ADVANCED DATA

ANALYTICS IN
TRANSPORT
Machine learning perspective

PHOTO BY JI-ELLE / PUBLIC DOMAIN

 (i) corporate decision making, planning, and managing
(better utilization of accurate prediction and detection
models aiming to better forecast traffic volume, traffic
AI APPLICATION conditions, and incidents)

IN (ii) improve public transport is also discussed (public

transportation is regarded as a sustainable mode of
TRANSPORT mobility)
(iii) connected and autonomous vehicles, which aims to
enhance productivity by reducing the number of
accidents on highways
TRANSPORTATION: A MULTI-DISCIPLINARY INDUSTRY

Design

Science Policy Making

Planning

Engineering Management
CUTTING-EDGE ISSUES
Cooperative Vehicle Ride sharing Multimodal
Highway Systems Transportation

This Photo by Unknown Author is licensed under CC BY-NC-ND This Photo by Unknown Author is licensed under CC BY-SA
SMART TRANSPORTATION SYSTEM
INTELLIGENT TRANSPORT SYSTEM (ITS)
 KEY Basic Layer: map service and LBS
Upper Layer: decision service and
COMPONENTS marketplace
MAP MATCHING

 Route Planning
 Map Matching
 Positioning
 Traffic

Challenges:
 Large-scale GPS data
 Low-quality GPS data for mobile
service
 Limited among of labeled data
ROUTE PLANNING
H. Wu, Z. Chen, W. Sun, et al. Modeling Trajectories with Recurrent Neural
Networks. IJCAI 2017.
Piotr Mirowski, et al. Learning to navigate in cities without a map. Arxiv 2018.
T. Weber, S. Racanière, D. P. Reichert, et al. Imagination-Augmented Agents for Deep
Reinforcement Learning. NIPS 2017.

Challenges:
 Large-scale transport network
 High query speed
 Accurate result
 Travel distance
 Travel Speed
 Travel Condition
 Driver/Passenger Preference
 Data Driven vs Heauristic
ESTIMATED TIME OF ARRIVAL (ETA)
TRAFFIC ESTIMATION AND FORECASTING
Data source: loop detector, GPS trajectory, camera
Target: traffic speed in real-time and for the future

Calculate ground-truth traffic data: multiple source data fusion,

spatial-temporal spline
TRAFFIC PREDICTION
Input: road network, past traffic information (speed observed, sensors)
Output: traffic speed for the next time steps
DYNAMIC RIDE SHARING

Yongxin Tong et al. A Unified Approach to Route Planning for Shared Mobility, VLDB, 2018.
INTELLIGENT ORDER-DISPATCHING SYSTEM
Every dispatching decision affects future supply (drivers) distribution
Maximize drivers’ collective income through optimized dispatching, while
ensuring good customer experience

Route Planning
 To minimize cost
 To maximize driver efficiency
 To optimize transportation efficiency

ETA
 To estimate the traveling time and
the waiting time of each ride
REAL-TIME TRAFFIC CONTROL SYSTEM
REAL-TIME TRAFFIC MANAGEMENT
DynaMIT 2.0 is a multi-modal
multi-data source driven,
simulation-based short-term traffic
prediction system developed by
SMART-FM. The system contains a
collection of data modules and
algorithms that estimate and
predict network states in real-
time.
1. the modeling of multiple modes
including public transit, mobility-on-
demand services such as Uber/Lyft,
car and ride sharing, etc.;
2. Online calibration of demand and
supply parameters using
heterogeneous data sources;
3. context mining and the scenario
analyzer for unstructured data, e.g.
traffic incident reports;
4. The strategy optimization module for https://its.mit.edu/dynamit
network control strategy optimization.
SMART TRANSPORT BRAIN
PLANNING, DESIGNING AND CONTROLLING TRANSPORTATION
NETWORK STRUCTURES
Incident detection
efficient multiple model particle
filter (EMMPF) to solve the problems
of traffic state estimation

Predictive Models
INCIDENT DETECTION
detect incidents on the highway using both simulated and field data

efficient multiple model particle filter (EMMPF) social media (Twitter, Weibo)
PUBLIC TRANSPORT

 BRIDJ (Boston)
 CB (China)
 Optibus (San Francisco (USA), London
(UK), Tel Aviv (Israel), and Düsseldorf
(Germany)) https://www.bostonglobe.com/business/2017/04/30/bridj-local-demand-bus-service-shutting-
down/56xoGs674wYgyUWdrD9EuO/story.html

https://www.traffictechnologytoday.com/news/public-transit/optibus-gains-additional-
investment-to-further-develop-its-ai-based-mass-transportation-platform.html
INTELLIGENT URBAN MOBILITY

Autonomous vehicles
AUTOMATIC ROAD ENFORCEMENT

Look in the slide notes below for

topics to consider talking about

PHOTO BY MARIORDO MARIO ROBERTO DURAN ORTIZ/LETICIA FERREIRA / CC BY-SA 3.0

ROAD CONGESTION IN URBAN AREAS

Road congestion in urban area is

largely determined by the equilibrium
between demand and supply
capacity. Transport planning usually
deals with the supply capacity,
whereas congestion management may
look for short-term demand incentives.
However, the problem is how to
evaluate the impact of transport
planning and congestion management
scheme to the entire urban traffic
system with higher accuracy,
improved efficiency and visual
demonstration. The ADAIT team aims
to develop a traffic simulation
platform to assist transport planners
and traffic engineers to evaluate the
holistic impact of their own proposals
or borrowed solutions from other cities
to the entire urban traffic system in
Sydney.

