Design and Implementation of Adaptive Accident-Avoidance System for Autonomous Vehicles 2022-28-0012

Accident Avoidance is an important part of transportation infrastructures, serves to secure pedestrians lives and possessions and keep traffic flow in order. The main aim of the proposed prediction system is to detect the pedestrian crossing by considering various behaviors of the pedestrians in different types of lanes. In this technique the pre-processing of the captured frame and probabilistic branching of the frame are taken care by the Convolutional Neural Networks (CNN) and Long Short Term Memory (LSTM). When the pedestrian crossing is detected, the distance is calculated based on the ratio between the area of ROI and the frame area. Based on the calculated distance, the brake is controlled automatically to stop the accidents in obstructed areas. The proposed technique is beneficial in road safety applications to scale back the unpredictability in determining the behavior of the pedestrians while crossing the road. This technique majorly comprises of Convolutional Neural Networks (CNN) which can produce higher range of accuracy of 96.7% than in histogram-oriented gradients (HOG) and support vector machine (SVM) which produces accuracy of 78-80%.