Program for 2017 IEEE 20th International Conference on Intelligent Transportation Systems (ITSC): Workshop

Time Bohyoh (7F) Cherie (2F) Hisui (2F) Kohaku (2F) Rainbow (7F) Ruby (2F)

Monday, October 16

09:00-12:12 WS09: Transportation 5.0 WS05-0: Deep Learning for Autonomous Driving WS03-1: Eco-drive at Intersections with Connected,Cooperative and Automated Vehicle Technology WS02: Human Factors in Intelligent Vehicles   WS06: Artificial Transportation Systems and Simulation
09:10-09:40 WS05-1: Deep Learning for Autonomous Driving  
09:40-10:10 WS05-2: Deep Learning for Autonomous Driving  
10:10-10:40 WS05-3: Deep Learning for Autonomous Driving  
10:45-11:10   WS03-2: Eco-drive at Intersections with Connected,Cooperative and Automated Vehicle Technology  
11:00-11:10   WS07-0: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility
11:10-12:10 WS05-4: Deep Learning for Autonomous Driving WS03-3: Eco-drive at Intersections with Connected,Cooperative and Automated Vehicle Technology WS07-1: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility
11:40-12:00   WS07-2: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility
12:00-12:20     WS07-3: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility  
12:20-12:40         WS07-4: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility  
12:55-13:00 WS10-0: International workshop on Large-Scale Traffic Modeling and Management          
13:00-13:25 WS10-1: International workshop on Large-Scale Traffic Modeling and Management   WS01-0: Industry Panel: Connected, cooperative and automated transport     WS04-0: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems
13:10-16:30 WS05-5: Deep Learning for Autonomous Driving WS01-1: Industry Panel: Connected, cooperative and automated transport WS08: Behavioural Change Support Intelligent Transportation Applications  
13:15-14:15   WS04-1: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems
13:25-13:50 WS10-2: International workshop on Large-Scale Traffic Modeling and Management  
13:50-14:15 WS10-3: International workshop on Large-Scale Traffic Modeling and Management  
14:00-15:20 WS07-5: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility
14:10-14:55 WS05-6: Deep Learning for Autonomous Driving WS01-2: Industry Panel: Connected, cooperative and automated transport
14:15-15:00 WS10-4: International workshop on Large-Scale Traffic Modeling and Management WS04-2: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems
14:40-15:10 WS10-5: International workshop on Large-Scale Traffic Modeling and Management WS05-7: Deep Learning for Autonomous Driving
15:10-15:30 WS10-6: International workshop on Large-Scale Traffic Modeling and Management      
15:30-16:30     WS01-3: Industry Panel: Connected, cooperative and automated transport   WS04-3: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems
15:40-16:10   WS05-8: Deep Learning for Autonomous Driving  
15:45-16:10 WS10-7: International workshop on Large-Scale Traffic Modeling and Management  
15:50-16:50 WS07-6: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility
16:00-17:30 WS01-4: Industry Panel: Connected, cooperative and automated transport
16:10-17:10 WS10-8: International workshop on Large-Scale Traffic Modeling and Management WS05-9: Deep Learning for Autonomous Driving
16:30-17:30   WS04-4: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems
16:35-17:00 WS10-9: International workshop on Large-Scale Traffic Modeling and Management  
16:50-17:00   WS07-7: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility
17:10-17:30   WS05-10: Deep Learning for Autonomous Driving    
17:30-17:35           WS04-5: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems

WS01-0: Industry Panel: Connected, cooperative and automated transport

Opening

WS01-1: Industry Panel: Connected, cooperative and automated transport

Paper presentations
Vision Based Lane Change Detection Using True Flow Features
Leveraging Cloud Intelligence for Hybrid Vehicular Communications

WS01-2: Industry Panel: Connected, cooperative and automated transport

Invited speeches 1

Invited speeched: Haesik Kim, VTT (14:10-14:25); Satoshi Nagata, NTT DOCOMO (14:25-14:40); Feiyue Wang , Huawei (14:40-14:55)