PHOTO BY MARIORDO MARIO ROBERTO DURAN ORTIZ/LETICIA FERREIRA / CC BY-SA 3.0

SMART MOTORWAY
The ADAIT team developed Data61’s indigenous smart motorway control system, and evaluated the
cost/benefit of such system for Sydney M4 as well as West Australia Kwinana Freeway. Microscopic simulation
demonstrated that travel time savings can reach 40% during peak hours, with capacity gain equivalent to
adding an extra lane. This means a saving of $22 million per annum on part of M4 alone, potential for more
than $500m p.a. applied to whole system in Sydney.

Figure 7: Modelled freeway section on Sydney M4, with M4/M7 intersection on the background; this is one of the key congestion areas for westbound traffic during peak-hours
PHOTO BY MARIORDO MARIO ROBERTO DURAN ORTIZ/LETICIA FERREIRA / CC BY-SA 3.0
RMS KEY ROAD PERFORMANCE REPORT

This project fuses RMS data

from multiple sources to
produce meaningful contents
and support interactive user
interface. The back-end system
fuses SCATS and GPS data to
produce travel time distribution
for each of the 127 key roads
at 15-minute interval, 24/7.
The advanced reporting
architecture will be able to
generate offline and online
reports for road performance
evaluation. Figure 6 presented
the online reporting interface
for the RMS Key Roads
Performance.
Figure 6: Online report for Key Road Performance in NSW
PHOTO BY MARIORDO MARIO ROBERTO DURAN ORTIZ/LETICIA FERREIRA / CC BY-SA 3.0
SYDNEY CBD MOBILITY MODELLING

This is the NSW Roads and

Maritime Services’ project on
evaluating the impact of both
planned and unplanned closures
of George St on traffic volume,
congestion and travel time at
intersections and road links in and
around the Sydney CBD (Figure
5).
Focusing on quantify the impact
of a light-rail construction and
operation in the Sydney CBD on
traffic flows, we have developed
the state-of-the-art transport
zoning method and origin-
destination demand modelling
tool, which are integrated with
existing traffic assignment models.

Figure 5: A snapshot of Sydney CBD traffic based on degree of saturation

PHOTO BY MARIORDO MARIO ROBERTO DURAN ORTIZ/LETICIA FERREIRA / CC BY-SA 3.0

TRAFFIC WATCH

For incident management purpose, we

have developed the live system
displaying the current location and status
of all incidents from multiple sources
including social media. Figure 4 presented
the web service and interface for the
real-time view of current incidents in
Victoria. Heat map and cluster map were
also generated.
This system utilises advanced web
technologies and state-of-the-art machine
learning (ML) algorithms such as Support
Vector Machines and Conditional Random
Fields to classify the tweets relevant to
traffic incidents.

Figure 4: Incident detection and visualisation

PHOTO BY MARIORDO MARIO ROBERTO DURAN ORTIZ/LETICIA FERREIRA / CC BY-SA 3.0
INCIDENT DURATION PREDICTION

The research project named

“Decision Support for
Incident Management” (also
known as Machine Learning
Assessment of Road
Incidents) with NSW
Transportation Management
Center mainly focused on
machine learning methods
for incident duration
prediction and outlier
detection. The best
performance was achieved
with Gradient Boosted Trees
accompanied by advanced
sampling algorithms.

PHOTO BY MARIORDO MARIO ROBERTO DURAN ORTIZ/LETICIA FERREIRA / CC BY-SA 3.0

PREDICTING NEAR FUTURE TRAFFIC JAMS AND CONGESTION
We have introduced an algorithm
based on the combination of association
rule mining and dynamic Bayesian
network to construct causality trees from
congestions and estimate their
propagation probabilities based on
temporal and spatial information.
Frequent sub-structures of these
causality trees reveal not only recurring
interactions among spatio-temporal
congestions, but potential bottlenecks or
flaws in the designs of existing traffic
networks.
This algorithm was then utilised to
discover real-time ‘hot spots’ for
congestions in the Sydney CBD,
Australia. Figure 2 shows the traffic
conditions for a typical morning in
Sydney CBD.

PHOTO FROM HTTPS://RESEARCH.CSIRO.AU/DATA61/ADVANCED-DATA-ANALYTICS-IN-TRANSPORT-MACHINE-LEARNING-PERSPECTIVE/

PREDICTING SECONDARY INCIDENT LOCATION
(WARNING AND PLANNING)

In the special conditions such as night time or

heavy rain, the occurrence of one incident
can be used to predict the secondary
incident nearby. We have developed an
algorithm based on frequent item set and
association rule mining to discover the
likelihood of incident propagations from
both spatial and temporal perspectives
which:
 Determine the main cause (incident) of the
dangerous conditions.
 Predict incident-prone conditions on sections of
road when first incident happened.
 Improve conditions by actionable suggestions to
prevent the secondary incidents from happening,
e.g. applying traffic management measures.

PHOTO FROM HTTPS://RESEARCH.CSIRO.AU/DATA61/ADVANCED-DATA-ANALYTICS-IN-TRANSPORT-MACHINE-LEARNING-PERSPECTIVE/

 ANN as a “block box” – overcome by combining NN
with traditional techniques as a hybrid solution
LIMITATIONS Ability to reach the finest optimum solution for the
raster algorithms AI tools
Computation complexity of AI Algorithms
COUNTERMEASURES
Look in the slide notes below for
topics to consider talking about

PHOTO BY FLOYDIAN (TALK) / CC BY-SA 3.0

 BY COUNTRY
Look in the slide notes below for topics to
consider talking about

PHOTO BY PROF TPMS / CC BY-SA 4.0

POSSIBLE DATA
SOURCES