluster the driving primitives into different car-following behavior patterns. Based on the identified patterns, the differences across the three car-following modes are further explored. The results show that the driving behavior under three car-following modes differs from each other. Specifically, compared to the HVs following HVs mode, in the AVs following HVs mode, AVs tend to maintain car-following behavior patterns with a longer distance gap and adopt a lower speed to ensure safety. This study offers a comprehensive understanding of car-following behavior in mixed traffic, which will facilitate the development of human-like driving strategies for AVs and enhance the safety of mixed traffic.

Urban link travel time estimation using large-scale taxi data with partial information

In this paper, we study the problem of origin-destination (OD) travel time estimation where the OD input consists of an OD pair and a departure time. We propose a novel neural network based prediction model that fully exploits an important fact neglected by the literature -- for a past OD trip its travel time is usually affiliated with the trajectory it travels along, whereas it does not exist during prediction. At the training phase, our goal is to design novel representations for the OD input and its affiliated trajectory, such that they are close to each other in the latent space. First, we match the OD pairs and their affiliated (historical) trajectories to road networks, and utilize road segment embeddings to represent their spatial properties. Later, we match the timestamps associated with trajectories to time slots and utilize time slot embeddings to represent the temporal properties. Next, we build a temporal graph to capture the weekly and daily periodicity of time slot embeddings. Last, we design an effective encoding to represent the spatial and temporal properties of trajectories. To bind each OD input to its affiliated trajectory, we also encode the OD input into a hidden representation, and make the hidden representation close to the spatio-temporal representation of the trajectory. At the prediction phase, we only use the OD input, get the hidden representation of the OD input, and use it to generate the travel time. Extensive experiments on real datasets show that our method achieves high effectiveness and outperforms existing methods.

The ubiquitous massive mobile phone data generation presents new opportunities to determine the requirements of transportation, disaster management and public health care systems. Currently, data from mobile phone records can help in identifying the location of the users while they are making trips. Generally, this estimation is achieved using traditional data collection methods; however, these methods are difficult to apply in developing countries with rapidly growing cities owing to the high population and limitation in conducting a survey. Call detail records (CDRs) are used as base data because they are valuable data sources and can reduce the cost and time limitations. The aim of this study is to estimate origin-destination (OD) trips from each zone by using the CDRs. The OD trips are estimated by using the CDRs of one week taken from Myanmar Post and Telecommunication mobile operator for over 1.9 million users per day in Yangon, the economic center of Myanmar. The OD trips are estimated from CDRs based on the location of the base station in a limited time window and time frame. If the same mobile users is observed in two different the ones within the time limit, it is assumed that the mobile user is coming out from the first zone and the trips represents an originating trip. This trip would be the destination trip for zone where the mobile user enters. In this study, the originating (outgoing) and destination trips (incoming) from each township on a weekday and weekend are determined. These data are useful for infrastructure development and urban transportation planning.

The continuous growth of automobile traffic on urban and suburban arterials in recent years has created a substantial problem for transit, especially when it operates in mixed traffic conditions. As a result, there has been a growing interest in deploying Transit Signal Priority (TSP) to improve the operational performance of arterial corridors. TSP is an operational strategy that facilitates the movement of transit vehicles (e.g., buses) through signalized intersections that helps transit service be more reliable, faster, and more cost-effective. The goal of this research was to quantify the mobility and safety benefits of TSP. A microscopic simulation approach was used to estimate the mobility benefits of TSP. Microscopic simulation models were developed in VISSIM and calibrated to represent field conditions. Implementing TSP provided significant savings in travel time and average vehicle delay. Under the TSP scenario, the study corridor also experienced significant reduction in travel time and average vehicle delay for buses and all other vehicles. The importance and benefits of calibration of VISSIM model with TSP integration were also studied as a part of the mobility benefits. Besides quantifying the mobility benefits, the potential safety benefits of the TSP strategy were also quantified. An observational before-after full Bayes (FB) approach with a comparison-group was adopted to estimate the crash modification factors (CMFs) for total crashes, fatal/injury (FI) crashes, property damage only (PDO) crashes, rear-end crashes, sideswipe crashes, and angle crashes. The analysis was based on 12 corridors equipped with the TSP system and their corresponding 29 comparison corridors without the TSP system. Overall, the results indicated that the deployment of TSP improved safety. Specifically, TSP was found to reduce total crashes by 7.2% (CMF = 0.928), FI crashes by 14% (CMF = 0.860), PDO crashes by 8% (CMF = 0.920), rear-end crashes by 5.2% (CMF = 0.948), and angle crashes by 21.9% (CMF = 0.781). Alternatively, sideswipe crashes increased by 6% (CMF = 1.060), although the increase was not significant at a 95% Bayesian credible interval (BCI). These results may present key considerations for transportation agencies and practitioners when planning future TSP deployments.

