Ben Ayed et al. have discussed different cases of the application
of big data technology to transportation in their paper [11].
We summarise the presented cases in Table II. Many of these
make use of commercially available technologies. Commonly
used technology for data capture include: GPS for capturing
location data; video surveillance and image processing technology
for capturing and recognising car license plate data;
traffic sensors for capturing traffic density and flow; ship sensors
for capturing water quality data for navigable waterways;
weather sensors for capturing current weather conditions;
RFID sensors for capturing movement of RFID-tagged items
in logistics; and various software technologies for processing
the collected data. Data processing commonly makes use of
Hadoop and related technologies: Hadoop Distributed File
System (HDFS) for data storage; HBase distributed database;
Hadoop MapReduce for processing data; and others.
The main steps of big data processing for transportation
applications are depicted in Fig.5. These are the familiar steps
of big data processing, beginning with a map operation to
sort/filter input data, then a shuffle (or group) operation to
redistribute the mapped data by output key, and finally a
reduce operation to summarise the data on each node. Input
data includes CCTV road traffic streams, GPS data of moving
vehicles, traffic and weather sensors, and others. The output
depends on the transportation application and may include
assessments of traffic congestion, traffic flow, transportation
efficiency, pollution levels, as well as forecasts of traffic and
pollution trends, among others.
Technical solutions can differ in their degree of complexity.
Many existing applications are very simple. For example, in
Beijing (as well as in other major cities around the world)
road-side electronic displays such as the one in Fig.6 show
traffic conditions of nearby road sections, and have been in
use for several years. Thus a driver may react immediately
and choose the least congested road that leads in the general
direction of travel. However, these systems only cover a
small area and lack the foresight to direct drivers away from