A Comparative Study of Artificial Intelligence based Vehicle Classification Algorithms used to Provide Smart Mobility

Abstract

Due to the rising number of vehicles on the road and the limited resources supplied by current infrastructures, traffic problems are becoming more prevalent. Signalized junctions are the prime locations of congestions where commuters need to wait for long time in front of the signals to get their turn to move. This leads to several issues including wastage of time, additional fuel consumption, and green gas emissions. Optimization of traffic signals based on traffic behavior is widely explored topic in which vehicle detection and classification is one of the leading areas of research of Intelligent Transportation System (ITS). Among the technologies Artificial Intelligence (AI) has emerged as a giant in which vehicle classification has developed as a prominent subject of study because of its usefulness in several applications such as traffic control and surveillance, security systems, traffic congestion, avoidance, and accident prevention. Numerous algorithms and techniques for classifying vehicles have been proposed and implemented so far globally which mimics human intelligence. The goal of the paper is to familiarize the reader with the existing AI-based vehicle classification algorithms and to give a comparison of various vehicle detection and classification methods. The existing vehicle classification algorithms are summarized under two categories based on input type i.e., image or video. When the technologies such as AI, image processing, data mining and sensors are combined, the ITS can observe the road, initiate autonomous vehicle detection and thereby control traffic on road efficiently.