Capacity uncertainty on urban road networks: A critical state and its applicability in resilience quantification

Origin-Destination (OD) trip matrices describe the patterns of traffic behavior across the network and play a key role as primary data input to many traffic models. OD matrices are a critical requirement, either in static or dynamic models for traffic assignment. However, OD matrices are not yet directly observable, thus, the current practice consists of adjusting an initial or a priori matrix from link flow counts, speeds, travel times and other aggregate demand data. This information is provided by an existing layout of traffic counting stations, as the traditional loop detectors. The availability of new traffic measurements provided by ICT applications offers the possibility to formulate and develop more efficient algorithms, especially suited for real-time applications. However, the efficiently strongly depends, among other factor, on the quality of the seed matrix. This paper proposes an integrated computational framework in which an off-line procedure generates the time-sliced OD matrices, which are the input to an on-line estimator. The paper also analyzes the sensitivity of the on-line estimator with respect to the available ICT traffic measurements.

With the advent of Advanced Transportation Management and Information System (ATMIS), much attention has been paid to the estimation of dynamic or time-variant OD matrices, as development of improved methods for the derivation of OD-based real time traffic information is vital for analysis of transportation systems and various ATMIS strategies involving traveler information systems and route guidance, dynamic traffic assignment, and adaptive traffic signal control, among others. This study performs a systematic simulation investigation of the performance and feasibility of anonymous vehicle tracking in signalized networks using the Paramics simulation model. Previous research experience with vehicle reidentification techniques on single roadway segments is used to investigate the performance obtainable from tracking individual vehicles across multiple detector stations through a network to obtain real-time OD path flow information such as travel time and volume. The findings of this and subsequent studies serve as a logical and necessary precursor to possible field implementation in signalized networks of vehicle reidentification techniques.

Total shoulder arthroplasty (TSA) is a common treatment to decrease pain and improve shoulder function in patients with severe osteoarthritis (OA). In Canada, patients requiring this procedure often wait a year or more. Our objective was to determine patient preferences related to accessing TSA, specifically comparing out-of-pocket payments for treatment, travel time to hospital, the surgeon's level of experience and wait times. We administered a discrete choice experiment among patients with endstage shoulder OA currently waiting for TSA. Respondents were presented with 14 different choice sets, each with 3 options, and they were asked to choose their preferred scenario. A conditional logit regression model was used to estimate the relative preference and willingness to pay for each attribute. Sixty-two respondents completed the questionnaire. Three of the 4 attributes significantly influenced treatment preferences. Respondents had a strong preference for an experienced surgeon (mean 0.89 ± standard error [SE] 0.11), while reductions in travel time (-0.07 ± 0.04) or wait time (-0.04 ± 0.01) were of less importance. Respondents were found to be strongly averse (-1.44 ± 0.18) to surgical treatment by a less experienced surgeon and to paying out-of-pocket for their surgical treatment (-0.56 ± 0.05). Our results suggest that patients waiting for TSA to treat severe shoulder OA have minimal willingness to pay for a reduction in wait time or travel time for surgery, yet will pay higher amounts for treatment by an experienced surgeon.

An optimal global algorithm for route guidance in advanced traveler information systems

Travel time reliability assessment has been widely used in recent years to evaluate the performance of transportation networks and measure the operation level of transportation systems. Weather, as one of the important factors influencing travel time reliability, affects the relationship between the supply and requirement of urban road networks. Considering the traffic characteristics under different traffic conditions, a study on the influence of weather on travel time reliability under different conditions is proposed to predict the probability of travelers completing their trips within the expected time under different weather conditions. Based on the urban road network data and cab trajectory data of Harbin city, this paper correlates the floating vehicle location with the road network information through a hidden Markov model to reduce the influence of vehicle trajectory errors on the calculation results of path travel time. To analyze the entire distribution of extreme travel time and its impact on the reliability of travel time under various traffic situations, it captures the tail features of the travel time distribution based on extreme value theory. Then, to increase the predictability of each quantile, it combines a deep-learning LSTM model and a quantile regression model to create a probabilistic travel time prediction model utilizing combined layers. The proposed model is compared with the linear quantile regression and neural network quantile regression models, and the model is evaluated in terms of point prediction results and probabilistic prediction results, respectively, to ensure the accuracy of predictions from the model. As a result, the prediction accuracy of the model in this paper is greatly improved, and the degree of violating quantile constraints is greatly reduced.

