Real Time Optimized Traffic Management Algorithm

With the development of the modern world, the usage of vehicles has been increased. The need of a vehicle has become a basic need for a family. Therefore, most of the houses among the world own at least one vehicle. Also with the rapidly growing population, the number of vehicles increases. This has caused one of the major problems in many countries all over the world, including Sri Lanka.

Road traffic is a main problem that needs to be solved soon. It effects everyone’s day to day life. Some countries have had terrible traffics for more than one day. There have been some incidents during heavy traffic, such as pregnant women who got stuck in traffic, delivered their babies on the road. Such incidents make people’s lives more uncomfortable.

In Sri Lanka, the most effected traffic areas can be seen around Colombo, Gampaha etc. Mostly traffic starts in the morning because school students and people who work, travel at that time. Again it happens in the evening as well.

To reduce road traffic, the government of the particular country should construct highways, flyovers, wider roads, bridges, place traffic light systems etc. Other than that, people could use public transportation instead of their personal vehicles. Unless they have more than two passengers for their vehicle they should use public transportation. In that case, the number of vehicles can be reduced during traffic time.

When it comes to Sri Lanka, most of people prefer their personal transportation due to the uncomforting experience in public transportation. So, that solution would not be more effective in Sri Lanka. Even though there are some other solutions that has been given already, such as traffic light systems, widening roads, constructing highways and flyovers, the problem in road traffic still remains same.

Engineers and developers all over the world do researches in this phenomenon, expecting a better solution. Some students of SRM University in India has done a research in traffic management. They have assumed Wireless Sensor Networks to be the source of input in their traffic management system. According to their research paper there are two main methods in traffic signal control.

Periodic Signal Control

It uses a predefined sequence and duration of green light for various lanes. It has limited efficiency when it comes to real time traffic scenarios.

Dynamic Signal Control

It adapts the traffic signal in accordance with the dynamic traffic flow.

There are four steps in their proposed algorithm.

• Determining queue length (volume) of traffic.
• Determining the presence of emergency vehicles as well as the scenario.
• Assigning green light to most suitable phase.
• Calculating green light duration for each phase.

  

They have identified a problem which is based on handling prioritized vehicles (Emergency Vehicles). They have proposed a technique as following.

In that technique, they have offered a solution to tackle the worst case condition for managing emergency vehicles.

First, it categorizes emergency vehicles based on standard policies. Then it considers the distance of each emergency vehicle from the intersection point. They have also included an approach for more optimized emergency vehicle management by collaborating Code Messaging Service (CMS) and a GPS Enabled Mobile Device.

They have given their problem formulation as follows         :

• The flow of traffic through the intersections decides the signal phase selection.
• To formulate a solution, assuming that the left-turn is always permitted.
• All eight pairs combine in various combinations which results in 12 possible cases.

They have assumed that the speed of every similar vehicle type is constant, all the vehicles have an installed vehicle node. When assigning green light to most suitable phase, the queue length is calculated as per the input given by the WSN.

They believe that they have tried to provide a solution to a very new problem of Emergency Deadlock Condition.

Their proposed network architecture has 3 Incoming Sensor Nodes (ISN) placed on each incoming lanes. Overall there are 24 sensors.

In our project, we have different and better features than their project. Unlike them, we use image processing, instead of using WSN. In this method, CCTV cameras are used to take road footage to check out the traffic at any instant on that lane. It will take the real time values.

According to this method, the cameras detects the vehicles and then it will give the density of vehicles during a specific time period. In that case, it is possible to detect the vehicle density in all the road of a 4 way junction. Then the traffic light system can generate the light cycles based on the priority.

We will consider the time, to categorize the light cycles. In Sri Lanka, we can identify roughly about 3 types of traffics. 

• 6.00am -8.30am (School + Office)
• 12.30pm – 2.30pm (School)
• 4.00pm – 7.30pm (Office)

In between the above time periods, the traffic seems to be less. So we will arrange four types of light cycles to generate each time period.

Other than that, our traffic management system keeps logged data, in accordance with the vehicle density at each time period. Then it can analyze the logged data and fix a specific light cycle to be generated at a specific time.

Currently, there aren’t any traffic light systems which have our proposed features here in Sri Lanka. Therefore, our project will introduce a successful solution for the traffic in our most populated areas.

Reference:

Partha Sarathi Chakraborty, Prajeeth Nair, Pranshu Raj Sinha and Ishan Kumar Behera, Department of Computer Science and Engineering, SRM University, India. “Real Time Optimized Traffic Management Algorithm” International Journal of Computer Science & Information Technology (IJCSIT) Vol 6, No 4, August 2014.