Smart Sensing for Traffic Monitoring
Growth in urbanisation, particularly in emerging economies, is causing increased traffic congestion and affecting environmental conditions in cities. Cities need to manage this growth in traffic in an efficient way. Intelligent infrastructure for traffic monitoring and sensing offers a potential solution, and so this book explores the prospective role of this approach in managing congestion, the established and emerging related technologies, and routes to effective implementation. Intelligent infrastructure will also play an important role in the future automation of vehicles. Onboard sensor technology is well-established, but higher levels of vehicle automation are difficult to achieve without additional sensing technology on the infrastructure side. Thanks to recent advances in device and software technology, new and innovative approaches to intelligent sensing infrastructure, both fixed and onboard, can be implemented. The development and deployment of automated vehicles is a hot topic, with early versions potentially entering the market within 5-8 years, so as an overview of developments in the field this book is timely and relevant. The book systematically covers the key elements of intelligent infrastructure. It begins with the architectures and projects in key regions, and then continues with coverage of novel technologies for vehicle, bicycle and pedestrian detection using different kinds of recognition technology. The third part describes technology for detecting traffic conditions such as incidents, illegal parking and adverse weather. Smart Sensing for Traffic Monitoring offers methodically presented information for researchers, practitioners, and advanced students with an interest in the technologies behind intelligent infrastructure and traffic monitoring.
Inspec keywords: image sensors; Global Positioning System; road vehicle radar; pedestrians; image recognition; object detection; optical radar; traffic engineering computing; road safety; bicycles; intelligent sensors; stereo image processing; computerised monitoring; cameras; cloud computing; intelligent transportation systems; video surveillance; road traffic
Other keywords: traffic state sensing; roadside unit; vulnerable road users; LIDAR system; cloud applications; incident detection; on board unit; intersection vehicle detection; Japan; traffic counting; phased-array weather radar; Singapore; cooperative intelligent transport systems; stereo camera; smart mobility; heavy precipitation sensing; GNSS-based traffic monitoring; surveillance camera; crowd density estimation; cycle traffic; RADAR system; smart sensing; Europe; OBU
Subjects: Computer vision and image processing techniques; General electrical engineering topics; Computerised instrumentation; Video signal processing; Image sensors; Image recognition; Computerised instrumentation; Intelligent sensors; Control engineering computing; Satellite communication systems; Road-traffic system control; General and management topics; Intelligent sensors; Signal processing and detection; Information networks; Radar equipment, systems and applications; Traffic engineering computing
- Book DOI: 10.1049/PBTR017E
- Chapter DOI: 10.1049/PBTR017E
- ISBN: 9781785617744
- e-ISBN: 9781785617751
- Page count: 253
- Format: PDF
-
Front Matter
- + Show details - Hide details
-
p.
(1)
-
Part I: Regional activities
1 Japan perspective
- + Show details - Hide details
-
p.
3
–19
(17)
The article discusses the history of intelligent transport system (ITS) development in Japan, focusing on the following technologies: infrastructure sensors and driving assistance using vehicle-to-infrastructure (V2I); expressway and ordinary road case studies; and driving safety assistance using vehicle-to-vehicle (V2V) communication.
2 European perspective of Cooperative Intelligent Transport Systems
- + Show details - Hide details
-
p.
21
–41
(21)
This chapter presents C-ITS development and deployment from the European perspective. The first section provides a brief overview of the history of C-ITS development and deployment in Europe, through EU (European Union)-funded projects and national initiatives, which are joint research and/or innovation actions of industry partners, authorities and academia, at either a Europe-wide or a national scale. Following sections successively introduce the ITS deployment platform initiated by the European Commission (EC), the C-Roads initiative and main activities of the EU member states, C-ITS architecture, and C-ITS services, use cases and operational guidelines developed in Europe.
3 Singapore perspective: smart mobility
- + Show details - Hide details
-
p.
43
–60
(18)
Together with the public-transport -focused land transport strategy for Singapore, ITS has been a useful and effective technology to keep traffic on the road network flowing. The respectable traffic situation that Singapore enjoys today is evident from various third -party reports such as the INRDC 2017 Global Traffic Scorecard and the McKinsey & Company's report on urban transportation systems of 24 global cities. The former ranked Singapore 181st out of 200 cities for congestion, with less than 10 h spent in congestion during peak periods in 2017. McKinsey & Company's report has Singapore as a top city in its overarching urban mobility ranking. Nevertheless, there is still much that can be done to improve the travel experiences of commuters in Singapore.
-
Part II: Traffic state sensing by roadside unit
4 Traffic counting by stereo camera
- + Show details - Hide details
-
p.
63
–80
(18)
Computational stereo vision introduced by Marr is one of the powerful methods to extract three-dimensional (3D) information using multiple images. By applying stereo vision, many kinds of traffic monitoring systems have been developed. In this chapter, we discuss configurations and characteristic of stereo-based traffic monitoring systems especially for the roadside unit.
5 Vehicle detection at intersections by LIDAR system
- + Show details - Hide details
-
p.