실시간 자료를 반영한 통행시간 예측 기법은 다양하지만 관련 연구 검토 결과 과거이력데이터가 충분하다면 타 모형에 비해 K 최대근접이웃(K-Nearest Neighbors)의 정확도가 우수하므로 본 연구에서는 이에 대한 적용 방법 도출 및 가능성 평가를 목적으로 한다. 본 연구에서는 KNN의 입력 자료로 TCS 교통량 및 DSRC 구간통행시간의 실시간 및 과거 이력자료, 경로통행시간 이력자료를 활용하였다. 통행시간 예측치는 TCS 교통량 및 DSRC 구간통행시간의 실시간 자료와 유사한 경로통행시간을 탐색한 후 이를 가중평균하여 산출하였다. 예측 기법을 적용한 결과 DSRC 구간통행시간의 가중치가 증가할수록 정확도는 증가하였으며, 이는 실시간 교통상황 변화를 DSRC 구간통행시간이 잘 반영하기 때문이다. 그러나 TCS 교통량을 기반으로 한 경우 역시 정확도의 차이가 크지 않으며, 변화 추이도 유사하게 나타났다. 이러한 결과를 볼 때 향후 대용량의 과거이력자료가 축적될 경우 예측오차는 더욱 감소될 것으로 기대된다. There are various methodologies to forecast the travel time using real-time data but the K-nearest neighborhood (KNN) method in general is regarded as the most one in forecasting when there are enough historical data. The objective of this study is to evaluate applicability of KNN method. In this study, real-time and historical data of toll collection system (TCS) traffic flow and the dedicated short range communication (DSRC) link travel time, and the historical path travel time data are used as input data for KNN approach. The proposed method investigates the path travel time which is the nearest to TCS traffic flow and DSRC link travel time from real-time and historical data, then it calculates the predicted path travel time using weight average method. The results show that accuracy increased when weighted value of DSRC link travel time increases. Moreover the trend of forecasted and real travel times are similar. In addition, the error in forecasted travel time could be further reduced when more historical data could be available in the future database.

Traffic states prediction in air transportation systems is a challenging problem and has not been fully explored because it is subject to many more highly correlated factors and a more complicated traffic management scheme compared to urban transportation systems. It becomes a more formidable task when facing a multi-airport system (MAS), in which several major airports are closely located and tightly coupled with each other through limited terminal airspace. In this work, we propose a novel method using a time series model and recurrent neural network to make the estimated time of arrival (ETA) for a flight to an MAS, which can be potentially utilized for flight delay prediction and congestion analysis. The experiment utilizes two months of 4D trajectories data from Beijing Capital International Airport (PEK) to Shenzhen Bao’an International airport (ZGSZ). The entire prediction work is decomposed into two sub-problems, en-route travel time prediction which is from flight origin to the entering gate of MAS, defined as the location is 300km from the airport in MAS, and terminal maneuvering area (TMA) travel time prediction which is from the entrance to flight’s destination. The auto-regressive integrated moving average (ARIMA), a time series prediction model, is used to predict travel time in en-route under given the flight departure time. Bidirectional long short term memory (LSTM), a recurrent neural network, is developed to forecast travel time in the arrival approach by utilizing spatio-temporal features. To design the input features, we use density-based spatial clustering (DBSCAN) with the help of the Voronoi diagram to extract spatial information from every historical flight trajectory of aircraft operated in an MAS, then select the observation time window to capture the temporal information for each flight. The Multivariate Stacked Fully connected-Bidirectional LSTM (MSFCB-LSTM) model is constructed to make shortterm forecasting using spatio-temporal features we designed when the flight’s entering MAS time is given. For TMA travel time prediction, a case study of Guangdong-Hong Kong-Macao Greater Bay Area (GHM-GBA), a typical MAS which contains five major airports closely located within 120km, is carried out using actual historical 4D trajectory data. Finally, Using two months 4D trajectories data, PEK to ZGSZ, the result exhibits the best accuracy, a measurement we define for prediction, of the longterm prediction of ETA given departure time is 92%, and mean absolute error (MAE) is 6.09 minutes.

Phase diagram in multi-phase heterogeneous traffic flow model integrating the perceptual range difference under human-driven and connected vehicles environment

Livable cities involve such quality-of-life factors as transportation, convenience of daily life, education, and a safe and stable built and natural environment. The livability of a city also has social and psychological dimensions, such as emotion and perception. This paper explores how the advantages of new transportation technology can be realized under mixed-traffic conditions, while meeting the requirements for safety, convenience, and high educational standards. One of the key challenges to realizing the advantages of a new technology, such as autonomous shuttle buses, under mixed-traffic conditions and under adverse weather conditions is how to strike a balance between innovation and adherence to essential urban demands. Ensuring that the introduction of advanced technologies like autonomous shuttle buses does not compromise safety is of the utmost importance. Under mixed-traffic conditions, autonomous shuttle buses must negotiate interactions with human-driven vehicles, pedestrians, and bicyclists. Predictable and safe driving behavior is vital to gaining the confidence and acceptance of all road users. Additionally, adopting vehicle-to-everything (V2X) communication enables real-time information exchange with other vehicles and infrastructure, which enhance safety. Autonomous shuttle buses also face crucial challenges in adverse weather conditions. Robust sensor systems are required to accurately perceive the environment despite rain, snow, fog, and other weather-related constraints. Continual learning and software updates ensure the system’s ability to adapt to changing weather conditions. To overcome these challenges, collaboration with regulatory bodies and traffic management authorities is essential. Furthermore, public awareness campaigns and educational initiatives can inform the public about the safety measures and benefits of autonomous shuttle buses, fostering acceptance and reducing concerns during the transition.