Intelligent System for Accident Detection
Vaishali Nirgude, Aaman Chaudhary , Dheeraj Choudhary , Hemendra Chaturvedi
Computer Department
Thakur College of Engineering and
Technology
Mumbai, India
Abstract— A significant number of fatalities resulting from road accidents occur every day throughout the world. The two effective ways to lower fatalities due to road accident are building intelligent systems to identify accidents and secondly the deployment of first responders on the scene. In recent times we have seen the deployment of such systems in cars. Although being efficient these methods turn out to be costly, require constant servicing and do not serve people other than the one driving the car. On the other hand, the development in smartphones has made them an easy and handy tools to detect accident using sensors already present in them. The majority of accident detection solutions for smartphones rely on the high speed of the vehicle (extracted from the smartphone's GPS receiver) and the G-Force value recorded via the accelerometer and the gyroscope sensor. According to numerous sources, 90% of car accidents on the road take place while the speed is low. Therefore, this work focused on low speed car accident detection in addition to high speed accident detection. It can be difficult to tell if a user is inside or outside of a vehicle, walking or slowly jogging, which is the main challenge in low-speed accidents. In this work, a proposed method that distinguishes between the speed fluctuation of a low speed vehicle and a walking or slowly moving person is used to reduce the impact of this obstacle.
Impact Statement — The proposed system consists of two phases; the detection phase which is used to detect car accident in low and high speeds. Second one being the notification phase which immediately after an accident sends a detailed notification containing location information to the close relative/friend of the person.
Index Terms— AI; intelligent transportation systems (ITS); cognitive science; deep learning; IoT; ResNet; InceptionResnetV2; accident detection; sensors