Abstract

it becomes imperative to understand the human behavior in an environment with growing number of intelligent and autonomous systems. Public awareness of safety and privacy of individuals have increased steadily in recent years due to the presence and mobility of people in public places. Therefore, excessive surveillance cameras have been deployed in the public places, which results in huge surveillance data. Trajectory predictions based on observational data is an important area of research in the field of computer vision and behavioral analysis of crowds. The aim of this paper is to suggest an effective direction prediction algorithm. Therefore, we use ensemble learning techniques that play a great role when the best performance on a predictive modeling is the most important outcome. We use different ensemble models and choose the one that gives highest accuracy. We test models on bidirectional pedestrian flow moving in a corridor. Accuracy of different ensemble models on trajectory data range from 83% to 97%, where these ensemble models give high accuracy and less error as compared to simple machine learning algorithms.