WS01-3: Industry Panel: Connected, cooperative and automated transport

Invited speeches 2

Invited speeches: Yasin Ferdous Ahmed Khan Ericsson (15:30-15:45); Onur Altintas, Toyota (15:45-16:00)

WS01-4: Industry Panel: Connected, cooperative and automated transport

Industry Panel

Panelists: Tim Leinmüller, Denso / Onur Altintas, Toyota / Satoshi Nagata, NTT DOCOMO / Feiyue Wang, Huawei / TBD China Mobile / Yasin Ferdous Ahmed Khan, Ericsson / Koichi Sakai, University of Tokyo / Meng Lu, Dynniq

WS02: Human Factors in Intelligent Vehicles

Hypovigilance in Limited Self-Driving Automation: Peripheral Visual Stimulus for a Balanced Level of Automation and Cognitive Workload
Adaptive Behaviour Selection for Autonomous Vehicle Through Naturalistic Speed Planning
Cost-Efficient Brainwave Controller for Automated Vehicles Route Decisions
Driver's Decision Analysis in Terms of Pedestrian Attributes -A Case Study in Passing by a Pedestrian-
Coffee break
Effect on Driving Performance of Two Visualization Paradigms for Rear-End Collision Avoidance
Individualized Driver Action Anticipation Using Deep (Bidirectional) Recurrent Neural Network
Evaluating Passenger Characteristics for Ride Comfort in Autonomous Wheelchairs
A Collision Mitigation Strategy for Intelligent Vehicles to Compensate for Human Factors Affecting Manually Driven Vehicles

WS03-1: Eco-drive at Intersections with Connected,Cooperative and Automated Vehicle Technology

Signal Plan Stabilization to Enable Eco-Driving
Cluster-Wise Cooperative Eco-Approach and Departure Application Along Urban Signalized Arterials
Connected and Autonomous Vehicles Coordinating Method at Intersection Utilizing Preassigned Slots
Coffee break

WS03-2: Eco-drive at Intersections with Connected,Cooperative and Automated Vehicle Technology

Invited Session : An Extensive Investigation on Cooperative Eco-approach Controller Under Partially Connected and Automated Vehicle Environment

This study evaluated the performance of a cooperative eco-approach control system at signalized intersections under partially connected and automated vehicles environment. The controller was recently published and is the first eco-approach controller that is able to function with existence of surrounding human-driven traffic. The purpose of this work is to further conduct an extensive test on the controller in order to identify the room for improvement. Two different networks were tested, including an isolated signalized intersection and a corridor with two signalized intersections. The measurements of effectiveness (MOE) adopted are throughput, fuel consumption and CO2 emissions. All the before-and-after MOEs have been tested for statistical significance using t-tests, and the Pearson correlation of the MOEs with respect to various traffic-related and infrastructure-related factors was assessed to identify factors that significantly impact controller's performance. The results indicate the eco-approach control generally reduces fuel consumption and emission without harming mobility, and can be equipped at isolated signalized intersections or on signalized arterials. The performance of eco-approach control is affected by several traffic-related and infrastructure-related factors, including market penetration rate of connected and automated vehicles, congestion level, green ratio of traffic signal and minimum speed of connected and automated vehicle. Excessive small distance between adjacent intersections causes significant gating effect and has negative impact on the performance of eco-approach control.

WS03-3: Eco-drive at Intersections with Connected,Cooperative and Automated Vehicle Technology

Oral Session 2
Multi-Camera System for Traffic Light Detection: About Camera Setup and Mapping of Detections

WS04-0: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems

Welcome and opening remarks

WS04-1: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems

Keynote: Smart move

Abstract: In our research we look at demand-responsive urban transportation: next-gen smart mobility. The focus is on concepts of economic feasibility, and on raising its attraction and trust. I'll give an overview of the ongoing work.

