Incident Detection Systems: a Review of the Algorithms

Only available on StudyMode
  • Topic: Fuzzy logic, Neural network, Road transport
  • Pages : 17 (4232 words )
  • Download(s) : 61
  • Published : April 10, 2013
Open Document
Text Preview
Advanced Traffic Engineering and Management

Title of the Term Project:

Incident Detection Systems: A Review of the Algorithms

Table of Contents

Table of Contents1

Background of the Incident Detection Systems2

Significance of the Research in Incident Management Systems3

Literature Review4

Desirable Properties of Incident Detection Algorithms8

The Benchmark Algorithm9

The Fuzzy Logic Based Algorithm10

Future Research Area15

References15

Background of the Incident Detection Systems

Statistics indicated that the UAE looses about AED 5 billion a year to road congestions [1]. The UAE police reports [2 and 3] reveal that the total number of traffic accidents was 10135 in 2008, compared with 8828 in 2007, and 8843 in 2006. The increase in the number of traffic accidents is approximately 15% between the years 2007 and 2008, and 4% between the years 2006 and 2007. Studies in US and France showed that the probability of death in a car accident increases by a factor of 7 when the time taken for assistance to arrive exceeds 20 minutes [4]. Also, first response has become a common component of Emergency Medical Services [5]. Virtually, all algorithms developed to date to recognize the incidents have two major limitations: high false-alarm rates and threshold calibration requirements. Empirically threshold values once exceeded by the measured traffic quantities, the occurrence of an incident are indicated. The false-alarm rates and detection rates clearly depend on the choice of threshold values.

With the statistics and their implications in terms of the economical, social and psychological losses, the UAE authorities have initiated a nationwide effort to establish a strategic traffic safety plan. Amongst the objectives of this plan is the utilization of the advanced technology in developing the activities and plans for traffic operations, education, engineering, and environmental issues associated with the strategic safety plan.

The primary objective of this term paper is the review of some existing and emerging incident detection algorithms from traffic operations perspectives. The tasks of this term paper could be summarized as follows:

a) Brief introductions on the underlying mechanisms of the algorithms developed so far

b) Details of a fuzzy logic based algorithmic procedure and architecture for a reactive highway (freeway or urban settings) monitoring and incident detection in the context of Intelligent Transportation Systems are presented.

c) Some thoughts on the future extensions of the existing models.

Significance of the Research in Incident Management Systems

Traffic congestion has also been a critical issue for all road users in the big cities of the UAE. Motorists have expressed high levels of frustrations and considerable loss of productivity due to congestion. As stated earlier, about AED 5 billion a year is lost due to road congestions [1]. Therefore, significant amount of costs savings of the society would be possible if there in an efficient incident management system with efficient incident detection algorithms. Road incidents are very frequent in the major UAE roadways and it has a major share for the overall yearly road congestions. For example, using the same 2008 economic cost data of the UAE, if the yearly road congestions of the major roads are possibly reduced by at least 1%, 5%, 10%, 15% and 20%, then the respective costs savings would be 50, 250, 500, 750 and 1000 million AED.

Incident detection is the foundation of incident management. Apart from the economic savings, an efficient incident detection algorithm could be used to guide the congested traffic for rerouting efficiently. Other benefits of the incident management system results in improved public health and safety benefits, increased responder service and safety, improved energy consumptions and...
tracking img