81
–96
(16)
Intersection monitoring using LIDAR (light detection and ranging) system has been performed to a limited extent until recent years because of the higher cost of the LIDAR hardware compared with cameras and the reason that the higher cost did not promote the development of point cloud technology. However, opinions on LIDAR system are changing. Autonomous vehicles have been adopting LIDAR system as key sensors, and the robust and reliable measurements from the LIDAR system are highly evaluated. It is boosting the development of point cloud technology. There is also expectation that a high volume market for the autonomous vehicles will lower practical LIDAR price in the future. Vehicle detection at intersection by the LIDAR system is beginning to appear as a new trend.
6 Vehicle detection at intersection by RADAR system
- + Show details - Hide details
-
p.
97
–111
(15)
Millimetre -wave radar is mainly used as a vehicle onboard sensor for monitoring the front of a car and contributes to the spread of ADAS (Advanced Driving Assist System) such as Active Cruise Control and Pre -Crash Safety. In addition, it is considered that when compared with visible camera and LiDAR (laser imaging detection and ranging) that have the similar sensing function, it is superior in environmental robustness and velocity measurement. Furthermore, since it can be mounted inside the emblem or bumper, the fl exibility of installation and design is also a feature of millimetre -wave radar.
-
Part III: Traffic state sensing by on board unit
7 GNSS-based traffic monitoring
- + Show details - Hide details
-
p.
115
–123
(9)
Global Navigation Satellite System (GNSS) probe -based traffic information services are now commonly used across the world. For the purpose of this chapter, we will use the term `GNSS' stands for Global Navigation Satellite System which is a standard generic term for satellite navigation systems. GNSS includes all global navigation systems, including GPS, GLONASS, Galileo and other regional systems. GNSS data is highly accurate, and this service provides a scalable delivery platform that will continue to improve as more data become available.
8 Traffic state monitoring by close coupling logic with OBU and cloud applications
- + Show details - Hide details
-
p.
125
–148
(24)
Traffic state monitoring is important for improving and maintaining smooth traffic. Various factors that affect traffic can be considered as states. These factors to be detected can be not only conventional traffic congestion but also the volume of pedestrians at stops or pedestrian flooding into roads and blocking vehicles' paths and bicycles weaving through traffic. The conventional approach to sensing is to deploy roadside units such as loop coil, sonic sensor, or camera sensor. The new approach is based on a vehicle probe system gathering mainly location and speed data. However, a more sophisticated approach is an image-recognition-based probe system. This approach can directly sense traffic states. Various sensing targets can be detected by developing image-processing logic specific to targets in collaboration with cloud applications. In short, the image-recognition-based onboard unit's (OBU's) probe system has flexibility for further possibilities.
-
Part IV: Detection and counting of vulnerable road users
9 Monitoring cycle traffic: detection and counting methods and analytical issues
- + Show details - Hide details
-
p.
151
–169
(19)
This chapter considers the importance of cycle counting and monitoring and the nature of cycle traffic. It provides an overview of current methods, including surface, subsurface and above -ground detection methods. It discusses some of the analytical issues in connection with the data and provides a forward view into the future of cycle counting.
10 Crowd density estimation from a surveillance camera
- + Show details - Hide details
-
p.
171
–191
(21)
This chapter presents an approach for crowd density estimation in public scenes from a surveillance camera. We formulate the problem of estimating density in a structured learning framework applied to random decision forests. Our approach learns the mapping between image patch features and relative locations of all the objects inside each patch, which contribute for generating the patch density map through Gaussian kernel density estimation. We build the forest in a coarse-to-fine manner with two split node layers and further propose a crowdedness prior and an effective forest reduction method to improve the estimation accuracy and speed. Moreover, we introduce a semiautomatic training method to learn the estimator for a specific scene. We achieved state-of-the-art results on the public Mall and UCSD datasets and also proposed two potential applications in traffic counts and scene understanding with promising results.
-
Part V: Detecting factors affecting traffic
11 Incident detection
- + Show details - Hide details
-
p.
195
–219
(25)
All transport networks are subject to unexpected disruption. The identification that disruption has taken place is usually given the broad title of 'incident detection'. Taking the first question, we can consider an 'incident' to be an unplanned event that results in a set of consequences. These impacts might be a reduction in safety or increase in the level of hazard to road users, including maintenance and vehicle recovery workers who are present in the road; a loss of highway capacity that leads to vehicles being delayed and traffic congestion or, of increasing concern and importance, an unacceptable increase in air pollution. The latter is also a consequence for nearby residents and workers and so affects people who are not travelling. An individual incident may create all of these consequences.
12 Sensing of heavy precipitation—development of phased-array weather radar
- + Show details - Hide details
-
p.
221
–228
(8)
With the advancement of society and global warming in recent years, tragic accidents caused by atmospheric phenomena such as tornadoes and heavy rainfall are on the rise. Development of phased -array Doppler meteorological radar that can observe meteorological phenomena such as cumulonimbus clouds that cause such local heavy rain and tornado in three dimensions with high resolution in time and space is progressing. In such a radar, the time required for observation is dramatically improved in comparison with the conventional method by using the electronic scanning method as compared to the conventional mechanical scanning method, and the world's best performance is realized. In this chapter, we outline the phased -array meteorological radar and its observation results and introduce future approaches to the use of local governments and others.
-
Back Matter
- + Show details - Hide details
-
p.
(1)