Bio: Stephan Winter is Professor and Discipline Leader, Geomatics, at the Department of Infrastructure Engineering, The University of Melbourne. He is a world-wide recognized researcher in spatial information science, specializing in human spatial cognition and interaction, and intelligent mobility / transport. He leads the smart city agenda of the Melbourne School of Engineering. He is member of IEEE.

WS04-2: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems

Session 1: Demand estimation for MoD
Analysis on Supply and Demand of Shared Autonomous Vehicles Considering Household Vehicle Ownership and Shared Use
A Critical Analysis of Travel Demand Estimation for New One-Way Carsharing Systems

WS04-3: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems

Session 2: Operations of MoD
DeepSpeedometer: Vehicle Speed Estimation from Accelerometer and Gyroscope Using LSTM Model
Rebalancing Shared Mobility-on-Demand Systems: a Reinforcement Learning Approach
Stackable VS Autonomous Cars for Shared Mobility Systems: a Preliminary Performance Evaluation

WS04-4: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems

Session 3: Simulation of MoD
Agent-Based Simulation of a Shared, Autonomous and Electric On-Demand Mobility Solution
Development of a Simulation Platform to Implement Vehicle Routing Algorithms for Large Scale Fleet Management Systems
City-wide Shared Taxis:A Simulation Study in Berlin

WS04-5: Modelling, Analysis and Control of Intelligent Mobility-on-Demand Systems

Closing remarks

WS05-0: Deep Learning for Autonomous Driving

Opening Remarks

WS05-1: Deep Learning for Autonomous Driving

Invited Talk #1 : Virtual Worlds for the Perception and Control of Self-Driving Vehicles

Deep learning has emerged as a key enabling technology for developing autonomous driving under two main paradigms. On the one hand, we can find modular approaches with explicit tasks for detecting the free road, the dynamic objects, etc. and then plan for a safe vehicle maneuver according to particular control laws.  These tasks rely on deep models. On the other hand, there are end-to-end driving approaches able to output vehicle motion commands by processing the raw data with a deep model, without explicit tasks for detecting the free road, the dynamic objects or running a specific control law. In the former case, perception and control are separated, in the latter they are not; but in both cases some sort of ground truth (GT) is required for training and testing the self-driving AI agents. In fact, deep models are very data hungry (raw and GT). Our research group at the CVC/UAB has been investigating during the last eight years how virtual worlds and simulation can help training and testing advanced driver assistance systems (ADAS) first, and self-driving AI agents nowadays. In this talk we review all this research presenting last news about our well-known SYNTHIA environment, as well as new simulation environments such as CARLA. Focusing on how these environments can contribute to both self-driving paradigms, as well as to other tasks related to vision zero traffic accidents.

WS05-2: Deep Learning for Autonomous Driving

Invited Talk #2 : From Deep Learning to Autonomous Driving

Deep Learning and Autonomous Driving are emerging research topics that become more and more interweaved. Besides continuously upcoming new achievements in learning, we see successful approaches in the domain of autonomous vehicles reaching from learning individual components of the overall system, over several components at once, up to directly learning vehicle control commands from visual sensor input. However, when bringing these approaches to real world autonomous driving, the question on how to safely incorporate those techniques into production-grade vehicles arises. This issue can be considered manageable when learning techniques are used for highly dedicated perception tasks with a single learning step, but becomes more complex with increasing responsibilities of the learning system. If vehicles are controlled by learning-based approaches directly, rare failures will have immediate impact and thus more severe consequences. This emphasizes the importance for research towards the additional integration of expert knowledge in order to constrain vehicle behaviors in terms of safety and reliability.  his presentation will outline the power and computational expressiveness of deep learning approaches in autonomous driving. Furthermore, the potential of current end-to-end learning concepts for vehicle control using supervised and unsupervised methods will be discussed. This will be followed by potential methods to combine such learning algorithms with model driven and probabilistic approaches in order to gain comprehensiveness and accountability. Experiments and results from real world scenarios with our autonomous research car CoCar will be shown.

