Article Preview
Top1. Introduction
The number of vehicles operating in the world is increasing day by day to meet this ever-increasing demand. This increased load of vehicles is putting an extra pressure on the system to maintain the roadways leading to wear and tear of the roads. It is getting difficult for the government bodies to maintain each and every road because of the immense traffic. Potholes and distress on road are the major causes of road accidents. Along with this, the other factors such as lack of human attention, weather conditions, automobile error etc. contribute to many fatalities. There are millions of reported road accidents in the world in which thousands of people lose their life every day. In a country like India, it was reported in 2016, that over 150 thousand people died in road accidents (Guardian, n.d.; Suong & Kwon, 2018). In another report, it was stated that over 30% of the accidents caused in India were due to the poor road conditions (that is, potholes on the road, disintegration of road, etc.). Potholes are the circular shape openings or depressions that can be caused by water in the underlying soil surface and weathering of the road surface. They appear when the topmost surface layer of the road has been eroded by busy traffic, thus exposing the concrete base or the structure underneath. Potholes are hazardous for drivers and vehicles when moving at high speeds because it is very hard for the driver to see the potholes on the road surface. Moreover, they possess a threat to the car tyres as they can rupture from the sudden impact leading to the accidents. Due to their small shape and structure, they are not easy for authorities to find and repair. Many accidents occur due to the lack of a driver’s attention on the road. As stated in the reports, using mobile phones contributes to the distraction of driver that causes accidents. Talking on phone, drinking while driving, operating the radio, texting on phone, talking with passenger and the driver reaching behind for some purpose are few of the reasons why a driver might get distracted causing him not to pay focus on the road that results into an accident. It is also the responsibility of the respective Governments to make the road conditions safer for its citizens. The Municipal Corporations of the various Indian cities like Mumbai, Hyderabad invested Rs. 1095 Crores and Rs. 2800 Crores respectively to eliminate the potholes in the roads (Abadi et al., 2016; Mahamkali, 2015). Techniques, such as Machine Learning, Artificial Intelligence, Computer Vision and Internet of Things can be used in the vehicles to avoid and prevent these situations (Huang et al., 2017). Vehicles should come with some sort of integrated infrastructure in a way that this paper suggests in order to reduce the chances of accidents due to human error & road conditions along with capturing the location of these potholes leading to more efficient repair(Huang et al., 2017). This system uses an approach of determining the state of roads as well as the state of driver simultaneously & producing alerts in real-time which even helps in providing data of the roads for improvement along with significant help in driving & ensuring safety (Canny, 1986).
Some of the research gaps found from the existing research are discussed here followed by the motivation and novelty in the proposed work. The major research gap is that many of the research papers discuss only about the techniques to detect the pothole with measurement of one or two performance metrics, but they do not include all the performance metrics (Redmon & Farhadi, 2017). Another gap is related to the partial (and not full) representation of the image i.e. it doesn't look at the whole picture at once and involves several complex phases (Chebrolu & Kumar, 2019). Therefore, SSD and YOLO methods have been used to work on this problem. Third challenge is that the approach used in the previous research works does not have machine learning at its heart (Karthika & Parameswaran, 2018). The last identified gap is that a lot of work has been done for the detection of potholes but very little work has been done to reduce the accidents caused due to potholes and poor road conditions and to provide the pothole’s geographic locations (Redmon & Farhadi, 2017). All these research gaps provided the motivation to the authors to overcome these issues/challenges to ensure the safety and the maintenance of roadways. The novelty of the proposed work is that it is able to identify the road conditions/potholes and warn users, along with the real-time calculation of the exact position in terms of latitude and longitude and there is not any such model yet available.