Article Preview
TopIntroduction
In metropolitan cities, traffic congestion is increasing rapidly, thereby results in chronic situations in dense downtown areas. Traffic signals play a significant role in the urban transportation system. They control the movement of traffic on urban streets by determining the appropriate signal timing settings. Adaptive traffic signal controllers are principle part of smart transportation system has a primary role to effectively reduce traffic congestion by making a real time adaptation in response to the changing traffic network dynamics. The unmanaged traffic has a number of bad consequences, like delays, fuel consumption and pollution, road rage and problems for emergency vehicles which need early clearance. With the advent of deep learning and neural network applications, it has become easier to solve real life problems more easily. The overall methodology followed by researchers till now is the video camera is implanted on the road to capture traffic conditions and then the algorithm tries to sense the degree of congestion on the road in real time. Based on the traffic condition the algorithm then tries to allocate time for the traffic lights to glow. Hence, the main difference is made by the algorithm used in the process, which decides the accuracy and processing time to define how adaptive the system is. To develop a robust and efficient algorithm for classification of road condition is the biggest challenge now.
Building up a component to foresee the constant traffic stream in urban areas that reduces the trip time utilizing data-mining calculations enhances the exactness, versatility, and flexibility of intelligent traffic applications. This technique consolidates a few versatile data mining strategies, for example, decision tree, association rules, and neural network applications. These methodologies utilize some traffic parameters and authentic information as input. Past traffic information was utilized to foresee the transient traffic stream utilizing the Artificial Neural Network (ANN) (Kumara et al., 2013). The model uses traffic volume, speed, thickness, time and day of the week alongside the speed of every class as data parameters. Video observation information is utilized by Dhingra et al., (2019) for grouping of street traffic and further used Convolution Neural Network (CNN) for classification. Traffic density estimation is acquired from traffic intersection pictures utilizing different artificial intelligence based techniques (Nubert et al., 2018) (combined with CV apparatuses with classes for the thickness of blockage). Another approach was followed by Diao et al., (2019) to create a classification model to foresee the momentary traffic volume in heavy transportation frameworks. The authors introduced a novel hybrid model to precisely figure the volume of travelling streams multi-step ahead. Far reaching factors were viewed as, for example, transient, origin-goal spatial, recurrence and self-likeness, and recorded probabilistic conveyance points of view.