WS05-3: Deep Learning for Autonomous Driving

Invited Talk #3 : Faster Convolutional Architecture Search for Semantic Segmentation

Designing deep learning architectures is a complex task and requires expert knowledge. Convolutional Neural Networks involve different network topologies, layers, layer parameters. In order to automate the design process our approach is based on MetaQNN, the Q-learning agent design architectures by selecting CNN layers. We extend the approach for semantic segmentation task, where architecture involves encoder-decoder layers. To speed up the search process, we use a Hyperband-like technique. Our experiments are evaluated on CamVid urban street scene semantic segmentation dataset. The architectures designed by the Q-learning agent for semantic segmentation task are better than some commonly used hand-designed architectures with similar number of parameters.

WS05-4: Deep Learning for Autonomous Driving

Oral Paper Session #1
Learning Temporal Features with CNNs for Monocular Visual Ego Motion Estimation
Speed and Steering Angle Prediction for Intelligent Vehicles Based on Deep Belief Network
Adding Navigation to the Equation: Turning Decisions for End-to-End Vehicle Control

WS05-5: Deep Learning for Autonomous Driving

Poster Session
Intent Prediction of Vulnerable Road Users from Motion Trajectories Using Stacked LSTM Network
Feature Detectors for Traffic Light Recognition
An LSTM Network for Highway Trajectory Prediction
Deep Fully Convolutional Networks with Random Data Augmentation for Enhanced Generalization in Road Detection
Deep Convolution Long-Short Term Memory Network for LIDAR Semantic Segmentation
Fast Semi-Dense 3D Semantic Mapping with Monocular Visual SLAM

WS05-6: Deep Learning for Autonomous Driving

Invited Talk #4 : Deep Learning Implementation and Optimization on DRIVE PX Autonomous Driving Platform

DL (Deep Learning) offers a superhuman image recognition performance today, and this is the strong motivation to apply this technology for autonomous driving perception. Additionally the capability of driving smoothly in various conditions is expected to be implemented by the End-to-End DL. NVIDIA has developed an autonomous driving platform Drive PX that include both of those DL features. This platform also implement massive parallel computing that is necessary to execute human written algorithms for image processing, localization, path planning etc. In this talk, we will introduce our DRIVE PX platforms based on Parker and Xavier SoCs (System On a Chip), DL and GPU computing implemented on those platforms and key technologies including TensorRT, DLA(DL Accelerators) that enables the DN execution in a practical performance/watt level.

WS05-7: Deep Learning for Autonomous Driving

Invited Talk #5 : Towards Deep Understanding of the Vulnerable Road User

Daimler introduced an advanced set of driver assistance functions starting from 2013 in its Mercedes-Benz S-, E-, and C-Class models using stereo vision. Already included was a pedestrian safety component which facilitates fully automatic emergency braking. During the last years our research focused on developing the next-generation of active pedestrian and cyclist safety systems. These systems extract higher-level visual cues and use more sophisticated motion models for path prediction - enabling a deeper traffic situation understanding. The potential to react earlier in dangerous traffic situations, without increasing the false alarms, make such systems essential for autonomous driving in our cities. In this talk we provide an overview over current research projects and show how these works can greatly improve the perception of difficult traffic situations by making use of modern machine learning methods.

WS05-8: Deep Learning for Autonomous Driving

Invited Talk #6 : Deep learning journey: from driver assistance to driver replacement

According to recent expectations, fully autonomous commercial vehicles should start rolling in the streets in less than three years. This is the fruit of at least a decade of research and development work. In this talk, we will trace this interesting journey focusing on its machine vision aspect. It all started with detecting individual objects from images and videos - especially pedestrians. Soon after, with the successful introduction of convolutional neural networks, pixel segmentation of multiple objects simultaneously was deemed possible to assist drivers in unfavorable weather conditions. Furthermore, another line of research focused on monitoring the driver to detect alertness and concentration. Recently, deep learning has been active in estimating the correct steering angle from camera feeds. Hopefully, integrating all these components together can help enjoying the ride safely in autonomous cars be very soon.

WS05-9: Deep Learning for Autonomous Driving

Oral Paper Session #2
Probabilistic Vehicle Trajectory Prediction over Occupancy Grid Map via Recurrent Neural Network
A Survey on Leveraging Deep Neural Networks for Object Tracking
Imitation Learning for Vision-based Lane Keeping Assistance

WS05-10: Deep Learning for Autonomous Driving

Closing Remarks & discussion

WS06: Artificial Transportation Systems and Simulation

Short-term Traffic Flow Forecasting Based on Wavelet Transform and Neural Network
Hybrid Calibration of Agent-Based Travel Model Using Traffic Counts and AVI Data
Teaching Self-Driving Cars to Dream: A Deeply Integrated, Innovative Approach for Solving the Autonomous Vehicle Validation Problem
Automated Scenario Generation for Regression Testing of Autonomous Vehicles
Efficient Expression and Deep Analysis Platform of Massive Traffic Video Data
Coffee break
Relaying Algorithm Based on Soft Estimated Information for Cooperative V2X Networks
Behavioral Trajectory Planning for Motion Planning in Urban Environments
Market-based Approach for Cooperation and Coordination Among Multiple Autonomous Vehicles

WS07-0: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility

Opening Session

WS07-1: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility

SESSION I - Demand Patterns

"Future Public Transport Fare Structures: Behavioural and Social Impacts of IST"

Automatic Fare Collection (AFC) systems such as smart cards, or mobile phone based charging are allowing us to collect long-term data records of user behaviour. This information can be used to optimize a transport network including network structure and fares, though understanding and predicting demand elasticities remains a challenging topic. At the same time, new AFC technologies are not only a data source but also allow operators to introduce complex fare structures including spatial, temporal and user group distinguished charging. The talk will discuss some of these trends and possibilities, such as changes from flat or zonal to distance-based fares and price capping. Potential impacts for single travelers as well as specific parts of a hypothetical city will be discussed to illustrate that some cities understandably hesitate to use the full potential of now possible charging structures.

WS07-2: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility

SESSION I - Demand Patterns
Daily Metro Origin-Destination Pattern Recognition Using Dimensionality Reduction and Clustering Methods

WS07-3: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility

SESSION I - Demand Patterns

Predicting the Next Trip of Individual Public Transportation Passengers * by Zhan Zhao, Haris N. Koutsopoulos and Jinhua Zhao

WS07-4: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility

SESSION I - Demand Patterns
Short & Long Term Forecasting of Multimodal Transport Passenger Flows with Machine Learning Methods

WS07-5: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility

SESSION II - Operations and Traffic Dynamics
Novel C-ITS Support for Electric Buses with Opportunity Charging
Adjusting Bus Timetables Considering Observed Delays and Passenger Numbers
Fault Diagnosis Method of the On-board Equipment of Train Control System Based on Rough Set Theory
Non-stationary Traffic Flow Prediction Using Deep Learning

WS07-6: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility

SESSION III - Network Dynamics
Resiliency Assessment of Urban Rail Transit Networks: A Case Study of Shanghai Metro
Using Mobile Phone Data Analysis for the Estimation of Daily Urban Dynamics
Simulation of Demand and Supply of Urban Rail in a Multimodal Environment

WS07-7: The Third International Workshop on Intelligent Public Transports -- Toward the Next Generation of Urban Mobility

Closing Session

WS08: Behavioural Change Support Intelligent Transportation Applications

Implementation of a Mobility Behavior Change Support System in Manila Philippines
Exploring Influential Factors on Transition Process of Vehicle Ownership in Developing Asian City- A Case Study in Bogor City Indonesia
Pricing as a Tool to Influence Behaviour
Personalized Persuasion Services for Route Planning Applications
Classifying Aggressive Drivers for Better Traffic Signal Control
Coffee break
Impact of Autonomous Vehicles on Pedestrians' Safety
Assessing Passenger Feedback Reliability in Crowd-Sourced Measurement of Transit Ride Quality
Reliable Feeder Bus Schedule Optimization in A Multi-mode Transit System
Incident Detection Using Data from Social Media

WS09: Transportation 5.0

Transportation 5.0 in CPSS: Towards ACP-based Society-Centered Intelligent Transportation
CPSS Models and Spatiotemporal Collaborative Optimization of Urban Public Transport Dynamic Network
A Hybrid Deep Learning Approach for Urban Expressway Travel Time Prediction Considering Spatial-Temporal Features
DeepTrend: A Deep Hierarchical Neural Network for Traffic Flow Prediction
Risk Diagnosis Model for the Disaster Spread of Highway Traffic Network
Memetic Algorithm with Adaptive Local Search for Capacitated Arc Routing Problem
Lane-changing Rules Based on Cellular Automaton Traffic Flow Model Under V2V Environment
Improved Single Image Haze Removal for Intelligent Driving
Scene-Specific Pedestrian Detection Based on Parallel Vision
Visual Lane Detection Based on the Hierarchical Grouping Structure
The ParallelEye Dataset: Constructing Large-Scale Artificial Scenes for Traffic Vision Research
An Improved Method of Real-time Camera Pose Estimation Based on Descriptor Tracking
Accelerate the Autonomous Vehicles Reliability Testing in Parallel Paradigm
An Accelerated Testing Approach for Automated Vehicles with Background Traffic Described by Joint Distributions

WS10-0: International workshop on Large-Scale Traffic Modeling and Management

Openning

WS10-1: International workshop on Large-Scale Traffic Modeling and Management

Macroscopic fundamental diagram for coordinated ramp metering

Recent studies on aggregated traffic modeling of urban networks have shown the existence of a well-defined and low-scatter Macroscopic Fundamental Diagram (MFD) that links network aggregated flow and density. However, the MFD of freeway networks typically exhibits high scatter and hysteresis loops that challenge the monitoring capability and control performance of MFD-based congestion management schemes in freeways. This paper investigates the effect of density heterogeneity and capacity drop on characteristics of freeway MFD based on field traffic data. In addition, we introduce a model to capture the evolution of density heterogeneity which is essential to reproduce the dynamics of freeway MFD accurately. The proposed model is incorporated in a hierarchical control structure for coordinated ramp metering on a freeway stretch with multiple congestion bottlenecks and on- and on-ramps. At the upper level, a model predictive control (MPC) approach is developed to optimize total inflow from on-ramps to the freeway stretch, where the prediction model is formulated based on the proposed MFD model. The lower level controller distributes the optimal total inflows to each on-ramp of the freeway based on local traffic state feedback. The proposed ramp metering framework shows desirable performance to reduce the total time spent and eliminate congestion. The control approach is compared with other coordinated ramp metering controllers based on the MPC framework with different prediction models (e.g. CTM and METANET). Simulation outcomes highlight that the MFD-based hierarchical controller (i) is better able to overcome the modeling mismatch between the prediction model and the plant (process model) in the MPC framework and (ii) requires less computation effort than other nonlinear controllers.

WS10-2: International workshop on Large-Scale Traffic Modeling and Management

Hierarchical management of large-scale networks via path assignment and regional route guidance

Alleviating congestion via manipulation of traffic flows or assignment of vehicles to specific paths has a great potential in achieving efficient network usage. Motivated by this fact, this paper proposes a hierarchical traffic management scheme that combines a network-level regional route guidance model predictive control (MPC) scheme with a region-level subregional path assignment mechanism. The route guidance MPC optimizes network performance based on actuation via regional split ratios, whereas the path assignment mechanism recommends paths for vehicles to follow, realizing the regional split ratios sent by the MPC in order to achieve said performance. Simulation results in a detailed 49 subregion model indicates potential of the proposed scheme in achieving coordination and efficient use of network capacity, leading to increased mobility.

WS10-3: International workshop on Large-Scale Traffic Modeling and Management

A macroscopic approach for optimizing road space allocation of bus lanes in multimodal urban networks through simulation analysis: An Application to the Tokyo CBD Network
A Macroscopic Approach for Optimizing Road Space Allocation of Bus Lanes in Multimodal Urban Networks Through Simulation Analysis: An Application to the Tokyo CBD Network

WS10-4: International workshop on Large-Scale Traffic Modeling and Management

3D clustering methods to estimate travel times at large urban scale

This presentation is about partitioning link speed data every 10 min into 3D clusters that proposes a parsimonious sketch of the congestion patterns at the whole city scale. Days with similar patterns are then gathered using consensus clustering methods to produce a unique global pattern that fits multiple days. We show that Amsterdam's network over 35 days can be synthesized into only 4 consensual 3D speed maps with 9 clusters. A new travel time estimation method can be derived from such patterns. By matching the current observation to historical consensual 3D speed maps, an efficient real-time method is designed. It successfully predicts 84% trips travel times with an error margin below 25%.

WS10-5: International workshop on Large-Scale Traffic Modeling and Management

A Case Study of a Continuous Flow Intersection and Its Impact on Public Transport
A Case Study of a Continuous Flow Intersection and Its Impact on Public Transport

WS10-6: International workshop on Large-Scale Traffic Modeling and Management

Off the Shortest Path: Betweenness on Street Network Level to Study Pedestrian Movement
Off the Shortest Path: Betweenness on Street Network Level to Study Pedestrian Movement

WS10-7: International workshop on Large-Scale Traffic Modeling and Management

Multi-Agent Based Road Traffic Control Optimization
Multi-Agent Based Road Traffic Control Optimization

WS10-8: International workshop on Large-Scale Traffic Modeling and Management

Bus Arrival Time Prediction at Any Distance of Bus Route Using Deep Neural Network Model
Bus Arrival Time Prediction at Any Distance of Bus Route Using Deep Neural Network Model

WS10-9: International workshop on Large-Scale Traffic Modeling and Management

ANALYSING THE IMPACTS OF TRAFFIC STATES AND OD-PATTERNS ON THE PARTITION OF LARGE-SCALE URBAN ROAD NETWORKS

With the expansion of cities and the increasing of traffic demand, the scale of the physics of congestion increases. The macroscopic fundamental diagram (MFD) is widely applied for understanding these physics, and serves as a tool monitoring and control of large-scale networks. The MFD is also becoming a topic of increasing concern. The existence of a well-defined MFD requires spatial homogeneity in traffic distribution in a region, which necessitates transportation network partitioning an essential step in MFD-type applications. This paper moves forward with this direction and investigates the partitioning criteria and methods. Firstly, two partition factors, namely link average density and average speed are defined; Secondly, an extended NCut algorithm is built up and partitions are performed based on the two factors respectively. Then, both partitions are applied and tested in the Sioux Falls network. We employ a variance coefficient to evaluate the performance and a diversity coefficient to qualify the sensitivity of the proposed partitions to the variation of traffic demand. The results reveal that (1) the two methods shall not be regarded as equivalent for obvious differences quantified up to 30%, (2) the density-based method shows advantages in effective and steady identification of traffic patterns under slow-varying demand scenarios and (3) a threshold on the variation of demand can be found, exceeding which the demand profile experiences significant change and the speed-based partition method outperforms the density-based method with respective to partitioned homogeneity. This paper is among the few pioneer works targeting the issues of network partition based on the MFD state variables.